Space Combat Mechanics Guide

[visage]

<edit- Updated to accommodate changes in 1.2b

9/06/2004 edit: Visibility finally figured out and included>

I'm writing this document as an attempt to consolidate what is known about space combat in MoOIII into a single coherent whole.

This post will explain how combat works in some detail. The post that follows it will contain charts (or links to charts) of the various ingame values that are relevant to space combat. The third post will touch upon various implications brought out by the information in the first two.

I will be assuming that you the reader have a minimum of familiarity with MoOIII. :)

If you find errors, let me know. I care more about factual errors than typos, of course. :)

If you want to snarf the information here, go right ahead. If you're going to snarf the actual text, it'd be nice if you actually attributed it.


Starships versus System Ships versus Orbitals

Starships are the default for ship stats. Hull size for hull size, orbitals cost 60% of the AUs and get 150% of the space, System ships cost 75% of the AUs and get 133% of the space.


Armor, Shields, and Hit Points

Things that get shot at by default have Hit Points equal to their space. For some reason, ships have space available equal to one less than their listed space; the amount of damage to distroy a hull with a listed space of 200 is 200 points, but you can only put 199 space of stuff in the hull (and Orbitals and System ships have more HP than starships). Fighters have HP equal to their base space (modified by weapon mods), which is 66% of the space the interceptor form takes up and 33% of the space that the SCF form takes up; this is entirely dependent upon the weapon type of the fighter. Missiles have HP equal to the space a single missile takes up (without counting in the +1 per-rack surcharge; missile mods do modify missile HP, though).

The number of armor points a ship, fighter, or missile has is based on its armor type and armor thickness. Armor for fighters and missile always has the same "thickness." Armor also has a deflection value, established by armor type and modified by thickness. Armor never takes up space, regardless of its thickness.

Shields have three combat-relevant stats, apart from their size and cost. Each shield technology has a stopping percentage (StpPow), a number of shield points (ShdStr), and a recharge rate (Recharge), and except for the Damper Field they all default to consuming 5% of the hull space of the ship they're on (so shielding an Orbital takes up more space than shielding a Starship). Shield generator sizes modify the shield points, the recharge rate, and the space, but don't change the stopping percentage. Missile and fighter shields have their own shield generator "size."

Note: Shields and armor for fighters and missiles don't take up space.


How damage is applied to a target:

1) Take a particular weapon hit (so, a volley from 2 LFG's that fire 5 times each time they fire would go through this 10 times).
2) First, apply the effects of shields. The shield absorbs a fraction of the damage equal to its stopping percentage.
Subtract that fraction from the shield points remaining on the shield.
The remainder, plus any absorbed damage that was in excess of the remaining shield points, is passed on to the next step.[/list=a]3) Now, apply the effects of armor. [list=a] Pick a random number between 0 and 1. If the random number is greater than the ArmrPier value of the weapon, skip to step 4.
If the amount of damage remaining is less than the deflection value, the remaining damage is wasted. If shields were damaged, display a blue number equal to the number of shield points lost.
Take a random number between 0 and 1. If the random number is less than the ratio of the remaining armor to max armor, the remaining damage hits the armor: subtract that many armor points. Display a yellow number equal to the amount of damage that hit armor. If all the armor is destroyed at this point, pass on to hit points all the damage that hit armor (not just that in excess of the amount needed to destroy the armor).
If the random number is not less than the ratio, go on to the last step.[/list=a]4) The damage falls on the hit points of the target. Subtract the remaining damage from the hit points. Display the lesser of that number and the number of HP that were present to take the damage. If hit points are reduced to or less than 0, the target blows up.


In actuality, you only display damage per-volley, based on the furthest penetration achieved by that volley.

Note: the shield recharge rate is misrepresented in game as the number of shield points regenerated per 5 seconds; the actual rate is per second.

Note: When you wipe out a target's armor all the damage that hit the armor is passed on to the Hit Points of the target, not just that which is in excess of the armor that was remaining. This strongly suggests to me (though I haven't tested it) that wiping out all the shields points has similar behavior.

Direct Fire Weapons

Accuracy: The change to hit at or below the minimum range (AccDisSt) for a DF weapon is the listed maximum accuracy for that weapon. The chance to hit scales down linearly to 60% of the maximum value at the maximum range (AccDisEn); beyond this range the weapon will refuse to fire. So, a weapon with an accuracy of .7, or 70%, has a 42% chance of hitting at its maximum range. Some weapons (Stellar Converter, Mauler) have their minimum range set to just under their maximum range so as to produce an effect of no dropoff in accuracy. Having an accuracy of greater than one (once accounting for range dropoff) is treated like having an accuracy of 1. (Note: See the visibility section below for an unused mod to accuracy.)

Damage Dropoff: Most DF weapons don't do their maximum damage when they hit their target. There are two elements to this.

First, there's a range-based reduction. At or under a weapon's minimum range for damage purposes (DmgDisSt), this reduction is zero; the Base Damage is equal to the weapon's maximum damage (NearDamg). The Base Damage scales down linearly from there to the minimum damage (FarDamag) over the course of twice the weapon's listed maximum range for damage purposes (DmgDisEn). Thus, the Base Damage for a weapon is its minimum damage at 2*DmgDisEn + DmgDisSt; the default weapons set always has the maximum range for accuracy and damage the same, so the actual minimum Base Damage for a weapon is close to the average of the minimum and maximum damage, shifted towards the maximum damage based on the relationship between the minimum range and the maximum range.

The actual damage done is the Base Damage reduced by a random factor. To determine the random factor, take the smaller of the Base Damage and the damage delta (DmgDelta) stat. Multiply this by a random number from between 0 and 1. Subtract the result from the Base Damage to determine how much damage is actually dealt.

By default, the damage delta stat is equal to the difference between the maximum and minimum damage stats.

Mass-classsed weapons (Mass drivers, Rail guns, and the like), have their minimum range for damage set to be just under the maximum range for damage, so they have no dropoff due to range. Furthermore, they have their damage delta set to 0, so there is no randomized reduction.

Mounts change all of the stats mentioned above except for the damage delta stat. Thus, mounts with larger multiples on the damage stats (e.g. the Spinal series) suffer much smaller proportional reduction in damage dealt from the random factor. (Obviously, mass-classed weapons don't benefit from this effect.)

Note: The code is capable of handling cases with stat configurations that are far from the defaults. DmgDelta can be much larger than the NearDamg; FarDamag can be larger than NearDamg; AccDisEn does not have to be the same as DmgDisEn; etc. These cases all behave as described above.


Weapon Stats: Range and damage weapon stats, as well as DmgDelta, are discussed above. Other weapon stats:[list] FireDlay: Seconds between weapon fires. There seems to be a floor on this stat; reports suggest it is around 0.5 or 1 second.
Multfire: How many times the weapon fires when FireDlay has elapsed. All Multfire shots hit the same target.
ShldPen: When hitting shields, the shield's stopping percentage is multipiled by this value; lower is better. If StpPow * ShldPen is greater than one, it is treated as if it is one.
ArmrPier: Is a multiplier against the chance to hit armor; the lower the ArmrPier, the greater the chance that a particular shot ignores armor and hits Hit Points directly. Mounts: Weapon mounts can change the four range stats, the two damage stats, Dmgdelta, size, cost, ShldPen, MacxAccry, but not MultFire.

Mods: Weapon mods are capable of altering: MultFire, the four range stats, FireDlay, the two damage stats, size, DmgDelta, and ShldPen.



IF Weapons

All default missiles have 95% accuracy at all ranges, regardless of what you change about their range stats; they always do max admage, similarly regardless of what you change about their range stats.

The rate of fire displayed in the ship design screen is not the actual rate of fire, but the multiplicative modifier to the warhead's rate of fire, which are all 45 seconds by default. The number of racks you have is the number of missiles you fire in a volley; the number of missiles per rack is the total number of volleys you can fire. The space a missile rack takes up is proportional to 1 plus the number of missiles in it.

Once you have the tech for missile armor and shields, they are installed by default on all missiles in new designs, but the checkboxes aren't selected. To turn them off, you have to select then deselect the checkboxes.

Missiles have Hit Points equal to the space they take up per missile (including chassis and weapon mod effects).

MultFire for missiles just generates more missiles; it doesn't require all the missiles to hit the same ship (or missile or fighter).

Fighters

As far as I know, the following fighter stats have no in-game effect: ShieldDmg, MaxShrnk, MinFactr, MinimSize, HullDMod, DamagCap.

Fighters have HP equal to the base space, the amount they take up before the *=1.5 multiplier for Interceptors or the *=3.0 multiplier for SCFs is applied (fighter weapon mods do add to HP, though). The only difference between SCFs and Interceptors is that SCFs take up twice as much space and get a 40% increase in max damage and a 20% increase in min damage. Damage calculations for fighter weapons are handled in the same fashion as for direct fire weapons. However, all fighter weapons are listed with 3000 minimum range, and almost all fighter combat happens under that distance.


Carriers wait 90 seconds between launching successive waves of fighters.

I have noticed no difference between SCF AI and Interceptor AI. It's conceivable that Interceptors will occasionally deign to shoot at missiles and fighters while SCFs won't.

It is commonly reported in these forums that Interceptors will not participate in planetary bombardment, but SCFs will.

Modding Note: I've found the following fighter weapon stats to be not modifiable by the fighter chassis: Armor points, DmgDelta, DmgDisSt, MultFire, ShldPen, MaxAccry. Weapon stats I've found the chassis to modify: UniSpace, SystCost, NearDamg, FarDamag, FireDlay.


ECM/ECCM/Sensors/Cloak

Thanks to Bhruic, we finally know what's going on in visual detection! There are several significan bugs, but it's there.

Sensors and ECCM multiplicatively reduce the OffTgtRg stat. ECM and Cloak multiplicatively increase the DefTgtRg stat. (Values in TechTables which increase OffTgtRg or decrease DefTgtRg are ignored.) Cloak also increases the Cloaking stat.

Major Bug: The space combat code ignores the first non-UniSpace non-SystCost tag in the ShipItem and ShpMItem objects. Thus, the default TechTables versions of Sensors/ECM/ECCM/Cloaks have no effect upon visibilty. A modder can trivially fix this by merely repeating their values, however.

Bug: While they're listed as ShpMItem objects, meaning that they can appear multiple times on a hull, you only get the effect of the single best modifier to each of OffTgtRg and DefTgtRg that appears in a given Task Force. A single ship in an Armada with a Sensors mk V is just as effective as an entire recon armada filled to the gills with 10 Sensors mk V each.

Periodically, every TF in a combat attempts to detect every enemy TF in that combat. Apparently, each wave of missiles and fighters counts as a TF in all ways for this. The period is approximately 10 seconds, but apparently with some element of randomness. If any of your TFs see an enemy TF, they all do.

The chance for a TF to see another TF is found by the following process (this isn't the precise programmatic sequence, but it is functionally identical):

EW_Mod = 0.75/(OffTgtRg*DefTgtRg)
If EW_Mod < 0, set it to the range between the TFs.
If EW_Mod > 1, set it to 1 (which means if it was < 0, it's now 1).
D is the distance between TFs.
Range_Mod = (D-8500)/50000
Chance to see the TF is EW_Mod - Range_mod.

In other words, assuming the appropriate bounds on OffTgtRg*DefTgtRg, your chance of seeing another TF with a given TF is:
0.75/(OffTgtRg*DefTgtRg) + 0.17 - D/50000

Written in terms of squares on the combat map rather than weapon range units (and assuming that a square is 3400), and looking at it from a different direction:

If there's no EW (OffTgtRg, DefTgtRg), you have a 0% chance of seeing an enemy TF at a range of 13.5 squares, and hit 100% chance of seeing it at -1 squares (which you of course can't reach). Each square of distance changes your chance of seeing the enemy TF by 6.8%. As 1/(OffTgtRg*DefTgtRg) grows, up to a maximum of 1.33, both of those distances increase, up to a maximum of 3.67 squares. As 1/(OffTgtRg*DefTgtRg) shrinks, those distances decrease.

Effects on Accuracy: For purposes of accuracy, range is multiplied by DefTgtRg. Note that modding TechTables to fix visibility will mean that ECM (which the AI empires don't use) will completely break game balance and will make ships nigh-impossible to hit with DF weapons at high TLs.

Shipbuilding AI

What the autobuild will pick for you, and thus what your AI opponents use.

Armor and Shields
Autobuild always uses the highest-tech class of armor and shields available. The generator size and armor weight used are as follows: Class 0: No armor, no shields. Colony and Outpost ships are Class 0.
Class 1: Minimal armor, minimal shields. The AI will take use your smallest shield generator and uses a buggy algorithm (you have N armor weights available; if you have more than 2N hull sizes available, assign the heaviest available armor weight to everything; if you have fewer than 2N hull sizes available, assign the second-lightest (counting "none" as the lightest) armor weight to everything; if you have 2N hull sizes available, everything gets the second-lightest weight available except the two smallest hull sizes which get the heaviest) to select armor weight. IF, CV, Recon, and Transport are Class 1.
Class 2: Moderate armor, moderate shields. The AI will use the second-smallest shield generator. For armor weight, see below. Point Defense ships are Class 2.
Class 3: Maximum armor, maximum shields. It will use your third-largest shield generator, and will select armor as indicated below. LR, SR, and Planet Destroyer ships are Class 3.For Class 3 ships, armor weight is selected as follows: Ignoring the "no armor" weight and the first two hull sizes (Lancer and Cutter) available, map your weights to hull sizes as evenly as possible, rounding such that the heavier weights get hulls allocated before lighter weights do. The first two hull sizes get the second-lightest armor (using the default game files, they'd get Very Light). So, if you had all size armor weights in addition to "no armor" and you had all the hull sizes except Behemoth and Leviathan, you'd have: Lancers, Cutters, Corvettes get Very Light; Frigates and Destroyers get Light; Light Cruisers get Medium; Cruisers and Battlecruisers get Heavy; Battleships and Dreadnoughts get Very Heavy; Superdreadnoughts and Titans get Ultra Heavy.

Orbitals are considered one hull larger than they are; Class 2 ships are considered one hull smaller than they are.

Note: "Highest Tech" above actually means "furthest down the listings in TechTables." Since the Damper Field is listed after all other shields, it will be chosen preferentially over all other shields. Changing its position in the sequence will change the preference. Similarly, larger shield generator sizes and heavier armor weights are actually "later listed" in the code; changing sequence will change preference. Nothing actually evaluates the stats of the various weights or techs. "No armor" is merely the first-listed armor weight.


ECM/ECCM/Cloak

If a ship is marked Stealth, it'll put in the best cloak/ECCM available; best in this case is the one with the best (largest) DefTgtRg.

LR, SR, and PD ships get one Sensor or ECCM, the one with the smallest OffTgtRg (i.e. the best) that fits within 25% of the space left for PD weapons.

Recon ships seem to get maxed out on last-listed ECCM in TechTables (last-listed object with OffTgtRg). Note that multiple copies of EW items do not increase effectiveness.

Weapon and Mount choices

I've not investigated anything about the decisionmaking process of picking which Planet Destroyer weapon to mount on a Planet Destroyer ship. :) Each of SR, LR, and PD ship-types are listed in TaskForceRules as trying to optimize a particular characteristic (NearDamg, AccDisEn, and FireDlay, respectively). For simplicity, I'm going to talk about the decisions based on SR/LR/PD. I've not found another stat that you can plug in there that will work (including FarDamag, which is listed as what IF TF's are supposed to optimize) for DR autobuild choices.

The AI will add mods to weapons according to the following scheme: Point Defense weapons: No mod that increases space or FireDlay gets used. Other mods that reduce either or both get used.
Short Range Weapons: No mod that decreases NearDamg gets used. Other mods that reduce space or increase NearDamg get used.
Long Range Weapons: No mod that decreases AccDisEn gets used. Other mods that reduce space or increase AccDisEn get used.


The Autobuild AI will consider the effects of the above mod choices when making decisions according to the following rules:

Point Defense ships: For the type of the primary weapon, it picks the weapon with the smallest FireDlay (in general, every weapon except the New Orion weapons has the same FireDlay of 2.5 seconds). From any ties there, it chooses the weapon with the highest base tech level. From any ties there, it picks the weapon lowest on the TechTables list. It uses the available mount highest on the list in TechTables. Once it has stuffed as many of those in as it can, it will try to fill in the remaining space by going down its list of "best PD weapons" and finding the best one that will fit at least one weapon in.

Long-Range ships: For the type of the primary weapon, it picks the weapon with the highest AccDisEn. From any ties there, it looks like it prefers the weapon with the smallest InfoWeap number (the order weapons are listed in the weapons table). For the mount, it selects the mount furthest down the TechTables listing (since TechTables lists the PD, then Light, then Standard, then the Heavy Series, then the Spinal series, it will only pick Heavy series mounts for LR ships if there are no Spinal mounts available). If the selected weapon/mount combo is too big to fit, it will try smaller weapons. 25% of space is reserved for filling with weapons chosen according to the PD criteria above; Sensors consume PD space.

Short-Range ships: For the type of the primary weapon, it picks the weapon with the largest NearDamg. I haven't been able to determine what governs choices between ties in this case. If the selected weapon/mount combo is too big to fit, it will try smaller mounts. It follows the same scheme for weapon mount choice and secondary weapons as it does for LR.

Indirect-Fire ships: It takes the last-listed warhead you've got and puts it on the chassis furthest down TechTables. It then fills in space with PD nukes. I haven't yet tested how it decides to use PD nukes; I assume the PD chassis is selected as the first listed in TechTables, nukes are probably chosen as the first warhead listed in TechTables (and if there were lower-listed warheads with smaller space, they'd presumably be used in any case where the PD nukes wouldn't fit). It builds 5-missile racks.

Carriers: It selects the fighter weapon listed last in TechTables, and fills the carrier with SCFs of this type. It then fills the rest of the space with Laser Interceptors. Predictably, the SCF/interceptor choice is made by position in TechTables, and it's also the case that the weapon used for "padding" is chosen by taking the first figther listed in TechTables. It looks like for the primary fighter, it'll add all available mods and for the "padding" it will ignore mods


Engine Speeds

Colony ships, Transport ships, and Outpost ships get autobuilt with 60% of their max speed. IF ships get autobuilt at 80% of max speed. Everything else gets full system engines.



PD bug


Before the 1.2 patch, there was a severe problem that ships usually failed to acquire the first inbound volley of missiles and/or fighters before they hit/start firing. The best workaround that was found is to make sure that every TF you want to have firing PD has a missile rack with missiles remaining.

The 1.2 patch has largely, but not entirely, alleviated this problem.

Presumably modding Sensors to work will improve the behavior of PD to some degree.

TF AI

This information is based on preliminary testing only. Others have done their own tests and found results that contradict my own, so take the following with a grain of salt:

TF AI is done by TF type, and does not care about the ship designations within the TF. (This assertion in particular is disputed by the tests performed by Blaze/Renaux.)

LR and SR AI: On Watch mode, fighters and missiles are prelaunched (fired before any targets are detected). On Control mode, it auto-launches and auto-fires (fires on targets as soon as detected and in range, without any orders from the player). In Watch mode or when directed to attack a target, it will attempt to maintain a particular target range. If its max AccDisEn is longer than than that of its target, it will attempt to maintain its target's max AccDisEn; if its max AccDisEn is shorter than that of its target, it will attempt to maintain approximately 40% of its max AccDisEn.

IF and CV AI: On Watch mode, it prelaunches fighters and missiles; on Control mode, it autolaunches and autofires at a range that seems based on its DF max range. When in Watch mode or when ordered to attack, it will compare its DF combat stats with the DF stats of the target as to whether it should charge or run away from the target.

Colony and Outpost AI: On Watch mode, it prelaunches fighters and missiles; I expect it autofires in Control mode. In Watch mode, I've observed well-armed TFs maintain a moderate distance between them and their target while firing; I expect that this is triggered by the DF stat comparison leading the AI to believe it was more powerful than the target. It seems likely that similar behavior will be seen in Control mode.

Recon AI: Seems to take the firing behavior of LR and SR and the movement behavior of IF and CV.


The DF stats comparison mentioned above seems to use the same calculations as are used to generate the TF stats that display in the lower right corner during combat, so it is possible to fool the AI into thinking that a TF is better or worse than it actually is. My only hard data point on this is reducing FireDlay to below the apparent effective cap. I currently expect that this is the reason behind the kludges used for Mass-classed weapons, which all have unused FarDamag stats. The same calculations are apparently also used in determining AI retreat behavior and which options the AI defaults to for the pre-combat options ("Blockade Planet", "Intercept Fleet", etc).

[visage]

Data Appendix

<edit - While I updated the first post with the last code patch, I'd forgotten to update this, so for the past year it's been saying that AP doesn't work and similar. Oops.>

Included are the raw stats for various elements discussed in the header post. I'll probably add interpretive tables over time.


Hulls

Hull Num Space Cost TechLvl
------------- ---- ------ ------- -------
Lancer 00 50 50 0
Cutter 01 70 71 0
Corvette 02 100 102 0
Frigate 03 140 149 0
Destroyer 04 200 219 0
Light Cruiser 05 285 325 0
Cruiser 06 405 487 6
Battle Cruiser 07 575 729 11
Battleship 08 815 1132 17
Dreadnought 09 1155 1756 23
SuperDN 10 1635 2736 29
Titan 11 2310 4323 34
Behemoth 12 3265 6853 40
Leviathan 13 4615 11122 45


Armor

Armor Num Points Defl Cost TechLvl
----- ---- ------ ---- ----- -------
Zortrium 01 100 2 10 0
Duranium 02 200 6 30 10
Titanium 04 400 10 70 20
Neutronium 07 800 14 150 30
Adamantium 10 1600 18 310 40


Armor Weights

Armor Weight Armor*= Defl*= Cost*= TechLvl
------------ ------- ------ ------- -------
None 0 0 0 0
Very Light 0.5 0.5 0.75 0
Light 1 0.66 1 0
Medium 2 1 3 11
Heavy 4 1.5 7 21
Very Heavy 8 1.75 12 35
Ultra Heavy 16 2 18 45
Missile/Fighter 0.5 0.25 0 5


Shield Techs

Shield Class StpPow ShdStr Rechrge Cost TechLvl
------------ ------ ------ ------- ---- -------
I 0.5 50 5 50 04
II 0.54 75 8 63 10
III 0.58 125 13 79 15
IV 0.62 200 21 100 20
V 0.66 325 34 126 25
VI 0.7 525 55 159 30
VII 0.74 850 89 200 35
VIII 0.78 1375 144 251 40
IX 0.82 2225 233 317 45
X 0.86 3600 377 400 50
Damper Field 0.75 32000 32000 750* 45

*: takes up 3* space of others



Shield Generator Sizes

Generat. Size ShStr*= Rech*= Cost*= % HullSpace TechLvl
------------- ------- ------ ------ ----------- -------
Small 0.66 0.5 0.7 2.5 0
Standard 1 1 1 5 0
Large 1.5 1.5 2 10 16
Missile/Fight 0.33 1 0 0 16

Note: Damper field takes up 3* as much space as normal



Direct-Fire weapons

Weapon NearDam FarDam DmgDsEn MaxAcc Cost Space ShldPen TechLvl Number Mods
------ ------- ------ ------- ------ ---- ----- ------- ------- ------ -------
Laser 7 1 7167 0.7 5 10 1 1 0 M1, M2, Imp, AP, AF, Cont
Mass Dr 14 3* 5733 0.5 9 22 1 1 9 M1, M2, Imp, AP, AF
Fusion 31 2 6167 0.8 21 15 1 5 1 M1, M2, Imp, Cont, Env
Quark 15 3 8017 0.7 10 12 0.9 5 14 M1, M2, Imp, AP, Cont
Hard Be 15 7 11333 0.7 10 7 1 10 15 M1, M2, Imp, AF, AP, Cont
Neutron 33 3 11628 0.7 22 15 0.9 15 2 M1, M2, Imp, AP, Cont
Rail Gu 30 7* 9067 0.5 20 13 1 10 16 M1, M2, Imp, AF, AP
Gravito 49 3 13433 0.7 33 18 0.9 20 3 M1, M2, Imp, AP, Cont
Hellfir 67 10 8944 0.8 45 19 1 15 17 M1, M2, Imp, Cont, Env
IonPul# 33 4 15500 0.7 22 10 1 20 11 M1, M2, Imp, AF
Parti$# 49 10 13433 0.9 33 18 0.1 20 12
Phasor 48 5 17583 0.7 32 14 1 25 4 M1, M2, Imp, Cont, AF, AP
Plasma 213 6 13111 0.8 142 44 1 30 5 M1, M2, Imp, Cont, Env
Gauss 97 9* 14067 0.5 64 27 1 25 10 M1, M2, Imp, AP, AF
LFG# 2.5 36^ 6364 0.9 24 10 0.7 25 23
DEB 107 20 17044 0.7 70 25 1 35 19 M1, M2, Imp, AP, Cont
Disrupt 209 20* 17400 0.5 139 54 1 35 6 M1, M2, Imp, AP, AF
Death#$ 157 50 18850 0.9 105 45 0.8 35 13
Disint 105 23 21750 0.7 70 25 1 35 19 M1, M2, Imp, AF, AP, Cont
Megabol 314 23 14500 0.8 209 59 1 35 22 M1, M2, Imp, Cont, Env
Tachyon 231 27 20656 0.7 154 61 0.9 40 20 M1, M2, Imp, AP, Cont
DarkMat 453 30* 20733 0.5 302 95 1 45 21 M1, M2, AF, AP
Mauler$ 679 1! 15889 1@ 453 113 0.5 45 7 M1
Stella$ 1000 200! 28000 1@ 666 495 1 50 8

The above is listed in TechTables order
Default stats: FireDlay: 2.5; MultFire: 1; DmgDisSt: 4500; AccDisSt: 4500; AccDisEn = DmgDisEn.
DmgDelta is by default the difference of NearDam and FarDam.
*: DmgDelta is 0, and DmgDisSt is set to DmgDisEn-1, so damage is always NearDamg
!: DmgDelta is 0.
@: AccDisSt is set to AccDisEn-1 so accuracy is always MaxAcc.
$: Mauler's FireDlay is 4; Stellar Converter's FireDlay is 8; Particle and Death are 2.
^: MultFire = 5; FarDamag is equal to NearDmg, so DmgDelta is 0.
#: Innate armor piercing. 0.7 for LFG and Death Ray, 0.8 for Ion Pulse Cannon, 0.9 for Particle Beam.



DF Weapon Mounts

Mount Dam*= Range*= Dlay*= Space*= Cost*= TechLvl
------- ----- ------- ------ ------- ------ -------
Point Defense (*) 0.75 0.75 0.8 1 9
Light 0.8 0.8 0.8 0.8 0.75 0
Standard 1 1 1 1 1 0
Heavy 1.5 1.3 1.2 1.75 2 12
Very Heavy 2 1.7 1.4 2.75 3 28
Ultra Heavy 2.5 2.1 1.6 4 4 39
Spinal 2 1.5 2.2 1.5 3 0
Improved Spin 3 2 3.5 2 5 27
Ultra Spinal 5 2.5 5 3 7 39

This above is listed in TechTables order.
Damage multipliers are to both NearDamg and Fardamag.
Range multipliers are to all four range stats.
*: The Point Defense Mount has a NearDamg stat of 0.5 and a Fardamag stat of 0.33;
it gives a MaxAccry bonus of *= 1.5



DF Weapon Mods

Mod TechJump Effect
---------- --------- -------
Autofire +4 MultFire *= 3, Space *= 2
Continuous +4 MaxAcc *= 1.5, Space *= 1.25
Armor Penetrat +4 ArmrPen *= 0.75, Space *= 1.5
Enveloping +4 ShldPen *= 0.5, Space *= 1.66
Improved +13 NearDamg *= 1.5, FarDamag *=1.1
Miniturize 1 +4M Space *= 0.8
Miniturize 2 +8M Space *= 0.8

TechJump indicates how many tech levels above the weapon you find
the mod. 'M' indicates that the mod is found in the Mathematics
tech group.



Warheads

Warhead Number Damage Cost Space TechLvl
--------- ------ ------ ------ ----- -------
Nuclear 01 56 14 6 0
Anionic Energy 02 82 15 6 7
Neutronium 03 121 17 7 15
Hercular 04 178 19 8 18
Merculite 05 262 21 8 19
H.E. X-Ray 06 386 23 9 26
Scatter Pack 07 568 25 10 31
Ionic Pulsar 08 836 28 11 35
Energy Pulsar 09 1231 31 12 41
Omega 10 1811 34 14 45

All warheads have FireDlay = 45, MaxAcc = 0.95.


Missile Chassis

Chassis Dlay*= Cost*= Damag*= Space*= TechLvl
------- ------ ------ ------- ------- -------
Point Defense 0.1 0.75 0.25 0.25 0
Rocket 0.5 1 1 1 0
Light Missile 1 2 2 1.25 18
Heavy Missile 1.33 3 4 1.75 28
Torpedo 1.66 5 5 2.5 37



Fighter Weapons

Weapon NearDam DmgDelt MaxAcc Cost Space ShldPen TL Mods
--------- ------- ------ ------ ---- ----- ------- ---- ----
Laser 4 3 0.7 4 6 1 1 AF
Mass Driver 7 0 0.5 7 13 1 1 AP
Fusion 15 4 0.8 15 14 1 5 Env
Neutron 17 8 0.7 17 7 1 15
Graviton 25 12 0.7 25 9 1 20
Phasor 24 15 0.7 24 7 1 25 DualPod
Ion Pulse Canno 16 16 0.7 16 5 0.5 20 DualPod
Particle Beam# 25 20 0.9 25 9 0.1 20
Gauss Cannon 48 0 0.5 48 17 1 25
Death Ray#* 78 50 0.9 78 22 0.7 35
Plasma 107 25 0.7 107 41 1 30 DualPod
Disruptor 105 0 0.5 105 32 1 35

The above are listed in TechTables order
Default stats are FireDlay 2.5 and Multire 1.
*: FireDlay of 2.
#: Innate Armor Piercing: 0.8 for Particle Beam and 0.7 for Death Ray.


Fighter Mods

Mod TL Size*= Cost*= Effect
--------- ----- ------- ------- -------
Auto Fire +7 2 1 MultFire *= 3
Armor Penetrat +3 1.5 1 ArmrPier *= 0.75
DualPod +4..7M 1.8 2.25 MultFire *=2


Fighter Chassis

Chassis Size*= Cost*= Effect
--------- ------ ------ ------
Interceptor 1.5 1
Space Control F 3 2.5 NearDamg *= 1.4, FarDamag *= 1.2

The above is listed in TechTables order.



System Engines

Engine Cost Size*= Speed TL
-------------- ------ ------ ------ ---
Thrusters 25 0.25 1500 0
Improved Thrus 27 0.235 1800 5
Hydrogen Fuel 30 0.2325 2100 10
Impulse Engine 32 0.218 2400 15
Iridium Fuel C 35 0.2155 2700 20
Dotomite Cryst 37 0.201 3000 25
Uridium Fuel C 40 0.1995 3300 30
Reajax Fuel Ce 42 0.185 3600 35
Trilithium Cry 50 0.183 3900 40
Transwarp Driv 52 0.169 4200 45


Interstellar Drives

Drive Cost Size*= Speed TL
--------- ----- ------ ------ ---
Retro Engine 100 0.25 85 0
Nuclear Engine 126 0.235 105 5
Sub-Light Drive 158 0.2325 133 10
Fusion Drives 200 0.218 168 15
Impulse Drives 252 0.2155 211 20
Ion Drives 318 0.201 266 25
Anti-Matter Dri 400 0.1995 336 30
Inter-Phased Dr 502 0.185 422 35
Hyper Drives 634 0.183 532 40
Warp Factor X 800 0.169 672 45



Mandatory Ship Systems

System Cost Size*=
--------- ----- -------
Bridge 25 0.02
Crew Quarter 10 0.05
Life Support 40 0.02


Electronic Warfare

System Size Cost OffTgtRg DefTgtRg Cloaking TL
----------- ----- ----- -------- -------- -------- ---
Sensor I 15 30 0.85 0
Sensor II 20 90 0.79 13
Sensor III 35 150 0.63 17
Sensor IV 55 210 0.46 30
Sensor V 90 480 0.25 41
----------- ----- ----- -------- -------- -------- ---
ECM I 15 30 1.12 0
ECM II 20 60 1.17 12
ECM III 35 120 1.33 20
ECM IV 55 240 1.6 28
ECM V 90 480 2.2 37
----------- ----- ----- -------- -------- -------- ---
ECCM I 15 30 0.87 0
ECCM II 20 90 0.81 12
ECCM III 35 150 0.67 20
ECCM IV 55 210 0.5 28
ECCM V 90 270 0.3 37
----------- ----- ----- -------- -------- -------- ---
Cloaking De 15 30 2.2 1.5 18
Phased Cloa 20 30 3.3 1.8 29
Reactive Cl 35 30 4.95 3 38
Ghost Devic 55 39 7.425 7.4 45
----------- ----- ----- -------- -------- -------- ---





Other useful tables I've generated:

Point Defense effectiveness at various TLs and ranges: http://web.mit.edu/munch/www/moo3/newpdnums.html

Same as above, but without LFGs clogging up the numbers: http://web.mit.edu/munch/www/moo3/newpdnolfg.txt

The rest of these tables were done an awfully long time ago and I haven't checked that they're accurate, but I think that they're accurate in terms of what they claim to be:

Damage efficiency of weapon/mount combinations at maximum range: http://web.mit.edu/munch/www/moo3/fardammetric.html

Damage efficiency of weapon/mount combinations at 50% of maximum range: http://web.mit.edu/munch/www/moo3/middammetric.html

Percent damage done at maximum range, by weapon/mount: http://web.mit.edu/munch/www/moo3/fardropoff.html

Percent damage done at 50% maximum range, by weapon/mount: http://web.mit.edu/munch/www/moo3/middropoff.html

[visage]

Interesting Implications

<edit - Note that this was all written pre-code-patch (1.2? something like that) and thus there are things that are now inaccurate. I may rewrite it someday, but don't hold your breath. :) >

The following is conclusions I've drawn from the information above. If you think I've drawn an incorrect conclusion, point it out. I'll be glad to fix mistakes in my reasoning. I'll try to avoid wandering into speculation that isn't driven primarily by the specific mechanics discussed above (for example, I'll try to stick to things like "At range X and tech level Y, the following weapon/mount combos seem to offer the best damage efficiency once shields and deflection are counted in" instead of offering up the l33test TF I can design).

I'll be adding sections over the next week or so, I expect... which means I might be done in time to have it all be obsoleted by the code patch. :)


Point Defense

First off, the fact that weapon mounts can't affect MultFire means that the PD mount is itself pretty worthless right now. Without that MultFire *= 2, it has no benefits over the Light mount.

For the rest of this, though, I'm going to be optimistic and talk as if QS is fixing the mount problem in the code patch (hey, QS, if you're listening: while you're enabling MultFire in mounts, could you make fighter chassis able to mod fighter weapon combat stats, and shield generators able to change StpPow? I'll be explaining why these would be useful in my upcoming post about making the AI better at combat through mods.). I'm also not going to be talking about the PD bug.

How do you kill an incoming missile or fighter? In general, a sequence of shots hit the target. Assuming that after getting past any shield's stopping percentage deflection doesn't stop the shot, each shot chews away at armor or hits internals. The first shot's going to hit armor; after that, each one has a chance of hitting internals equal to the fraction of the armor that's missing from previous hits. Generally, before the shields are depleted and long before the armor is run out, enough hits have gotten past armor to destroy all the internals and the target blows up.

I'm guessing that firing on a swarm of missiles/fighters leads to shots being randomly distributed between the targets. This guess is based on casually observing what happens in that case and when DF volleys fall on TFs. I haven't done the testing to try to distinguish between that case, shots being evenly distributed, and shots falling on single targets in sequence; the behavior I've seen seems to bear the most resemblance to the first, offhand.


The following elements are what's particularly relevant to a weapon/mount combo being good for point defense:

Getting Past Deflection: This is actually a significant concern. Most weapons (generally anything that's not mass-class) only average 35-45% of their maximum PD-mount damage when firing at max range; a PD-mount Ion Pulse Cannon only averages 3.366 damage per hit at max range. About half the shots that hit will do damage less than that. IPCs are TL20. At TL 20, shields have 0.62 StpPow and missile armor has a deflection value of 2 (or 2.5 if fractions are kept, or 3 if it's rounded up; I should test this more). Weapons with no damage dropoff (mass weapons and Lightning Field Generators) generally don't have to worry about this; plasma weapons with their high damage are better about it than beam weapons. In the lategame, when shields become very effective, particle weapons' shield piercing effects are actually fairly useful. Deflection is abig downside to using the PD mount (assuming, of course, that the code patch gets it working), since with most weapons a significant fraction of your shots are going to get stopped by it.

Efficient Packetization of Damage: Up to the threshold where a single shot getting past shields can kill all the internals (ignoring armor) in a single hit, what's important is efficiency in dealing out damage. Past that threshold, efficiency of producing independent shots is more important; the value of of damage output increases goes approximately with the square root. So, a weapon that does twice as much damage per hit as it takes to kill all the internals is only 40% more effective than a weapon that does precisely enough damage to kill all the internals. Now, most weapons do randomized damage, so the value of raising your expected value per hit a fair ways above the amount required to kill the internals in one shot is better than that (since you're increasing the probability that a given hit can kill in a single shot).

Shield Penetration: Since in general you're not going to run out the shields on a point defense target before you kill it, and since getting past deflection is actually an issue with PD mounts, weapons with shield penetration are fairly handy for point defense work.

Range: Long-ranged PD has two uses. Primarily, it allows your TFs to support each other. If your per-TF point defense effectiveness with weapon A is three times that with weapon B, but you've got ten TF's in the battle and they can all support each other with B but not A.... It can also have the effect of letting your PD get multiple volleys in before an incoming wave of missiles hits. Followup volleys have the advantage of firing at shorter range, reducing damage and accuracy dropoff; this means the followup volleys are generally far more effective.


If you go for single-TF-focused PD, nothing beats Lightning Field Generators. They don't have damage dropoff, they've got good shield penetration, their damage packetization is such that they never hit deflection (at least against AI designs) even with a PD mount, and they have MultFire *=5 to get an obscene damage output efficiency. Of course, their short range is such that you're not going to get much inter-TF PD support with LFGs.


It's often the case that your best PD is lower-tech weapons outfitted with useful mods like continuous and autofire, especially since right now you should be using the Light mount and thus aren't in nearly as much danger from deflection. (Once the PD mount is fixed, the benefit of keeping ahead of deflection while using its MultFire bonus might push more for using the highest tech available.)


I haven't done enough testing of missile velocities to determine what ranges and tech levels will lead to X number of volleys. If I do (or anyone else does) that testing, I'll expand this section to cover that.


Armor and Shields

Armor: Since QS halved all armor costs in the data patch, one really doesn't have to do any math on the subject of what kind of armor to put on a ship. Just pile it on. Sure, as the Armor/Internals ratio grows the per-armor-point value of the armor goes down, but unless you're building hulls far far below your technological limit, the cost of armor is a small fraction of the total cost of the ship and thus negligible compared to its benefit. Troop, outpost, and colony ships are often far far below your size limit, and they generally see less combat than the other ship types, so minimal armor on them makese sense. Every combat hull, though, should really max armor.

Shields: If you've got a starship that has the mandatory equipment and moves at top system speed, you've consumed between 59% and 42.8% of its space (depending on Tech Level). So, the difference between consuming another 2.5%, 5%, or 10% for shields is significant. At TL16, when you gain access to the Large shield generator, you're down to 47.4% of your space left before you add a shield or weapon. The large shield generator will consume 20% of your remaining space. Will it increase the ship's survivability by that much? The answer isn't straightforward, but here are some thoughts: Deflection: One element of a shield's usefulness is that its stopping percentage can drop the damage dealt by an attack to below the deflection threshold. When it does that, the hit does no damage to armor or internals. This is most relevant if you're fighting at extreme range, if your opponent is using Light or Standard mounts, and particularly against fighters. Most of the point of the Spinal series is to be able to do damage at long range without dropping below deflection levels.

Combat Intensity: The faster ships are blowing up, the less time the shields have to recharge on something that's taking damage. It takes 10 seconds to recharge a shield from 0 to max, in general (small shield generators take about 13). If you've got two LR TF's sniping at each other at long range, shields are unquestionably worth it; the shields will generally be able to recharge in between salvos. If you've got 10 TFs versus 10 TFs, whatever's taking fire isn't going to get any noticeable recharging in. A shield's strength points alone are generally not going to justify its space, you need to benefit from its recharge rate.

Force/Tech Ratios: If you expect to be steamrollering an enemy, shields will help keep down the attrition you suffer in the process. If your tech is significantly superior to your opponent's, the value of shields goes way up, as you start forcing them beneath the deflection threshold and benefit more from recharging. If you're overwhelmed in tech or force, shields aren't going to do you much good.

System Ships / Orbitals / IF / CV: System ships and Orbitals don't have warp engines. All four of the listed frequently have system speed significantly reduced. Both of these significantly increase the space available, and thus reduce the effective space cost of shields for these types of ships.
Damper Fields: Damper fields merit their own section apart from normal shields. They take up enough space and their shield points and recharge rate are sufficiently close to infinite that you should probably never install anything but a small one (I really hope QS closes this loophole; I'd prefer they did it by making Damper Field the largest shield generator rather than by writing code to ban it from using the small generator; in any case, fixing this is necessary to prevent the AI from shooting itself in the foot by building Large Damper Fields all the time in the late game). Damper fields very clearly do increase the survivability of a ship by an amount that justifies the 7.5% of the hull (approx 13% of the space after essentials). Looking at the example of Ultra Heavy Disintegrators firing at 50% of their max range against a leviathan with Ultra Heavy Adamantium, a damper field increases the survivability of the ship by approximately a factor of 2 over using a Class X shield with a Large Generator (assuming minimal recharging time). Ships destined for low-intensity combat might benefit from using a non-damper shield, but that dramatically reduces their suitability for high-intensity combat.



Direct Fire Weapons

Coming soon.

Missile Choices

Coming soon.

Fighters

Coming soon.

ECCM/Sensors/ECM/Cloak

Coming someday. :p

[Da_Blade]

Sorry visage, wanted to keep it free from replies, but it seems like it's gonna sink to bottom else :(


STICKY! :D

[DoubleSkulls]

Great work.

Sticky!

[Philosophy]

Please Sticky.

And visage, after you clean this up make sure someone gets it into an Encyclopedia mod.

[visage]

Round 1 edits done. I've added tech levels to most of the tables, and fixed a bunch of typos.


Is having the tables actually located in these forums worth having it spread the message width out, or should I just make all the tables located off-site and just link to them?

[Da_Blade]

No, i like them here, i hate having to click offsite links to get the point of a message ;)

Though putting them in a .txt somewhere so we get nice printable tables would be nice :)

[Ron_Lugge]

I have to concur with the sticky...

Though there were a few details that seemed off.

Don't system ships get 1.5X the space and orbitals 2X the space?

[visage]

Originally posted by Ron_Lugge
I have to concur with the sticky...

Though there were a few details that seemed off.

Don't system ships get 1.5X the space and orbitals 2X the space?

The 1.3x and 1.5x is what's in Spreadsheets, and I'm fairly certain I checked that ingame.

[Ron_Lugge]

Originally posted by visage
The 1.3x and 1.5x is what's in Spreadsheets, and I'm fairly certain I checked that ingame.

I'lll go doublecheck...

But I think thats wrong.

[visage]

Originally posted by Ron_Lugge
I'lll go doublecheck...

But I think thats wrong.

Okey-doke. Let me know if it is.

[Ron_Lugge]

Originally posted by visage
Okey-doke. Let me know if it is.

You're right, it is 1.3 and 1.5 (about...)

[visage]

I've discovered one whopper of an error in the original post: fighter/missile armor behaves like ship armor when damage falls on it. I should go looking through the armor/shields thread and figure out where my thinking otherwise comes from...


I think I've fixed the first post to account for this change in understanding. Let me know if I missed anywhere that there's a reference to "fighter/missile armor working differently than ship armor."

[Hachiman Taro]

This is possibly the most STICKY - able thread I've read. Has anyone PM'ed a mod about it?

[Ron_Lugge]

Originally posted by Hachiman Taro
This is possibly the most STICKY - able thread I've read. Has anyone PM'ed a mod about it?

*cough*justdid*cough*

Edit:

*cough*stickied*cough*

[sL|De]

finally visage!!!

and well worth the wait... ill have it on ooMoo by tommorow.. when the site is back up

once again, bravo :up:

[elmo3]

visage

Great work, thanks.

I did just get a page not found for this link though:

Damage efficiency of weapon/mount combinations at 50% of maximum range:
http://web.mit.edu/munch/www/moo3/middammetric.html

Any chance you could make the tables in the posts into linked pages too? I can't cut and paste them into Word. Might just be me but they won't come willingly. :)

elmo3

[visage]

Originally posted by elmo3
visage

Great work, thanks.

I did just get a page not found for this link though:

Damage efficiency of weapon/mount combinations at 50% of maximum range:
http://web.mit.edu/munch/www/moo3/middammetric.html

Any chance you could make the tables in the posts into linked pages too? I can't cut and paste them into Word. Might just be me but they won't come willingly. :)

elmo3

Okey-doke, I've moved "midmetric.html" to "middammetric.html" and stuffed the text of the tables into "tables.txt" in the same location.

<edit: moved tables to tables.txt>

[elmo3]

Great. Thanks!

[visage]

Originally posted by elmo3
Great. Thanks!

Something that just occured to me: The damage efficiency tables (the "metric" ones) are not interoperable. They each ignore accuracy dropoff; this is fine for the Far Damage one, since everyone operates at 60% of their max accuracy at max range (except maulers and stellars).

This is a bug for the mid-range table, since the actual accuracy dropoff at 50% of max range varies from weapon to weapon. I need to fix that... (basically, the shorter the max range is, the more accuracy dropoff you see at 50% of that max range, since you're scaling from min range to max range, and everyone's min range is the same)

[sL|De]

heres your guide... ooMoo style... a little more print friendly as well.

http://lacota.net/moo3outpost/guides/combatmechanics.html

[Morris13]

Would it possible to get some expanded information for the ECM/ECCM section? Half the information I've been able to find on this board is directly conflicting with the other half in terms of wether ECCM and ECM stack between ships or not, or wether they counteract each other on the same ship, etc etc.. That's about the only thing missing from an otherwise stupendously comprehensive guide.

[mrmike49]

Creating a "LR" task force with a number of PD ships, and at least 2 each of Carriers and "IF" results in EXCELLENT Point Defense results !!!

All ships have a few PD weapons also.

What happens when you target a planet/star base/fleet is that the TASK FORCE launches missles & fighters to hit that target,,

LEAVING ALL THE LR, SR, and PD weapons free to respond to incoming missles and/or fighters

Try it!!

[visage]

Originally posted by Morris13
Would it possible to get some expanded information for the ECM/ECCM section? Half the information I've been able to find on this board is directly conflicting with the other half in terms of wether ECCM and ECM stack between ships or not, or wether they counteract each other on the same ship, etc etc.. That's about the only thing missing from an otherwise stupendously comprehensive guide.

I spent a *lot* of time trying to figure out ECM and ECCM. That was what started me on testing combat mechanics. All I ended up with was a bunch of noise for data. It's believable that everything connected to ECM and ECCM will become much clearer with the code patch; I'm certainly hoping it will.

If the code patch comes out and there's no change to ECM/ECCM/etc, then I'll probably bite the bullet and see if I can't extract some real data. I'm certainly not going to complain if someone beats me to it. :)

[visage]

First installment of the "Implications" post is up. I was intending to post it all at once, but I realized that with the rate at which I was writing sections, I probably should just post them as I write them.

[Jazzlvr]

Does anyone remember the strategy guide for MOO 1? This is just the kind of deep technical detail I remember in that very thick tome. I'm not sure I've ever seen a strategy guide that fulfills its purpose better. QSI should have paid someone to write this sort of thing rather than have their fans do it post-publication, and have to worry about mis-interpretation of observed details; QSI could have given an author all the internal details needed to remove all doubt from the accuracy of this kind of superb work.

Thank you visage!

[clb]

It is worth noting that what the player will optimize for should change in the game. Early in the game you have what I will call condition A: Plenty of room in the taskforces, but limited production. For this situation the ships should be optimized for the most bang per buck.
Late in the game you have Condition B: Full Armadas, but plenty of production. Optimising for space (while using the largest hulls) is paramount here.
Note that certain systems and planets can reach a condition B for their system ships (only 18 show up at a time) and orbitals (3 or 4 + 3 per moon?) well before the empire as a whole does for starships. Note that the AI selection of Armor/Shields not only does not distinguish between the situations, but often makes choices contrindicated in *both* situations.

An aside:
Transports deliberately enter combat zones and are reusable, gving armor and shields more worth; after all, a colony or outpost ship is gone forever once used.

[visage]

I've done some more testing of weapon mods and the Autobuild AI's choices wrt to them; top post is updated to include the info. Basically:
PD: No mod that increases space or FireDlay gets used. Other mods that reduce either or both get used.
SR: No mod that decreases NearDamg gets used. Other mods that reduce space or increase NearDamg get used.
LR: No mod that decreases AccDisEn gets used. Other mods that reduce space or increase AccDisEn get used.





Originally posted by clb
It is worth noting that what the player will optimize for should change in the game. Early in the game you have what I will call condition A: Plenty of room in the taskforces, but limited production. For this situation the ships should be optimized for the most bang per buck.
Late in the game you have Condition B: Full Armadas, but plenty of production. Optimising for space (while using the largest hulls) is paramount here.
Note that certain systems and planets can reach a condition B for their system ships (only 18 show up at a time) and orbitals (3 or 4 + 3 per moon?) well before the empire as a whole does for starships. Note that the AI selection of Armor/Shields not only does not distinguish between the situations, but often makes choices contrindicated in *both* situations.

My experience is that there are only really a few things where cost diverges sufficiently from space to really consider building cheap:
a) Armor. But since they halved armor costs, it's really only a big deal if you're building pretty small ships.
b) PD-chassis missiles. You can make an enormously expensive ship by trying to use PD missiles as your primary offensive weapons.
c) The Spinal series. But I never really used those... 'Course, I can shave off 1/3 of the cost of ships the AI autobuilds in various of my savefiles by changing the IS and US mounts to VH and UH.

[Morris13]

There is another situation where you would want to consider building cheap. If you've got a widespread empire with a lot of potential attack points. maintaining fleets of your best and most powerful ships at every possible attack point can become very difficult to support. You can deal with this by using smaller groups of ships, however this can lead to situations where both AI players and human will notice the lower ship counts, and see a weak point. Since there's no way to tell what that armada is actually BUILT with until you're in combat with it, having a couple of armadas of LR destroyers can do a lot more to scare off potential attackers, be they human or AI, than a smaller group of ships with greater actual combat potential.

I have a tendency to use throwaway recon TFs to do probing attacks against enemy fleets to see what kind of firepower they're actually packing, but even in this instance it's hard to tell the difference between 12 Dreadnaughts loaded for bear, and 12 Battleships that are packing less firepower. I really miss the 'scan' option from Moo2 for exactly this reason.

Bottom line is that having large numbers of less powerful ships doesn't make you more EFFECTIVE in combat, but it will make you look a lot more imposing to anyone you're not actually fighting with yet, and that can buy you a lot of time to expand and consolidate.

[Craig P.]

For your efficiency tables, could you elaborate on the formula? I'm doing some damage profile workups, and my damage efficiency (avg dmg * acc / (delay * space)) at max range doesn't agree with yours, even though I presume we're calculating the same thing.

(I'm assuming that the standard mount applies a 1.0 modifier to everything in the weapons table.)

[visage]

Originally posted by Craig P.
For your efficiency tables, could you elaborate on the formula? I'm doing some damage profile workups, and my damage efficiency (avg dmg * acc / (delay * space)) at max range doesn't agree with yours, even though I presume we're calculating the same thing.

(I'm assuming that the standard mount applies a 1.0 modifier to everything in the weapons table.)

So, when you say average damage, are you including the damage dropoff calculations? Average damage at max range is the weighted average of maximum and minimum damages (neardam weighted by Maxrange + minrange, fardam weighted by maxrange-minrange), minus the smaller of 0.5 * DmgDelta and the weighted average.

I threw in a base normalization factor (*20?), and a range normalizer of /10k (applied once per level of exponentiation in the range in the metric).

It's certainly possible there's a bug in my code...

[Craig P.]

I tried to apply both the range dropoff and the random delta (assuming a uniform distribution, so that the average delta is exactly half the max), as well as the accuracy. I'm doing the entire range, but max range is an easy comparison.

Here's the gory details...
SlopeAcc and IntAcc from point-slope form. Same for SlopeDmg and IntDmg. x2 for SlopeDmg is 2*DmgDisEn, otherwise as you'd expect.

Acc = if(R<AccDisSt, MaxAccry, if(R>AccDisEn, 0.0, SlopeAcc*R+IntAcc))
Base Dmg = if(R<DmgDisSt, NearDamg, if(R>2*DmgDisEn, 0.0, SlopeDmg*R+IntDmg))
Base Delta = min(Base Dmg, DmgDelta)
Avg Dmg = 0.5 * (Base Dmg + (Base Dmg - Base Delta))
Avg Dmg (adj) (aka "Punch") = Avg Dmg * Acc * MultFire
Power = Avg Dmg (adj) / FireDlay
Punch Eff = Punch / space
Power Eff = Power / space

I think my Power Eff at max range should be the same as the one in your far damage metric, for one of those tables (I'm not sure which one). However, e.g. for Maulers I get 0.8836, which is a lot less than any of the standard mount numbers you have.

At the moment, I'm plotting damage profiles for individual weapons, but I was planning to work up a spreadsheet that would display a damage profile for the entire battery of a ship.


(Power is intended to represent average damage output per second, which could be very useful in a profile form assuming a ship will survive long enough to get off multiple shots. Punch would be more useful in a max range measure, which is when weapons are likely to open up, and less useful as a profile. Punch still has value because larger punch will take out more opponents at once, potentially reducing attrition, though I was figuring I'd probably design for power all things being equal.)

[Craig P.]

I've finished my first spreadsheet, which has power and punch profiles for every weapon, plus a couple of charts with compared power profiles, plus a chart with compared power efficiency profiles (not inluding the LFG because it's so much better than everything else).

This is, of course, pending the outcome of my discussion with visage to confirm that I haven't screwed anything up.

Anyway, PM me if you want me to send a copy. (directed to the general public)

Addendum: I've now finished the first pass of my primary project, which will graph power vs. range for the beam compliment of a ship. It's an Excel sheet. It's also kinda slow, so I'll see about farming the guts out to something that will generate quicker. Again, PM me if you'd like a copy. Note that the guts use VBA, so you'll need to use caution (or not get it at all) if you're worried about security.

[visage]

So, looking over what you've listed, it looks to me like you're relatively close to what I've got, minus the normalization factor.

For standard mount mauler at max range, before the * 20 normalization factor, I get 0.8055 (I still need to check that I've actually got a * 20 normalization factor in there). You got 0.8836.

Now, it looks to me like you missed one thing, that's how DmgDisSt figures in to damage dropoff calculations, but that should make your result less than mine, not more....

Base Damage drops off from NearDamg at DmgDisSt to FarDamag at 2 * DmgDisEn + DmgDisSt. Your SlopeDmg might be taking this into account, but it looks from what you're saying that it's not.

I'll probably look over my code tomorrow looking for bugs.

[visage]

Oh, and another factor I expect will cause some divergence in our numbers, but also in the opposite direction of what seems to exist: the code doesn't apply the dropoff for accuracy; instead it gives weapons with no accuracy dropoff a *5/3 bonus.

So, properly speaking, my normalization factor is 100/3. I'll update the tables tomorrow after a perusal for bugs, adding the normalization factors to the listing of the metric.

[Craig P.]

I'm using 2*DmgDisEn as one of the points for the damage slope. Should I be adding DmgDisSt to that?

Addendum: For Laser, at a range of 7,000 I get a base damage (before delta and accuracy) of just under 5.5 (about 5.47).

[visage]

Originally posted by Craig P.
I'm using 2*DmgDisEn as one of the points for the damage slope. Should I be adding DmgDisSt to that?

Addendum: For Laser, at a range of 7,000 I get a base damage (before delta and accuracy) of just under 5.5 (about 5.47).

I get 5.94.

NearDamg gets weighted by (2*7167 + 4500 - 7000)/14334. FarDamag gets weighted by (7000-4500)/14334.

Does your slope equation gives you 7 damage at 4500 and 1 damage at 4500 + 2 * 7

[Craig P.]

Looks like I need to revise the equations to make x2 be (RngDmgSt + 2*RngDmgEn) instead of 2*RngDmgEn as I use now. That'll lead to shallower damage slopes all around.

[ross154]

All greetings! I'm old fan of MOO series.
And, yesterday, I have got a third!:up:
I was struck with its huge amount of opportunities!
At me one question: whether is during fight an opportunity to do(make) a pause to estimate a developing situation, and then already to give orders to the ships?

[visage]

Minor updates to the header post; fighters get hp bonuses from mods that increase their size is the big one. There are other small ones probably scattered.


I also regenerated the middammetric and fardammetric files. They now list the normalization factor, actually accounts for accuracy dropoff, and fixes a bug in how I was calculating weights. Now that I think about it, though, the 50% range one is going to be inaccurate in two ways: maulers and stellar converters are going to be boosted a bit because I listed their accuracy as 1.67 in the data so I didn't have to deal with flagging them and didn't cutoff accuracies at 100%; and the accuracy used in that table is a bit lower than it should be, since I fed it 50% of total range for accuracy dropoff calculation, not accounting for the min range effect. I'll fix these at some point....

[Jazzlvr]

Originally posted by ross154
All greetings! I'm old fan of MOO series.
And, yesterday, I have got a third!:up:
I was struck with its huge amount of opportunities!
At me one question: whether is during fight an opportunity to do(make) a pause to estimate a developing situation, and then already to give orders to the ships?

There's no pause I've been able to find; it would be useful during the first few combats when you're not sure what you're doing, but overall you've got so few combat options I'm not sure it'd be generally useful.

[Craig P.]

Due to something raised in the heavy vs. spinal thread, I decided to evaluate PD mounts. I found that even if the mount got its intended multifire bonus, the damage output was still inferior to light mount - it went from atrocious to merely poor. Since indications are that multifire will still only hit a single target, the only advantage PD appears to get is that it has a very slight ROF bonus (0.75 fire delay vs. 0.80 for light mount).

The disadvantages of light mount are:
Slightly less range (about 6%)
Much less damage (by 59%)
In the same amount of spaceIn light of all of this, my conclusion is that the point defense mount should never be used, even if the multifire bug gets fixed -- I don't think it's even likely to increase the saturation level, because I don't think it's likely to make sense to mount something heavy enough to get one-shot kills.

Can anyone offer evidence to the contrary, or point out a flaw in my analysis?

(The fix to this would be decreasing the firedelay so that it's substantially less than that for light mount, or decreasing the damage penalty.)

[visage]

Originally posted by Craig P.
Due to something raised in the heavy vs. spinal thread, I decided to evaluate PD mounts. I found that even if the mount got its intended multifire bonus, the damage output was still inferior to light mount - it went from atrocious to merely poor. Since indications are that multifire will still only hit a single target, the only advantage PD appears to get is that it has a very slight ROF bonus (0.75 fire delay vs. 0.80 for light mount).

The disadvantages of light mount are:
Slightly less range (about 6%)
Much less damage (by 59%)
In the same amount of spaceIn light of all of this, my conclusion is that the point defense mount should never be used, even if the multifire bug gets fixed -- I don't think it's even likely to increase the saturation level, because I don't think it's likely to make sense to mount something heavy enough to get one-shot kills.

Can anyone offer evidence to the contrary, or point out a flaw in my analysis?


Assuming that the PD mount gets fixed and gets its multfire bonus:

Over in the Crisis of Difficulty thread I ran some number-crunching that showed cases where PD were better than Light, I believe.

In any case, the section in the third post about PD explains why PD can be better than light: if your shots kill the internals in one hit regardless of whether you use light or PD, you're better off using PD in general.

[Craig P.]

Originally posted by visage
Assuming that the PD mount gets fixed and gets its multfire bonus:

Over in the Crisis of Difficulty thread I ran some number-crunching that showed cases where PD were better than Light, I believe.

In any case, the section in the third post about PD explains why PD can be better than light: if your shots kill the internals in one hit regardless of whether you use light or PD, you're better off using PD in general.I remembered reading the Crisis of Difficulty thread, and going back to it, it roughly confirms my position -- the only time the 6% higher fire rate of PD helps is when you're approaching one-shot kills.

I probably need to look at this some more to try to figure out just when that's likely to happen (and what the exact impact of the multifire bonus would be -- I suspect that armor and shields make it more complicated than a mere doubling of damage output). My feeling is that it generally won't happen in the early game. I'm not sure about whether the AI's preference for heavy stuff will outstrip rising damage numbers in the mid and late game.

[visage]

Originally posted by Craig P.
I remembered reading the Crisis of Difficulty thread, and going back to it, it roughly confirms my position -- the only time the 6% higher fire rate of PD helps is when you're approaching one-shot kills.

Oh, there's no question that PD mounts as they currently brokenly stand are worthless. :)

and what the exact impact of the multifire bonus would be -- I suspect that armor and shields make it more complicated than a mere doubling of damage output)

For the non-pd case, multfire really is a multiplier on damage output; it's multiplying damage numbers that's the complicated operation.

For PD, it's still pretty close, with the only change being that overkill becomes more likely.

[Craig P.]

Fun with spreadsheets... I worked up a spreadsheet that I hope accurately calculates probabilities for the possible outcomes of one, two, or three shots (most common multifire choices) against an arbitrary missile setup.

Test case - ca. TL 15

Target - Neutronium warhead, rocket chassis, duranium armor, class III shield

Hellfire PD mount - 2.3% chance of a two-shot kill (working PD mount or second shot from non-working mount)
Hellfire light mount - 0% chance of a one-shot kill (10.7% chance of a second-shot kill)
Rail gun PD mount - 0% chance of a two-shot kill (0.1% chance on the third shot)
Rail gun light mount - 0% chance of a one-shot kill (7% chance of a second-shot kill)
Rail gun light mount AF - 13% chance of a three-shot kill (7% chance of wasting the third AF shot)

Conclusion - my best option would be to use light mount AF rail guns, regardless of whether the PD mount is fixed, because I've got a fighting chance at getting a kill on the first volley, and I'm not convinced missiles linger in the engagement envelope long enough to get another chance.

Also, when targetting missiles, unless the AI can get a second shot, it should wait until the missile is within the near accuracy range, because it can't afford to give up either damage or accuracy, both of which will degrade kill chances. I wouldn't guess that the combat AI is actually that smart, but you never know (maybe the PD bug was because of a broken implementation of something like this).

For a ca. TL 30 target, AF light mount rail guns still have a 4% chance at a three-shot kill. LFG looks like it's probably a no-brainer, between five shots of non-degrading base damage of 36 and modest shield penetration and very low space requirements. I'll still check out a couple of alternatives.

Target - HE X-Ray warhead, light missile chassis, Neutronium armor, Class VI shield

Gauss cannon:
PD mount has a 2.8% chance of a two-shot kill
Light mount as a 7% chance of a two-shot kill and a 12.6% chance of a three-shot kill.

Observation - sticking with more, older technology doesn't work out. Slightly more than twice the space buys you three times the effectiveness against this target. Still, if you've got 20 space to burn, the rail gun still has a chance at a kill.


Anyway, before I ramble too long, the upshot is that even at mid- to late-game tech levels, autofire light mounts look to me to be the way to go for PD weaponry.

And anybody, feel free to PM me if you want a copy of my spreadsheet.

[renaux]

now try that against PD missile chasis, and see just how ineffective most PD setups can be. I wrote a guide on missile boats, and some refinement to a missile only strat here (http://www.moo3.tk)

[Craig P.]

I thought about it some more, and I really need to run this out to the first six shots, so that I can check AF PD mounts (with working PD multifire) against AF light mounts. It'd also let me throw LFGs into the mix. I don't think it changes anything for my TL 15 setup, but by TL 30 the PD weaponry might be stout enough for six multifire shots at lower damage to beat three at higher damage.

[Craig P.]

Yikes, I just realize I made a major error figuring probabilities. Combining two uniform distributions most certainly does not yield a uniform distribution. However, I think this only causes a problem when there's partial overage on damage, as when there's no overage the average HP or armor left is the same. I think, though, that the math could get hairy enough that a Monte Carlo would make more sense -- I'll have to explore what Excel has available.

[DrPraetorious]

<I>I have noticed no difference between SCF AI and Interceptor AI.
It's conceivable that Interceptors will occasionally deign to shoot at missiles and fighters while SCFs won't.

All of this was done using Reconnaissance task forces, which I prefer (because they behave this way.) Indirect Fire and Carrier task forces behave differently, but I do not know how differently. Also, all of this was done with my controlling combat; the AI may behave very differently if you watch.

1) If a task force has both missiles and fighters, they always launch at the same time (the first volley, anyway,) and at the same thing.

2) The AI will SOMETIMES launch missiles/fighters before detecting any enemies, but usually not (and I do not know why.) If there aren't actually any enemies (because they were blown to shreds in a previous combat,) you can start combat anyway, and your ships will ALWAYS launch at nothing.

3) Even if you control combat, reconnaissance ships will launch missiles and fighters on their own, as soon as targets are detected (and sometimes before). Personally, I like this feature, which is why I always designate my missile boats as recon.

4) At least if they've been launched from recon ships, the size of the fighters/missiles makes no difference in terms of how targets are selected, or how new targets are acquired after the volley is launched. At first I thought it did after the data patch, but actually I was fooled by the long recharge time of non-point defense missiles.

5) I haven't been totally systematic about #4, but I can say with confidence that:
* If the TF (Recon, IF or Carrier, if other TF types behave differently, I couldn't say) identifies something as it's target, it will launch everything it has at it and start driving towards it if has out-of-range DF weapons of any kind (I like mixed armament vessels.) If the missiles recharge before reaching the target (typically a planet) the TF will launch the next volley. The AI has some capacity to change targets if attacked while the first target is still alive, but I've seen several instances of ships with point defense weapons flying through space, ignoring the fighters attacking them, in order to go shoot lightning fields at the enemy planet; this will presumably change when they fix the point defense bug.
* The mission of individual ships in the TF seems to make no difference. IF ships in a recon task force fire their weapons without prompting. Putting Point Defense ships in a task force doesn't get around the TFs inability to designate enemy missiles as targets.
* Interceptors will attack planets, Space Control fighters will attack missiles and other fighters, Missiles will attack anything, regardless of mount.
* The computer doesn't like to designate fighters or missiles as targets, it would rather fire at the enemy ships. It also has trouble changing targets - this makes point defense weapons on IF/Carrier TFs difficult to use. As far as I can tell, the point defense bug is all about target acquisition. The TF can't fire at the incoming missiles because it can't figure out it should designate them as a target. Various things (putting PD missiles on the ships, setting them to Patrol, etc.) have been proposed but I haven't had 100% success with any of them - the best way to get your ships to fire at enemy missiles is to blow up the entire enemy fleet - this always works.

[renaux]

Originally posted by DrPraetorious
[I]<snip>
* The computer doesn't like to designate fighters or missiles as targets, it would rather fire at the enemy ships. It also has trouble changing targets - this makes point defense weapons on IF/Carrier TFs difficult to use. As far as I can tell, the point defense bug is all about target acquisition. The TF can't fire at the incoming missiles because it can't figure out it should designate them as a target. Various things (putting PD missiles on the ships, setting them to Patrol, etc.) have been proposed but I haven't had 100% success with any of them - the best way to get your ships to fire at enemy missiles is to blow up the entire enemy fleet - this always works.

Actually, on LR TF's, composed of LR missile boats, they will target any missiles/fighters, if they target the attacking enemy TF, as long as your TF launches after the enemy launches. This gives missile boats an extremely high defensiv value, in addition to it's very high offensive value. I believe beam weapons work this way as well.

[DrPraetorious]

The retargeting AI is VERY simple. If you have missiles or fighters enroute to a target, and the target is destroyed, the missiles or fighters ALWAYS retarget on the NEAREST ENEMY THING, with no preferences for ships or planetary defenses over missiles or fighters.

So, if you launch four point defense volleys at the planet, and the first one destroys the planet while the other three are enroute (this is fairly easy to arrange), the others will immediately target the nearest thing - which is often a volley of missiles enroute towards your ships, if you flee screaming at full speed as soon as the combat starts.

Backward Utilised Tracking Torpedoes! Spathi, away!

Thanks for the tip about the LR AI, I hadn't known that.

[Zed-F]

Visage, I know that you tried a high-damage AF mount against a TF of 3 ships and saw only one ship destroyed per shot. Did you try the same against incoming volleys of missiles and fighters? If, say, you have a ship with 1 autofire mass driver up against a trio of laser interceptors (each interceptor taking one hit to kill), do you ever see more than one fighter destroyed in the same volley? What about trios of incoming PD nukes?

[visage]

Originally posted by Zed-F
Visage, I know that you tried a high-damage AF mount against a TF of 3 ships and saw only one ship destroyed per shot. Did you try the same against incoming volleys of missiles and fighters? If, say, you have a ship with 1 autofire mass driver up against a trio of laser interceptors (each interceptor taking one hit to kill), do you ever see more than one fighter destroyed in the same volley? What about trios of incoming PD nukes?

My original tests on the subject of multfire and targets was using a single LFG against volleys of incoming PD nukes. It would always kill only one. I've not tested fighters yet.

[DrPraetorious]

Missiles do help your ships designate missiles as targets, IF you don't give the task force another target, first - this is why PD missiles don't enable me to get around the point defense bug, because I use recon task forces which pick targets without prompting.

Other TF AIs (including LR, IF and TPort) will launch missiles at the enemy ships if explicitly ordered to do so, but you better be quick - otherwise they will autoacquire the enemy missiles as targets.

Again, the mission of individual ships makes no difference, only the mission of the entire Task Force.

So, Recon TF with missiles:
1) Automically targets detected enemies.
2) If the first thing it detects are enemy missiles, it will launch everything at the missiles. Otherwise (as is more likely,) it will launch everything at the enemy ships/planet.
3) Once a target has been acquired, it will pursue that target unless given a new order, or (?) if ships nearby suffer hull damage (?) or maybe blow up (?) or something.

Other TF with missiles:
1) Waits to be given order to target. When watching, sometimes targets enemies and sometimes not (don't know why it makes decision, mount of missile may enter into it.)
2) If it sees missiles and hasn't yet targeted an enemy, it will launch its own missiles defensively. This can be very annoying, if this isn't what you want.
3) Likewise, doesn't want to change targets until the first enemy it targeted has been destroyed.

I assume that Fighters work exactly the same way.

As a seperate problem, ships with no missiles (they need not be point defense missiles, rockets will also do) have trouble designating missiles or fighters as a target, even if they cannot detect anything else. Also, you only have to put one missile boat in the task force, and the entire point defense armament of the task force is enabled. You still have a problem switching targets.

Ships in *TF* with no missiles:
1) Sit there like idiots while getting blown to bits.

So, there are actually two, seperate, point defense related bugs:
1) Once you've started shooting at an enemy, it is very difficult to acquire a new target and start shooting at something else, even with weapons that cannot reach the original target. Giving ships the patrol order seems to solve this problem.
2) If you have no missiles, you can't acquire missiles and fighters as targets until after some part of your fleet has been hit (not sure that is what wakes TFs up, but it seems to be.) A ship fitted with missiles doesn't seem to have this problem, even if it has no missiles left to fire, works with either PD missiles or rockets.

Now I know where psychologists come from. Someone else was already doing the work with the hard numbers so they were forced into the vague, fluffy crap.

[dkass]

I was fooling around and I think the armor selection process used by the auto-build is more complex than you describe. In my current game, I did the following simple experiment. I auto-desgined a PD and a SR ship of each available size (LN through DN). I have access to all armor weights from none to heavy and access to titanium armor (and all weaker types).
I get the following types of armor on the ships (from smallest to largest):

PD: 5 x VL, 2 x L (CL and CA), 2 x M and H (for DN).
All ships had standard type-IV shields
SR: 4 x VL, 2 x L (DD and CL), 2 x M and 2 x H.
All ships had large type-IV shields (large being my best available mount, I also had small and standard).

I repeated the study in a saved game (90 turns earlier) when I only had zortium armor and was also missing the heavy armor (as well as the BB and DN class ships):

PD: 4 x VL, 2 x L (DD and CL), 2 x M
SR: 3 x VL, 2 x L (FF and DD), 3 x M

Its obvious that the ship size and best armor type, in addition to the ship mission, determine the type of armor used. But I'm not sure what is actually happening. It could be a simple, pre-programmed pattern, but it could also be something more sophisticated (say based on the value of the ship and a maximum % devoted to armor).

Interestingly, I also found that there seems to be a fractional space devoted to detection gear for both types. At some size (I forgot to note it down, but CA or BC I think), a detection rig would be added. For the smaller ships, it was my weakest one, but larger ships would get better ones. Note that PD and SR both got them (I forgot to check if there were differences in type or ship sizes).

[visage]

Originally posted by dkass
I was fooling around and I think the armor selection process used by the auto-build is more complex than you describe. In my current game, I did the following simple experiment. I auto-desgined a PD and a SR ship of each available size (LN through DN). I have access to all armor weights from none to heavy and access to titanium armor (and all weaker types).
I get the following types of armor on the ships (from smallest to largest):

PD: 5 x VL, 2 x L (CL and CA), 2 x M and H (for DN).
All ships had standard type-IV shields
SR: 4 x VL, 2 x L (DD and CL), 2 x M and 2 x H.
All ships had large type-IV shields (large being my best available mount, I also had small and standard).

I repeated the study in a saved game (90 turns earlier) when I only had zortium armor and was also missing the heavy armor (as well as the BB and DN class ships):

PD: 4 x VL, 2 x L (DD and CL), 2 x M
SR: 3 x VL, 2 x L (FF and DD), 3 x M

Its obvious that the ship size and best armor type, in addition to the ship mission, determine the type of armor used. But I'm not sure what is actually happening. It could be a simple, pre-programmed pattern, but it could also be something more sophisticated (say based on the value of the ship and a maximum % devoted to armor).

Interestingly, I also found that there seems to be a fractional space devoted to detection gear for both types. At some size (I forgot to note it down, but CA or BC I think), a detection rig would be added. For the smaller ships, it was my weakest one, but larger ships would get better ones. Note that PD and SR both got them (I forgot to check if there were differences in type or ship sizes).

I noticed the armor thing the other day... I posted about it somewhere; guess it wasn't here. :)

Thanks for the info... I'll have to see what I can make happen by modding values and see what it changes about autobuilds...

[renaux]

Originally posted by DrPraetorious
Missiles do help your ships designate missiles as targets, IF you don't give the task force another target, first - this is why PD missiles don't enable me to get around the point defense bug, because I use recon task forces which pick targets without prompting.

Other TF AIs (including LR, IF and TPort) will launch missiles at the enemy ships if explicitly ordered to do so, but you better be quick - otherwise they will autoacquire the enemy missiles as targets.

Again, the mission of individual ships makes no difference, only the mission of the entire Task Force.

So, Recon TF with missiles:
1) Automically targets detected enemies.
2) If the first thing it detects are enemy missiles, it will launch everything at the missiles. Otherwise (as is more likely,) it will launch everything at the enemy ships/planet.
3) Once a target has been acquired, it will pursue that target unless given a new order, or (?) if ships nearby suffer hull damage (?) or maybe blow up (?) or something.

Other TF with missiles:
1) Waits to be given order to target. When watching, sometimes targets enemies and sometimes not (don't know why it makes decision, mount of missile may enter into it.)
2) If it sees missiles and hasn't yet targeted an enemy, it will launch its own missiles defensively. This can be very annoying, if this isn't what you want.
3) Likewise, doesn't want to change targets until the first enemy it targeted has been destroyed.

I assume that Fighters work exactly the same way.

As a seperate problem, ships with no missiles (they need not be point defense missiles, rockets will also do) have trouble designating missiles or fighters as a target, even if they cannot detect anything else. Also, you only have to put one missile boat in the task force, and the entire point defense armament of the task force is enabled. You still have a problem switching targets.

Ships in *TF* with no missiles:
1) Sit there like idiots while getting blown to bits.

So, there are actually two, seperate, point defense related bugs:
1) Once you've started shooting at an enemy, it is very difficult to acquire a new target and start shooting at something else, even with weapons that cannot reach the original target. Giving ships the patrol order seems to solve this problem.
2) If you have no missiles, you can't acquire missiles and fighters as targets until after some part of your fleet has been hit (not sure that is what wakes TFs up, but it seems to be.) A ship fitted with missiles doesn't seem to have this problem, even if it has no missiles left to fire, works with either PD missiles or rockets.

Now I know where psychologists come from. Someone else was already doing the work with the hard numbers so they were forced into the vague, fluffy crap.

I disagree w/ a lot of what you say. Here's how LR works, If an enemy TF is detected, it will target that, if that enemy TF than launches fighters/missiles, your LR TF will then target the incomming threats, and ignore the enemy TF until the incomming threat is dealt w/. i have ships in my TF's that have missile, and ships that doen't, the act in much the same way, as well as those ships w/ carriers. If missiles/fighters are the first thing to be detected, than that is what gets targeted first. I am refering to controled battles exclusively, and the actions of my TF's is automatic.

[renaux]

Originally posted by DrPraetorious
The retargeting AI is VERY simple. If you have missiles or fighters enroute to a target, and the target is destroyed, the missiles or fighters ALWAYS retarget on the NEAREST ENEMY THING, with no preferences for ships or planetary defenses over missiles or fighters.

So, if you launch four point defense volleys at the planet, and the first one destroys the planet while the other three are enroute (this is fairly easy to arrange), the others will immediately target the nearest thing - which is often a volley of missiles enroute towards your ships, if you flee screaming at full speed as soon as the combat starts.

Backward Utilised Tracking Torpedoes! Spathi, away!

Thanks for the tip about the LR AI, I hadn't known that.

I have to disagree w/ this aswell, many times a volley of missiles I've sent out will be a fraction of a second late in destroying a carrier TF, and will turn around and target any fighters in route to my TF's, even though they are much further than other targets. I believe that a carrier TF and it's fighters are for the most part deemed the same target, w/ prefference to the fighters at almost all times. It is fun to watch those fighters get run down by a volley of missiles, and see there number decrease rapidly as they move closer.

[DrPraetorious]

Originally posted by renaux
I have to disagree w/ this aswell, many times a volley of missiles I've sent out will be a fraction of a second late in destroying a carrier TF, and will turn around and target any fighters in route to my TF's, even though they are much further than other targets. I believe that a carrier TF and it's fighters are for the most part deemed the same target, w/ prefference to the fighters at almost all times. It is fun to watch those fighters get run down by a volley of missiles, and see there number decrease rapidly as they move closer.

Well, okay, how do you reconcile that with these results?

The particular experiments I did:

1) Recon Task Force with Missiles Attacking Planetary Installation
The recon task force launches missiles and fighters at the planet. There are multiple volleys of missiles (point defense.)

The first volley bakes the planet.

The subsequent volleys retarget the closest thing, every time. Depending on how quickly the planet detected my own task force (and on the relative distances) this can either be the return volley of the missile base, the fighters from the fighter base, or it can be an orbital defense installation. Distance was the sole determining factor every time - in several instances my ships would get off two more volleys before the first connected and they would often retarget independently (so volley 1 hits the planet, volley 2 retargets at an orbital, and volley 3 retargets the planetary missile volley.)

So, if this works differently if the target is an enemy fleet, I find that very surprising.

If you just watch and do nothing (but in control combat mode) your fleet will ignore the first incoming enemy missile volley vis a vis the direct fire weapons the ships also have.

Once the ships have been hit once (by fighters or by missiles) they wake up and start using their DF weapons again. Multiple TFs can interact in some funny way in this respect but I haven't done thorough testing.

Setting the ships to "Patrol" also wakes them up.

2) Other Task Force types with Missiles Attacking Planetary Installations.
These work the same way, except that they will launch their missile volleys defensively (at enemy missiles) if the planet detects me and fires before I have a chance to give the order to fire at the planet.

As long as the incoming missiles are the designated target, my task force will shoot DF weapons at them. If I order the TF to attack the enemy planet, subsequent volleys target the planet and it fails to shoot DF weapons at incoming missiles, until it's been hit once, or I give it a new order ("Patrol").

3) Repeat with Rockets instead of PD missiles
I can't verify that the retargeting is the same, because I can't get multiple volleys off, but otherwise, the same.

3) Task Forces without missiles but with DF weapons (Recon, TPort or LR)
Ignore the first incoming volley of missiles or fighters. Once they've been hit once, they wake up and start firing back at incoming ships. Setting them to patrol doesn't help.

Once I give the attack order, they start flying towards the enemy planet. However, unlike a TF with missiles, they don't "forget" to shoot at missiles, and remain able to do so until they get within range of the planet (whereupon they shoot at the planet preferentially.)

...
I repeated these tests while attacking a SINGLE enemy IF TF.

The results were consistent, except that I cannot verify how retargeting works.

The recon ships shot at the ship and ignored it's tiny complement of missiles, unless I set them to patrol before the missiles arrived.

The IF ships launch their entire massive payload at the incoming missiles (not fighters, though; so I geuss that fighters and missiles don't always shoot at the same time.) If I deliberately only give the IF TF one missile to shoot, I find that after that missile has been launched the DF weapons work properly.

The ships without missiles sit there and allow the first tiny volley to hit them, then start shooting at the subsequent volleys. No orders I give change this behavior - telling them to Attack the enemy ship makes them move towards it, but does NOT cause them to ignore subsequent volleys.

[renaux]

it's possible that the fighter squadron from the planet was momentarily out of sensor range. When they entered your TF sensor envelope, they might have be reassigned a targeting priority lower than before they left sensor range. This is just theory since i don't have any knowledge of the battle you are refering too. I usually have atleast one fighter in my TF's so my sensor ranges are fairly comprehnsive.

I do find it odd though, you refer to being able to control your first target lock, in LR TF's the targets are set before i can react (<1 sec)

[DrPraetorious]

The LR task force in question doesn't have any actually long range weapons, just lightning fields. I'll try again with longer range weapons and see if it behaves like a recon TF.

[renaux]

if you want, you can check out my guide on space battles, this might give you some insight to where I'm comming from. I have played around w/ a few other strats, but this seems the most powerful www.moo3.tk

[visage]

Ok, I've generated a bunch of tables... these all have the same basic idea: monte-carlo to find out how many hits it takes to kill something, then from that determine the efficiency of the weapon. Note that this means that any overkill effects from multfire are not noticed....

First off, we have shipkilling with DF weapons ranked by space efficiency:
http://web.mit.edu/munch/www/moo3/spaceefficiency.txt

The same, but by AU efficiency:
http://web.mit.edu/munch/www/moo3/auefficiency.txt

Those two tables exclude LFGs, so here're the numbers for LFGs:
http://web.mit.edu/munch/www/moo3/lfg.txt


Now, pd against missiles, space efficiency:
http://web.mit.edu/munch/www/moo3/pdspaceeff.txt

That excluded LFGs, so the TL25+ tables regenerated but with LFGs:
http://web.mit.edu/munch/www/moo3/pdwithlfg.txt

...and the effectiveness of PD against pd-chassis missiles:
http://web.mit.edu/munch/www/moo3/pdpd.txt


Interestingly, the spinal series does very well for shipkilling at most ranges when looking at space efficiency. The heavy series rarely makes much of an appearance on those charts. Switch over to AU efficiency and you see a lot less of the spinals.

Not surprisingly, enveloping doesn't show up until damper fields come into play. I was somewhat surprised to see that enveloping didn't show up more in the PD tables...

PD-chassis missiles are significantly harder to kill per unit space. They're easier to kill per unit AU, of course.

[Vorlock]

This stuff is great, just about exactly what i was originally trying to figure out. I personally never care about AU efficiency, only space efficiency. Good work!

Ok, now what exactly is a monte-carlo ?

[DrPraetorious]

Originally posted by Vorlock
This stuff is great, just about exactly what i was originally trying to figure out. I personally never care about AU efficiency, only space efficiency. Good work!

Ok, now what exactly is a monte-carlo ?

Along with Gibbs' Sampling, it is one of the workhorse techniques of empirical statistics. Basically, you feed a calculation a whole bunch of random values and see what happens.

More info here:
http://www.decisioneering.com/monte-carlo-simulation.html

I thought I had too much spare time!

*Bows to visage*
I'm not worthy! I'm not worthy!

[visage]

Originally posted by Vorlock
This stuff is great, just about exactly what i was originally trying to figure out. I personally never care about AU efficiency, only space efficiency. Good work!

I certainly prioritize space over cost for weapons. Cost does sometimes become prohibitive, though. Before the data patch, IF ships that used only PD chassis missiles generally cost 2+ times more than an IF ship that used the latest tech missiles. If you take an autobuilt ship and rip out the spinal-series mounts and replace them with your most advanced heavy mount, you can often shave 33% off the cost of the ship.

Ok, now what exactly is a monte-carlo ?

Basically monte carloing something is using a large number of "generate and test" cycles rather than trying to explicitly solve the equations.

In this context (and for all I know I'm horribly misusing the term by applying it here), instead of finding the probability equations for the combat mechanics and using them to arrive at a calculated answer for the efficiency of a weapon, I've instead written a script that mimics the actions performed by the combat engine. The script runs through the process multiple times and then averages the results to produce an approximation of the result one would get by explicit calculation.


By the way: I screwed up in my description of the efficiencies in the files I linked to. I'll be fixing that.

[Ron_Lugge]

So spinals are more efficient space wise??

I guess that just threw a wrench in my gears... Be prepared for some cross linking to another thread...

http://www.ina-community.com/forums/showthread.php?s=&postid=3886666#post3886666

:mad:60:mad:second:mad:limit:mad:

[dkass]

Very nice tables, Visage.

Are the values at range 0 correct? They seem to be worse than the lowest range bracket in the low tech cases for the PD tables?

I do have two comments on the PD tables.

First, I think the tech 0 case should not have any armor on the missile. Missile armor isn't immediately available. I don't think this matters much, though.

Secondly, I find that most "good" PD happens at ranges from one to three squares (or something like 4000 to 12000). This is more or less regardless of the ranges of the weapons (excluding things like LFG). Up until level 15, the ranges you are using work great. But beyond that, it looks like the jump from range 0 to the first range bracket is overly larger. If you need to re-do the tables, you might consider fixing the lowest range brackets in the PD tables. Maybe 0, 4000 and 8000 (2500, 5000, and 7500 might be another good set). I realize this would create a variable number of range brackets.


On a more general note, you mention AF overkill. I wouldn't expect it to be a major factor in anything but the PD case...
Thanks again for doing these.

[visage]

Originally posted by dkass
Very nice tables, Visage.

Are the values at range 0 correct? They seem to be worse than the lowest range bracket in the low tech cases for the PD tables?
That doesn't sound good... I'll go look and see what's going on.


First, I think the tech 0 case should not have any armor on the missile. Missile armor isn't immediately available. I don't think this matters much, though.

Oops. I'd planned to restrict missile armor and shields to the appropriate TLs and completely forgot... I'll fix that and regenerate the relevant tables.


On a more general note, you mention AF overkill. I wouldn't expect it to be a major factor in anything but the PD case...
Thanks again for doing these.

Yeah, it really shouldn't be particularly relevant outside of PD.

[visage]

Originally posted by dkass

Are the values at range 0 correct? They seem to be worse than the lowest range bracket in the low tech cases for the PD tables?

Looking at the tables, I'm guessing that what's going on is that the noise in my methods is showing up in those cases; there should be no difference between range 2800 and range 0, for example. I'll check and make sure that it's actually this rather than an error handling ranges less than minrange (which is certainly possible).

I'll see if I can bump up the number of trials to a point where the noise is smaller without making it take obnoxiously long to run.

[Craig P.]

Monte Carlo is an attempt to use sampling of inputs on a probability distribution to arrive at an overall qualitative result that might not be feasible to calculate directly.

The reason it's attractive with this damage analysis, from my perspective: The damage from each shot falls on a uniform distribution, combining these is troublesome, and properly accounting for cutoffs due to exhausting shield points or armor points is even more of a concern. I'm not sure whether or not it can be done, yet -- if it can be done, I'd like to.

Side observation - I'm curious about whether there are any combinations of weapons and mounts that can get off a second shot at missiles between max range and impact. If not, the TacAI should hold off on shooting at missiles until they're within the max accuracy range in order to maximize the number of hits.


@dkass - Results shouldn't change out to the near damage/accuracy distance. It's entirely possible, for the late-game techs, that this might be as long as 15,000.

General observation - For missile PD (operating under my "you only get one volley" assumption), AF is a benefit up until you start getting a preponderance of first-shot kills. Even if the third shot is usually overkill, the third shot is free (not paid for in space).

[DrPraetorious]

Originally posted by Craig P.
Side observation - I'm curious about whether there are any combinations of weapons and mounts that can get off a second shot at missiles between max range and impact. If not, the TacAI should hold off on shooting at missiles until they're within the max accuracy range in order to maximize the number of hits.

Putting html tags in regular braces weirds me out.

What if the missiles are heading for another task force and are about to veer out of range? I'd be satisfied if the TacAI always shot, and worry about niceties of range control (LR task forces fighting SR task forces should run away while shooting,) later.

I use the data patch. The ship discussed below is outfitted with thrusters - if faster system drives speed up your missiles (I have not noticed such a thing,) that would change things, of course.

My volleys of point defense missiles, which should be (45 sec * 0.09 = 4 seconds, IIRC) 4 seconds apart, are seperated by between five and six grid squares, which is apparently about 20,000 WRU. If there's a "missile speed" somewhere in the spreadsheets I don't see it, but I'll call this 5,000 WRU/sec.

So, in the 2 second recharge time of a light mounted weapon, the missiles should move somewhat less than 12,000 WRU. This is certainly consistent with my (very limited) observations of which light mounted weapons end up getting off two distinct volleys (phasers and ion cannons.)

These ships have no fighters (I'd dig up another save game if Moo3 didn't crash when I alt-tabbed out,) but I recall fighters move at about two thirds that speed, about 7,500 WRU/sec?

Updated according to visage's observation that one grid square = 3,750 WRU. WRU = Weapon Range Unit.

[visage]

Originally posted by DrPraetorious
I use the data patch. The ship discussed below is outfitted with thrusters - if faster system drives speed up your missiles (I have not noticed such a thing,) that would change things, of course.

My volleys of point defense missiles, which should be (45 sec * 0.09 = 4 seconds, IIRC) 4 seconds apart, are seperated by between five and six grid squares, which, again IIRC, is 25-30 Gm. If there's a "missile speed" somewhere in the spreadsheets I don't see it (windows can, but I'll call this 7 Gm/sec.

I did some tests a week or two ago where I took starting system engines and boosted them to 15k speed. The ships I built certainly zipped around, but the missiles (or was I using fighters?) didn't appear to travel any faster.

Also, I think somewhere in the Armor/Shields threads I have some numbers indicating that DF reload times don't quite line up with the numbers displayed on the timer.

[Craig P.]

Originally posted by DrPraetorious
Putting html tags in regular braces weirds me out.

What if the missiles are heading for another task force and are about to veer out of range? I'd be satisfied if the TacAI always shot, and worry about niceties of range control (LR task forces fighting SR task forces should run away while shooting,) later.That's a good point. At the rate this is going, I can understand shooting first and asking questions later when there's a limited amount of processing time for the TacAI. It's just annoying to consider that a volley targetting a missile could throw away 90% of its effectiveness (throwing a number out there that might be an exaggeration) because the TacAI's not sophisticated enough.

(Then again, does it need to be that sophisticated? Given a fixed missile velocity and a change in bearing, it should be a relatively simple calculation to figure out when the missile will be within max accuracy range, and if it suddenly starts heading away, the guns that were waiting to fire can open up immediately. Of course, how well this works depends on the time slice granularity.)

[Craig P.]

Originally posted by DrPraetorious
So, in the 2 second recharge time of a light mounted weapon, the missiles should move 12-15 Gm.1 Gm = ??? range units

[visage]

Originally posted by Craig P.
1 Gm = ??? range units

I think I found that a square is appriximately 15000 range-units-for-engine-purposes and approximately 3750 range-units-for-weapon-purposes. If I don't mention it up in the Guide at top it should be findable in the Armor/Shields thread.

[DrPraetorious]

Thanks again, Visage. I fixed the numbers in the original post.

[clb]

There is, however, an effective cost per unit space - the cost of the rest of the ship divided by the total free spece. For example:

A Lancer at TL 0:
Hull cost: 50PP
Basic Equipment is 9% at 75PP
Retro & Thruster@1500: 50% at 150PP
Light Zortrium Armor: 0% at 10PP

cost is 235PP for 20 hull spaces
or a cost of 11.75PP/ unit space.

At the far end:

A Leviathan at TL 45:
Hull Cost: 11122
Bas. Eq. : 9% @ 75PP
WF X & TW@4200: 33.8% @ 905PP
UHvy Adamantium 0% @ 5580PP
Small Damper Field: 7.5% @ 525PP

cost is 18207 for 2293 hull spaces
or a cost of 7.94/ unit space.

but to work this into the space/cost trade off is complicated.:rolleyes:

For example, reducing system speed reduces this cost per unit space, although it does nothing for the cost as a whole.

[Ron_Lugge]

Visage; as a general rule I have LOTS of proccessor time sitting around and doing nothing. E-mail me the stuff I need and I'll be GLAD to run the "Monte Carlo" for a long time.

Just note that you won't get the results back TOO fast; but they will be reasonably accurate (I have this thing sitting around for several hours after I get home from school; surely I can run the program then!)

Note that I am NOT offering to help out with the data analysis; I'll just E-mail the results back to you (assuming you provide sufficient instructions to allow me to do so.)

[Craig P.]

Originally posted by DrPraetorious
So, in the 2 second recharge time of a light mounted weapon, the missiles should move somewhat less than 12,000 WRU. This is certainly consistent with my (very limited) observations of which light mounted weapons end up getting off two distinct volleys (phasers and ion cannons.)Ah, that's helpful. If you can get more specific (there might be some mountings right near the border), that would help even more. It complicates the analysis, too, because somewhere in the mid tech-range, there will be a full-range volley plus a short-range volley and a one-volley kill is no longer as important.


I'm going to take a shot at setting up some Monte Carlo stuff myself. I don't see that it should take that long to run, just thinking idly about it. I'd expect to be able to run a million shots in a minute or so, and I wouldn't think I'd actually need that many. If I do it, I'll also try to incorporate some statistical post-analysis to make sure I've actually got enough data for my conclusions.

[visage]

I've modified the script so that it also ranks by what I can Space Adjusted Cost, which is essentially doing what clb did above to determine the value of space and adding that to the AU cost of the weapon.

Tonight, I'll probably change it so that for PD it samples 75%, 37.5%, 18.75%, 9.375%, and 4.6875% of max range.

After that, i'll regenerate the tables.

[Ineffigy]

Do you know for a fact that the armor piercing does not work? Would it not just decrease the chance of hitting armour vs internals by .25 ? Basically, you have a higher chance at hitting the internals than striking the armour right? I would think that would be a hard thing to test for in game.

Ineffigy

[visage]

Originally posted by Ineffigy
Do you know for a fact that the armor piercing does not work? Would it not just decrease the chance of hitting armour vs internals by .25 ? Basically, you have a higher chance at hitting the internals than striking the armour right? I would think that would be a hard thing to test for in game.

Ineffigy

Well, I can't know for a fact, but if you look over in the Armor/Shields thread, you can see me explain the tests that have been done and why I at least am convinced that AP does nothing.

I recently revisited the issue, so the end of the thread should be the place to find the best discussion of it.

[Craig P.]

Just realized, although there are comments that shield penetration works as advertized, I don't see a mention of exactly how that is. Is it a straight multiplier on stopping power?

Also, I've implemented a Monte Carlo program and threw some stuff at it. PD mount (even with proper multifire) isn't looking too good in the face of missile shields and armor. So far, the only PD weapon worth a damn against an HE X-Ray / Light Missile / Neutronium / Class VI combo (assuming anything less than substantially more mounts than targets) is a light mount LFG, thanks in large part to its shield penetration. It knocks out its target in one volley about 10% of the time, and per the observation upthread about missile velocities most weapons up to this level will get only one volley.

I'd really like to see armor piercing and/or penetrating working, because I think they could help here. I'd also (still) like to see the TacAI shoot at closer range, because shooting at max range means large damage penalties for non-mass weapons and poor accuracy for everything.

[visage]

Originally posted by Craig P.
Just realized, although there are comments that shield penetration works as advertized, I don't see a mention of exactly how that is. Is it a straight multiplier on stopping power?

Yup. Multiply StpPow by ShldPen.


Yeah, the PD mount also still fares relatively poorly, unfortunately. Of course, the mod-workaround of reducing its FireDlay actually improves it. :)

[EmperorKosh]

I haven't read too much of this topic... only two posts on the first page but I hear visage is doing MOO3 combat testing and I heard some other guy talking about building cheap.

Well in my most successful game, (my second true game where I am a customized Imsaeis) I am at war with three alien civilizations. One is mentally retarded (the Psilons, go figure, the morons have military strength of 1 and 7 worlds when I have probably a military strength rating of about 500 and I have about 70 worlds), the other is a war monger and probably the most challenging out of the three (the Ithkul), and the other opponent are the rocks (the Silicoids).

Well anyways, it has appeared to me that a lot of the veterans on here play with really big ships later on in the game. Well, I myself in the game with my custom Imsaeis play with probably smaller size ships. The bulk of my spaceforce is composed with IF frigates, LR frigates, and SR frigates. I have LR available in Destroyer, Battleship, and Cruiser sizes as well but the amount of those I have is much less than the frigates. Oh yeah and I also use CV's that are lightcruisers/cruisers.

Anyway, on to my point. It's about turn 210 or so and my whole spaceforce is composed of literally over 1000 ships. I have found that my mainly frigate-composed-armadas are actually quite effective. Especially when you've got about 5-10 of them in one system. Together, they have a huge amount of firepower and I recieve almost 0 casualties through all of my battles because of their power. In my game with the Imsaeis, I could literally declare war on the whole galaxy and still win. (No joke).

Anyway, visage since you seem to be the one doing all this military testing and analysis; I encourage you to test the idea of quantity over quality. (LOTS AND LOTS OF small-medium size ships vs several biggen ship armadas). Also, I've found that throwing in a couple of huge ships in my armadas is also useful.

Anyway, I encourage all you military analysis peoples to test my theory. Lemme know how the theory sounds to you guys as well.

[Ron_Lugge]

Kosh, which would win:

My 10000+ dreadnoughts, or your 100,000 frigates? I'll give you a hint... only 180 ships can be in the fight for each side.

[Vorlock]

Myself, i always try to get max hull ships if possible, limited only by my manufacturing capacity. I always assume i will hit the 10 task force limit and thus want the biggest ships possible.

However, early game i use masses of small ships alot. It would be interesting to see a cost benefit analysis.

It should also be noted that if one guy has 100,000 dreads, the other guy should have like 1,000,000 or more frigates. Yes, most of the battles should go the dread guys way, but he will take losses and eventually might get killed off. The frigate guy should also be able to replace ships faster. What i am unsure about is at what point can the frigates simple not hurt the bigger hull size fleets at all. At that point battle is pointless. 100, 1000, or 10 million frigates wont matter.

[Da_Blade]

It all depends on the foe. Against a bigger foe i would recommend smaller mainly indirect fire ships. These can do quite a lot of damage and flee if they come too close. Smaller ships always get my preference too. I usually design small carriers and missile boats, though slightly larger then my recon and point defense ships. Then i have big SR/LR ships, which are expensive and only build by my top-line industry's. Because the other ships are small they are build in good proportions by the VRs, i don't have to manage the building of those, only the big ships.

Of course, against the current AI 200 ships is the same as 5 billion ships, as you'll seldomly have more then 10 TFs active anyways, and they are invulnerable for the AI's frigate SR detachments....

But against human players i like lots of small ships much better then big powerfull ships. You'll seldomly get to the point where you'll fight with 10 full armada's against each other anyways, so smaller ships are much better, since they are easier to replace and since you have much more of them in reserve you can pop up some ships wherever or whenever they are needed.

[DavidByron]

How many seconds are there between sucessive fighter waves?

[visage]

Originally posted by DavidByron
How many seconds are there between sucessive fighter waves?

*sigh* I checked this once, and I don't think I recorded the answer. The standard answer people throw out is 120, but I seem to recall finding it to be somewhat less than that.

[visage]

Originally posted by visage
*sigh* I checked this once, and I don't think I recorded the answer. The standard answer people throw out is 120, but I seem to recall finding it to be somewhat less than that.

Ok, just ran a quick test. It's approximately 90 seconds between fighter waves; or, at least, that's the difference I got between my first two waves in this test, and that matches my memory from the last time I looked at it. I'll have to add that to my local copy of the top post and get around to pushing the new version.

[DavidByron]

*sigh* I checked this once, and I don't think I recorded the answer

Well that ain't good enough -- you're the expert on this. You have a reputation to maintain! I expect you to, oh hang on......

Ok, just ran a quick test

Ok then! Thanks. :)

I guess a 13 minute turn around from request to results is adequate :D

[visage]

Originally posted by DavidByron

I guess a 13 minute turn around from request to results is adequate :D

We aim to please. ;)

[Ron_Lugge]

~90 seconds is pretty accurate. I tend to run 3 minute battles; I launch at the beggining of the ifght, the middle, and just at the end (barely; right when it says draw/just after)

[renaux]

my battles tend to run around 2 min max, but I go w/ IF exclusively (just about)

[EmperorKosh]

You can only have 180 ships per combat thing? Bullcrap, I've had 300+ ships in one system and I've been able to use all of them.


Yes Ron, my frigates would win. :p

[Ron_Lugge]

Originally posted by EmperorKosh
You can only have 180 ships per combat thing? Bullcrap, I've had 300+ ships in one system and I've been able to use all of them.


Yes Ron, my frigates would win. :p

Sounds like you were running a mod - mods can change the 10 tf limit, or the 18 ships per tf limit.

edit:

Note that large numbers of ships can be reported (ie, I once saw 400+ system ships reported) but only 19 system ships will be allowed in for some reason.

[Minor Annoyance]

Since we know AP doesn't work, has anyone found anything that can be edited to make them work? I would assume if there was, someone whould have done it already. But I'll ask anyway.

[Heisher]

Nopt sure..

butthanks for the guide!

[visage]

I've pushed my local copy of the guide to this thread again. The only major change that's coming to mind is that deflection blocks damage less than, not less than or equal to.

There may be other minor changes; I'm fairly certain I fixed a typo or two and the stuff about fighter mods affecting HP might not have been in the last rev.

I've got more tables generated for DF weapon efficiencies and will get around to posting them coherently soon. Perhaps I'll even get around to writing more of the "Implications" section. :)

[visage]

Good news on armor piercing: QS has become aware of the bug, and the fix is going into the patch.

Heh: now I wonder if we're going to be told how AP is supposed to work or if I'll end up testing the various theories again. :)

[Ineffigy]

It'll be faster if you figure it out. Oh wise an honable one ... (was that brown nosing enough?)

[visage]

Another update. This one fills in several of the blanks on how armor weights get doled out.

IF/CV/Recon aren't really covered because right now it's pretty buggy, AFAICT. I'll probably buckle down at some point and determine how it behaves in most cases. Right now, you get behavior like...

If you have all armor weights available, everything gets Very Light... except for lancers and cutters, which get Ultra Heavy.

If you have all armor weights except medium available, *everything* gets Ultra Heavy.


Given the VAI's predilection for small hulls pre-code-patch, it's no wonder that AI armor tended to be so poor.

[BlackHat]

What happens if you make up a new armor tech weight and stick it in the armor tables. (Something between None and Light, or Something between Medium and Heavy? or add an even heavier one to the end. How does that change the "picks"

If you have all armor weights except medium available, *everything* gets Ultra Heavy

Later,
BlackHat

[Vorlock]

Here is a suggestion for you Visage:

Given the sometimes not so intelligent behavior of the autobuild feature (which presumedly the AI uses as well), it would be nice to have a guide on how to mod your files so that the autobuilder would work more effectively. This would not only help people who dont feel like manually building ships from scratch, but would also help the enemy AI.

What i am thinking of is something like:

Change the armor table like THIS so that the more effective armors are choses.
Change the weapon mounts table like THIS so that the proper mounts are chosen once available. (IE light over pd, ultra heavy over standard spinal, etc.)
And anything else you have discovered that i forgot.

Whats good about this is you should be able to do it before the patch comes out and still have it work reasonably well AFTER the patch comes out.

[visage]

Originally posted by Vorlock
Here is a suggestion for you Visage:

Given the sometimes not so intelligent behavior of the autobuild feature (which presumedly the AI uses as well), it would be nice to have a guide on how to mod your files so that the autobuilder would work more effectively. This would not only help people who dont feel like manually building ships from scratch, but would also help the enemy AI.

How about this (http://www.ataricommunity.com/forums/showthread.php?threadid=292060) thread?

Now, it doesn't have "Visage's plan on fixing AI armor" in it yet, I'll admit, but I think it's reasonably clear on most other things.

I've started thinking about actually sitting down and publishing a mod that will do all this. Right now, I'm dithering about scope; my recent posts in that thread are me rambling about my thoughts on what should be done and soliciting input on where the line should be drawn at the extent of the mod.

Meanwhile, I'm also thinking about what I'd want to do if I were willing to revamp anything and everything that can be modded about space combat... I should have a post up reasonably soon rambling on that subject as well.

[Vorlock]

So, havent really had time to play the new, but was wondering about the big errors from the past.

Does armor peircing really work now? If so, what exactly does it do? It says "makes it 25% more effective" but thats kinda vague. :)

How about point defence? I never actually had the PD bug, but i'd love to hear that PD is more effective now than it was. Any thoughts?

[visage]

Originally posted by Vorlock
So, havent really had time to play the new, but was wondering about the big errors from the past.

Does armor peircing really work now? If so, what exactly does it do? It says "makes it 25% more effective" but thats kinda vague. :)

How about point defence? I never actually had the PD bug, but i'd love to hear that PD is more effective now than it was. Any thoughts?


They've changed a bunch of things... once the real patch is out, I'll try to identify all the changes and revise at least the top post of this thread.

I haven't tested to make sure that armor piercing is now working, but the way it's supposed to be working now is that the chance to hit armor once you've gotten past deflection is multiplied by your AP value. Normally, when you hit a ship with 150 HP and 50 armor with 10 points of damage, your first hit is guaranteed to hit armor, knocking it down to 40/50 armor. Your next shot would have a 4/5 chance of hitting armor. If it did, knocking armor down to 30/50, your next shot would have a 3/5 chance of hitting armor. If you've got AP of 0.5, your first shot would have a 50% chance of getting past the armor. If you hit the armor that time, there'd be 4/5 of the armor left, so you'd have a 2/5 chance of hitting the armor with your second shot.

PD seems to be working better. The PD mount itself is still broken, though.

[Beamup]

Originally posted by visage
The PD mount itself is still broken, though.

How so? Is the MaxAccy bonus it's supposed to now be getting not working?

[visage]

Originally posted by Beamup
How so? Is the MaxAccy bonus it's supposed to now be getting not working?

Precisely.

At least:
1) I tested mount mods to MaxAccry pre-1.2 and found it didn't work.
2) Tom Hughes tells me that they didn't make any changes to that in the code patch; that the code for it was there in 1.0 but unused.
3) Bhruic says that the savefiles show that PD's accuracy bonus doesn't exist.

[Lewit]

After reading over this well put together guide and the related threads, something came to my mind with the depth of analysis. I'm going to work on this a little bit tonight, but I thought I'd throw it out there as an interesting bit for you to work over as well.

There is a great deal of work here about the most efficient use of space for PD and attack at any given tech level, but I do recall you mentioning very early that once you reach a point after the "one shot kill" the numbers aren't as linear. What I'm proposing is to look over these same numbers with very specific tech levels in mind. So, for example, compare the point defense capabilities of each weapon against the shields and armor found at tech levels 5, 10, 15, etc... Maybe someone out there has already considered this, but I'm thinking this would be useful for maximizing your designs when fighting a lesser or (more importantly) a more superior oppontent. If lasers and mass drivers both had a +/- 7% chance or knocking out a higher tech missile, I'd like to stuff more of the space effiecient weapons in, despite the fact that one may be better if the tech levels were equal.

Anyway hope that is clear, and that it hasn't already been tackled. I'll run a few calculations tonight, but I'm working mainly in Excel on my roommate's computer from 1864, so it'll be pretty rough work.

[visage]

Originally posted by Lewit
After reading over this well put together guide and the related threads, something came to my mind with the depth of analysis. I'm going to work on this a little bit tonight, but I thought I'd throw it out there as an interesting bit for you to work over as well.

There is a great deal of work here about the most efficient use of space for PD and attack at any given tech level, but I do recall you mentioning very early that once you reach a point after the "one shot kill" the numbers aren't as linear. What I'm proposing is to look over these same numbers with very specific tech levels in mind. So, for example, compare the point defense capabilities of each weapon against the shields and armor found at tech levels 5, 10, 15, etc... Maybe someone out there has already considered this, but I'm thinking this would be useful for maximizing your designs when fighting a lesser or (more importantly) a more superior oppontent. If lasers and mass drivers both had a +/- 7% chance or knocking out a higher tech missile, I'd like to stuff more of the space effiecient weapons in, despite the fact that one may be better if the tech levels were equal.

Anyway hope that is clear, and that it hasn't already been tackled. I'll run a few calculations tonight, but I'm working mainly in Excel on my roommate's computer from 1864, so it'll be pretty rough work.

I've got a perl script that I'm constantly tweaking. Right now, it takes a TL, a metric for optimization (space, AU, or AU + Space*Cost_to_Build_Space), and whether it's for PD or shipkilling.

If it's for shipkilling, it constructs the largest hull, with best armor, shields, and sees how efficient every available weapon + mods (doing enveloping and armor piercing as combinatorics) + mount is at killing it at different ranges.

If it's for PD, it does the same for the biggest, most advanced, best-defended missile at that TL.

I've got some sample tables from an earlier version posted a ways earlier in this thread, I believe....

Yesterday, I created an offshoot version, that handles all mods in combinatoric-fashion, but only handles PD. That one I intend to use to figure out how valuable I should mark the different mods for purposes of training the autobuild about PD.

The first one's soon going to get an upgrade where for PD it marks the weapon/mount/mod combo that autobuild will use. I'll use that for tweaking the values until autobuild generally gets it right.

[Lewit]

Originally posted by visage
If it's for PD, it does the same for the biggest, most advanced, best-defended missile at that TL.

It sounds like that's exactly what I'm talking about. I only just now downloaded the table links from this tread, and was planning to go over them tonight (I think I have 8 to 10 from you in total); maybe it is in there. I like the idea of testing all the weapons at each tech level, which will essentially let you squeeze every inch out of a ship when you know your opponent's TL, and may have to play catch up.

After I check out these tables I'll get back to you, but it sounds like all the work has already been done!

[visage]

I was going to just crank out a table at every fifth TL for PD and another for shipkilling... but then the patch became imminent and I put it off until that came out. Now it's out and the PD mount is still broken, so I'm hoping they'll fix that before the non-beta patch comes out. (I know that Tom's intending to.)

[Inevitable]

PD isnt working now after the code patch?

[visage]

Originally posted by Inevitable
PD isnt working now after the code patch?

The 1.2b code patch apparently fixes the problem that PD was really bad at acquiring targets.

There was also the problem that the "Point Defense Mount" for direct fire weapons included a MultFire*=2 bonus in the data files which the game didn't actually process; thus the PD mount was broken and you always would rather have the Light mount.

In the 1.2b code patch, they frobbed a bunch of stuff about the PD mount, including giving it a 50% bonus to accuracy. Unfortunately, as far as I'm aware, mounts didn't and don't affect accuracy. So, once again, the PD mount is broken.

[Lewit]

Originally posted by visage
I was going to just crank out a table at every fifth TL for PD and another for shipkilling... but then the patch became imminent and I put it off until that came out. Now it's out and the PD mount is still broken, so I'm hoping they'll fix that before the non-beta patch comes out. (I know that Tom's intending to.)

I can see your reason for delay; a good deal of work will need to be recalculated with any changes made in the patch.

I took a look over your tables, and they are fairly close to what I was talking about yesterday. They do seem bias towards only selecting roughly equivalent techs against the defenses though, which isn't exactly what I was talking about.

To take an extreme case, imagine being at TL 50 against TL5 - there probably isn't much need to put the best guns on a ship since they would be overkill - I guess the question is where does the efficency start to drop off? Do I need to be churning out expensive ships to knock out that weak opponent, or can I get by sending all my big guns to the front line, and building a more cost savvy style for taking out the weaker race. I guess that is a bit extreme, since the worst span in technology I have in my current game is only about 5 to 10 levels below me, but I still need my best ships elsewhere and don't always have the reserves to move that last front forward instead of sitting on pure defense.

Anyway, I guess the question will have to wait until the patch since anything that gets calculated may need to get reworked anyway. Thanks for the imput, and great work on the numbers.

[Minor Annoyance]

Hey visage, was your Autobuild mod made before or after the code patch, and if it's before will you be changing it?

[visage]

I designed it before the code patch and implemented it afterwards, so it works with the code patch.

I'm still pondering the effects of the changes to, say, missiles, and want to explicitly test the new PD mount, etc.

[Ron_Lugge]

Originally posted by Lewit
I can see your reason for delay; a good deal of work will need to be recalculated with any changes made in the patch.

I took a look over your tables, and they are fairly close to what I was talking about yesterday. They do seem bias towards only selecting roughly equivalent techs against the defenses though, which isn't exactly what I was talking about.

To take an extreme case, imagine being at TL 50 against TL5 - there probably isn't much need to put the best guns on a ship since they would be overkill - I guess the question is where does the efficency start to drop off? Do I need to be churning out expensive ships to knock out that weak opponent, or can I get by sending all my big guns to the front line, and building a more cost savvy style for taking out the weaker race. I guess that is a bit extreme, since the worst span in technology I have in my current game is only about 5 to 10 levels below me, but I still need my best ships elsewhere and don't always have the reserves to move that last front forward instead of sitting on pure defense.

Anyway, I guess the question will have to wait until the patch since anything that gets calculated may need to get reworked anyway. Thanks for the imput, and great work on the numbers.

I've seen spans in the 30 TL range... :eek:

I had tech 50, they had about tech 20.

[visage]

I don't know whether it was working before the code patch, but mounts affect accuracy now. At least, now that I've gotten around to testing it, it certainly looks that way.

I'll be updating my simulator again, I guess, to handle mounts that apply different multipliers to near and far damage....

[Vorlock]

Doesnt that mean that the PD mount might not be "broken" anymore?

[visage]

Originally posted by Vorlock
Doesnt that mean that the PD mount might not be "broken" anymore?

It's not broken. However, my simulations show that it doesn't fare as well at most TLs as the Light, still...

[Lewit]

Originally posted by Ron_Lugge
I've seen spans in the 30 TL range... :eek:

I had tech 50, they had about tech 20.

You can see then why I'm thinking along these lines. Not that the poor saps need to get any more obliterated, but streamlining the empire to be churning out ships at the most effective rate is always a consideration. Its all about the little things... if it wasn't, would we even be having these discussions?

[hoof1]

I have a rather significant question about armor and deflection.

In the real world, you have armor that "resists" projectiles. For example, a particular armor plate can resist a particular projectile up until a certain striking velocity is achieved. Beyond this striking velocity, the round will penetrate. However, the act of forcing its way through the armor "slows" the round down and reduces its effectiveness once through the plate. Thus the "deflection" of this plate is in full effect, even if it is not enough to completely stop the projectile.

Take this Moo3 example. A ship equipped with class 10 shields and ultra heavy Adamantine armor is intact. It then gets struck by 257 points of damage. 86% of this damage is absorbed by the shields, leaving 36 points of damage left. The deflection value of Adamantine armor is 36 in the ultra heavy configuration. Since remaining damage is not more than the deflection of the armor, no damage occurs to armor or internals.

Now say that same intact ship is hit by 258 points of damage. 86% gets absorbed, leaving about 37 points of damage left to hit the armor. The deflection is 36. How much "damage" gets applied to the armor/internals? Is it 1 (37 - 36 = 1) or 37?

If I were writing Moo3, I'd have "deflection" be a raw subtraction of the damage that gets past the shields. In this way the deflection reflects armor technology for the last 15 or so centuries. Thus if the armor deflects x amount of damage, then x amount of damage is subtracted from stage 3 of visage's armor/shields description, before damage points are applied to armor/internals. I can think of no rational explanation for why a miniscule amount of extra damage would suddenly cause *all* the damage to be applied, vs none w/o that miniscule extra damage.

So which way does it work? Does deflection act simply as a test (if damage is less than deflection then no damage), or does deflection work like Moo2's shields (which subtracted directly off of damage before subtracting hitpoints)?

If it's merely a test, one only needs to overmatch deflection to ensure that deflection is meaningless. If you can get 37 points of damage past the shields in a single hit, it doesn't matter if you have very light Zortrium or ultra heavy Adamantium. If you get 37 points of damage through, all that matters is how many hitpoints the armor has and how much of it is "gone" before the hit lands (to determine if damage gets through to internals).

If deflection is simply a subtraction of damage, then suddenly the stronger/thicker armors mean more than simply more hitpoints, but even for weapons that exceed the deflection value, they will have a tougher time doing damage to the ship. As you can see, this is more "linear" and doesn't have a sudden effectiveness dropoff when you edge past deflection.

[visage]

Deflection is merely a test, not a subtractor.

However, it is a significant factor in the effectiveness of various weapons.

First off, shields multiply the effectiveness of deflection by reducing the amount of damage that gets to the point where you check against it.

Second, most weapons do damage that's randomized. It's not that your neutron blaster does 33 damage, it's that at point blank range it does between 3 and 33. And at its max range, a standard neutron blaster does 0-23 damage.

Of course, that's the whole point of the spinal series... DmgDelta isn't multiplied by mounts, so an Ultra Spinal Neutron Blaster does 135-165 damage at point blank range and 89-119 damage at max range.

Raising the lower end of the damage means that you're far less likely to lose damage to deflection after shields...

[hoof1]

Out of curiosity, how did you determine it was a test and not a subtractor? I apologize if you've already addressed this elsewhere in this thread. I'm at work and don't have time to read 5 pages of this thread :)

[Minor Annoyance]

I would assume because he was firing at armor with deflection of, lets say, 10 and he only ever saw damage equal or greater than 10.

[Beamup]

More likely he modded some weapon to do exactly 10 (or something) damage every time and perfect accuracy, fired it at some armor modded to have 10 (or something) deflection, and observed whether there was any damage or not.

[Minor Annoyance]

Well, I though the modding was implied...

[Beamup]

Originally posted by Minor Annoyance
Well, I though the modding was implied...

Ah, but the modding of the weapon would be the critical part of the test! Just observing what range of damage numbers appeared shooting at a given armor (which is what you seemed to be implying) would prove very little.

[dkass]

hoof1,

Think of the damage as "below the waterline" not above deck. Thus once the hull is breached, the reduction in KE due to the armor is mostly irrelevant. Note also that in real life, hits that are deflected can still do damage to the armor, weakening it for the next shot.

[visage]

Originally posted by Beamup
Ah, but the modding of the weapon would be the critical part of the test! Just observing what range of damage numbers appeared shooting at a given armor (which is what you seemed to be implying) would prove very little.

Actually, think what you'd see if you fired a weapon that randomly did between 1 and 20 damage against an armor with 10 deflection...

[hoof1]

dkass, that's how I figure armor worked in Moo3. Armor "points" were a rating of how "beat up" the armor system gets. Deflection is the amount that the armor can absorb indefinitely. Anything over that "beats up" the armor.

Your naval reference is one way to look at it. Armored ships (of ww2 and before) fired exploding rounds. The round only had to penetrate the armor, survive (so it could explode), and travel far enough inside to explode.

The flaw in the example above is that if Moo3 modelled that, then any shots that got past the armor into the internals should do almost full damage. Quite clearly it does not work that way. Moo3 seems to be modelled on the kinetic-energy theory that dominates solid shots vs armored targets concept (where protective systems reduce the projectile's kinetic energy and thus it's ability to do damage). But then it has a hitpoint system for the armor, which shoots that theory all to heck as well.

Tanks and battleships in the last century tended to be "destroyed" far before their armor systems were seriously compromised. A few small holes (where the killing shots passed through) aren't going to make the rest of the armor less effective. A tank or a battleship is akin to a Corvette in Moo3 equipped with Ultra Heavy Adamantine Armor, where any penetrating hit will severely damage/destroy the vehicle, yet the armor is not degraded significantly.

[hoof1]

visage, you'd see yellow numbers from 11-20 show up with no numbers below 11 in your example. Is that what you saw?

[Beamup]

Originally posted by visage
Actually, think what you'd see if you fired a weapon that randomly did between 1 and 20 damage against an armor with 10 deflection...

Thought of that and rejected it. AFAIK (and you've never addressed that I can find) nobody knows whether damage is really integral or fractional. So it would seem entirely possible to me that if you did that, some 10.2 damage hits would get through... and be displayed as 10 even if the deflection was <=. Or, if it rounded up, you'd never see 10 even if the deflection was < (though in that case the distinction between < and <= would be largely cosmetic). And even if damage is integral, such a test would still potentially fall prey to bad luck - it could never be 100% conclusive without infinite trials.

Whereas the test I proposed is 100% certain, totally unambiguous, with a single test, regardless of integral vs. fractional damage.

[visage]

I assume the thought behind the MoOIII armor system is essentially that each time you hit armor, you've created a weak spot. Hitting internals is passing through a weak spot you've already created.

The decision of deflection as threshold over reduction... *shrug*

[hoof1]

Another problem with determining whether deflection is a test or a subtraction is knowing how much damage the armor took vs how much was displayed.

If the damage display code had something like:

if (damage deflected)
{
Display Damage the Shield absorbed;
}
else
{
Display damage absorbed by armor;
ArmorHP = ArmorHP - (Damage - deflection);
}

In this pseudocode example, you'd not know that deflection was subtracted.

Also, if any deflected damage showed up as shield damage visually, then you'd have to ensure the target has no shields.

Can you run your deflection test again against targets with no shields? I suspect that for a deflection of 10, you'd start seeing yellow damage values from 1-20 in your example, if there are no shields.

[Beamup]

Originally posted by hoof1
The flaw in the example above is that if Moo3 modelled that, then any shots that got past the armor into the internals should do almost full damage. Quite clearly it does not work that way.

Actually it does... any shot that does any internal damage was totally unaffected by the armor.

Although there is one case I haven't seen explicitly addressed - suppose a 20-point attack hits a target with 10 armor points. Is there any internal damage?

[visage]

Originally posted by Beamup
Actually it does... any shot that does any internal damage was totally unaffected by the armor.

Although there is one case I haven't seen explicitly addressed - suppose a 20-point attack hits a target with 10 armor points. Is there any internal damage?

I'm under the impression I tested that case and found that it works as I describe at the top of this thread: subtract the damage done to armor and apply the rest to internals.

I don't actually remember the test in question, so perhaps I'm merely assuming here. It's easy enough to test...

[Beamup]

Originally posted by visage
I'm under the impression I tested that case and found that it works as I describe at the top of this thread: subtract the damage done to armor and apply the rest to internals.

I don't actually remember the test in question, so perhaps I'm merely assuming here. It's easy enough to test...

When I read what you wrote:

Originally posted by visage
If the random number is less than the ratio of the remaining armor to max armor, the remaining damage hits the armor: subtract that many armor points. Display a yellow number equal to the amount of damage that hit armor.

it seemed entirely unclear. It would (IMO) make it more clear if you put in after this, "Carry any damage greater than the remaining armor to step 4."

[The Bursar]

[QUOTE]Originally posted by visage

Warheads

Warhead Number Damage Cost Space TechLvl
--------- ------ ------ ------ ----- -------
Nuclear 01 56 14 6 0
Anionic Energy 02 82 15 6 7
Neutronium 03 121 17 7 15
Hercular 04 178 19 8 18
Merculite 05 262 21 8 19
H.E. X-Ray 06 386 23 9 26
Scatter Pack 07 568 25 10 31
Ionic Pulsar 08 836 28 11 35
Energy Pulsar 09 1231 31 12 41
Omega 10 1811 34 14 45

All warheads have FireDlay = 45, MaxAcc = 0.95.

[QUOTE]

These damages are prepatch, correct?
If so do I round the halved damages up or down?

[visage]

Originally posted by The Bursar

These damages are prepatch, correct?
If so do I round the halved damages up or down? [/B]

Nope, those are post-datapatch, and thus currently correct, AFAIK.

[visage]

Originally posted by Beamup
Actually it does... any shot that does any internal damage was totally unaffected by the armor.

Although there is one case I haven't seen explicitly addressed - suppose a 20-point attack hits a target with 10 armor points. Is there any internal damage?

I just tested this, and I'm a bit surprised:

In that case, you do 20 internals.

(First, took a ship with a 75-damage weapon and fired it at some Hawk Defense ships. It displayed 75 red, which sort of surprised me, so I went to test that it was correctly displayed. Fired a weapon that did 200 damage at a Hawk Defense. Hawk Defense blew up and a red 186 was displayed.)

[hoof1]

Just a sanity-check. Does a Hawk Defence *have* any armor? If not, that would explain why it took 75 red damage.

[Craig P.]

It should be very light Zortrium armor.

[Ron_Lugge]

Originally posted by Craig P.
It should be very light Zortrium armor.

Not necessaryily - you don't always start out with light/very light armor weights - I've had to wait for heavy/ultra heavy more than once.

[visage]

Originally posted by Ron_Lugge
Not necessaryily - you don't always start out with light/very light armor weights - I've had to wait for heavy/ultra heavy more than once.

Starting ships are predefined.

Unless I'm confused on this, they'll have Very Light Zortrium regardless of what your Tech Tree looks like.

If I'm confused about this, I guess I need to go back and test with a check to make sure that the forces I'm attacking actually have armor.

[Ron_Lugge]

Originally posted by visage
Starting ships are predefined.

Unless I'm confused on this, they'll have Very Light Zortrium regardless of what your Tech Tree looks like.

If I'm confused about this, I guess I need to go back and test with a check to make sure that the forces I'm attacking actually have armor.

Oh... Never noticed that. I just remember being PO'd over not having any armor at all untill late game.

[visage]

I just did some testing on TF AI. The chart that Blaze/Renaux put together was only about Control mode, and I'm of course most interested in Watch mode. :)

The first thing I discovered was that the AI is very happy to retreat now. :)

The second thing I discovered was that the AI's retreat logic seems to be based in the same numbers that are used for the lower-right-hand-corner display during combat. This means that if you use a weapon that has a FireDlay below the floor it thinks your weapon is far more powerful that it actually is...

So, most of my tests on this used a single-ship TF, where the ship had near-infinite armor points, a laser interceptor, a PD nuke, a spinal laser, and 3k system speed. The laser was modded to have 17k range and a FireDlay of 0.01. The ship was usually designated as an LR ship. (The extremely low FireDlay meant that the AI thought the ship was far more deadly than it actually was, meaning it would neither retreat against the apparently superior enemy forces of three Hawk Defenders nor would it blow them away instantaneously.)


My tests suggest that there are three different TF AIs.

A "Heavy Combat" AI: On watch mode, it prelaunches fighters and missiles; on combat mode, it auto-launches and autofires (as in, will launch on and fire on a target as soon as it's detected and in range). When in watch mode or when directed to attack a target, it will charge in and circle at short range. LR and SR seem to have this AI.

A "Medium Combat" AI: On watch mode, it prelaunches fighters and missiles; on combat mode, it autolaunches and autofires at about 50% of max DF range. When in watch mode or when ordered to attack, it will compare its DF stats with the DF stats of the target as to whether it should charge or run away from the target. IF and CV seem to have this AI.

A "Light Combat" AI: On watch mode, it prelaunches fighters and missiles; it will charge when it can't see the enemy and then try to maintain a moderate amount of range when engaging an enemy TF. I haven't really tested its Control mode behavior yet. It's also quite possible that they maintain much more range when they don't have firepower superiority. Colony and Outpost AIs have this behavior.

I'm not sure yet if Recon falls into one of the above categories, takes elements from two, or is its own. My initial impression is that it moves like Medium and fires like Heavy.


When I found that the CV and IF TFs were charging in instead of trying to move obliquely away from their target AIs, I wanted to see if this was because the ship in the TF was LR, or if it was because the ship outgunned the enemies. Using an identical ship that was designated IF had the TF still charging the enemy. Using a ship that wasn't armed with one of the super-lasers had the TF running away. Using such a ship in a LR TF had the TF still charging into short range. (It doesn't matter if the wimpy ship is IF or LR.)

It's conceivable that the behavior is somehow different for multi-ship TFs than the single-ship TFs I was testing.


Now, unless the TF AI assignments are hidden somewhere in the spreadsheets, these results are largely useless to me. I'd been hoping to find a behavior that could be traded with LR's to make the AI actually have LR TFs that try to maintain range.

Unfortunately, the AIs that do that are Colony and Outpost, and I'm betting that the hoops you have to jump through to get LR to use Outpost AI would mean that the AI doesn't recognize your new "LR" TFs as combat troops and tries to send them to colonize magnates...

[visage]

I've done a few more tests to refine some info and thus have some notes and corrections about the TF AI:

IF TFs seem to launch their missiles at targets within their DF ranges.

I'm not seeing any indication that the ship designations in the TF have anything to do with TF behavior... when I change the ship designations without changing ship stats, nothing changes. When I change ship stats without changing ship designations, behavior changes. When I change TF designation without changing ship designation, behavior changes.

Ships in IF TFs seem to fire their missiles at targets within their DF ranges regardless of whether they're IF or LR ships. Ships in LR TFs seem to fire all their weapon at anything they can regardless of whether the ship is IF or LR.

I'm also not seeing Recon automatically moving without orders in Control mode. ...

[Blaze]

Originally posted by visage
<snip>
I'm also not seeing Recon automatically moving without orders in Control mode. ...
That's odd, because I specifically remember seeing this. What's more, I remember others also reporting that they use recon TF's for the purpose of firing their missiles first, and then closing in for DF weapons automatically. It may be possible that this behavior only displays when attacking a planet, I need to test that when I get a chance. I assume that everything before this last line i quoted was refering to watched battles? Could you also clarify your above statements? I'm not sure if I'm understanding you correctly, and i'd like to do so before commenting.

P.S.
By interchanging the tags for ships in the spreadsheets:TaskForceRules.txt and wStrings:wsTaskForce.txt
I was able to get ship behavior completly swaped, and still maintain their placement in TF design. I'm not sure if this even applies to what you might be looking for, but that's just my 2 cents.

P.S.S.
Originally posted by visage
<snip>
Ships in IF TFs seem to fire their missiles at targets within their DF ranges regardless of whether they're IF or LR ships. Ships in LR TFs seem to fire all their weapon at anything they can regardless of whether the ship is IF or LR.
<snip>
Looking over your post more, I think you are refering to controlled combats. I have noticed a distinctive difference in how LR and IF mission ships behave. Here's how I set up my test. I had 2 LR TF's. The first one had just 1 LR mission missile boat w/ only PD missiles w/ 2 volleys. The second TF had an identical ship and also had a second ship. The second ship was an IF mission role, filled w/ PD nukes, and only 1 volley. I defended a planet that had no defenses. Both TF's shot their first 2 vollies. However the IF ship held onto it's single volley, and only fired when the other TF was engaged by fighters. The PD nukes from the IF ship targeted the fighters. Now w/ that said, I'm currently testing for yet another missile launching bug. It's starting to look like if I increase the missile vollies on the IF ship, that it might not be launching at all, or is launching w/ the LR ships. I'm going to be buisy over this next week, so I'm not sure when I'll be able to test this. Also, I do like that tidbit of info about at what range an IF ship will launch, depending on it's DF range. I'll also try to confirm this aswell.

[visage]

Originally posted by Blaze
That's odd, because I specifically remember seeing this. What's more, I remember others also reporting that they use recon TF's for the purpose of firing their missiles first, and then closing in for DF weapons automatically.

Yeah, having heard the same things, I was pretty surprised that my recon TFs just sat there in control mode.

It may be possible that this behavior only displays when attacking a planet

Which behavior? My testing was done by starting a game, building a ship, and sending it off to fight at an enemy homeworld. All the battles I was looking at were planet assaults.

I assume that everything before this last line i quoted was refering to watched battles? Could you also clarify your above statements? I'm not sure if I'm understanding you correctly, and i'd like to do so before commenting.

No, actually, that post was pretty much entirely about Control mode.

I'm not sure what you want me to clarify, but...

Looking at IF TF behavior in control mode:

My default set of tests involved the aforementioned LR ship as the sole ship in the TF (one missile rack with four PD nukes, usually an interceptor, and a very high-ROF long-ranged laser, Mk V sensor). It would autolaunch but not at the instant of detection, perhaps at some range that might have been the DF max range. When told to attack, it charged into close range and circled at close range.

Same situation as above, but with the ship designated IF, I got the identical behavior.

Replace the ship with one that's got an unmodded light mass driver as its sole DF weapon but still has the fighter and more missiles: it autofires at much much shorter range and runs away when told to attack the three hawk defenders. This happens whether the ship is marked LR or IF.

Put the weakened ship in an LR TF and, regardless of whether it's an IF or LR ship, it behaved like my other tests of LR: autolaunch missiles immediately on detection; charge to short range when ordered to attack.

P.S.
By interchanging the tags for ships in the spreadsheets:TaskForceRules.txt and wStrings:wsTaskForce.txt
I was able to get ship behavior completly swaped, and still maintain their placement in TF design. I'm not sure if this even applies to what you might be looking for, but that's just my 2 cents.

Sure, but that doesn't look like it'll help:
1) I could switch an IF or CV AI over to LR, but a beam-heavy ship is liable to try to close to point blank range anyways. I guess it might be an improvement in that the AI might very well try to maintain distance against a TF that's got heavier DF weaponry.
2) The AI behavior I actually want is in the colony and outpost AI. But just imagine what carnage it'll wreak upon the rest of the game's AI if we're lying to it about which TFs are outposts and which are LR...

[Blaze]

I did some further testing, and I did make a mistake. Recon TF's don't automatically move towards targets in controlled battles. I'll have to fix that in the guide. I've also noticed that IF ships will not fire unless there is some kind of sensor in the TF, even if you have a clear view of incomming missiles from many squares away. I did another test, an IF TF, w/ an IF ship and a sensor ship. on the IF ship, i put on missiles and 1 spinal hard beam w/ a range of 8500 Gm. the IF ship didn't fire until just about point blank range (the same range if there was no DF weapon) about how many squares is 4250 Gm? I don't really know. It's possible that the point blank range is the sensor limit (I used the first sensor tech), I'll try modding the sensor range if that's possible. I am curious to take a look at your savegames, and the mods you use.
Also, I've confirmed another part of the missile launch system bug. Apparently in a LR TF w/ LR and IF ship, if the IF ships have more than one volley of missiles, I don't believe they fire at all, or don't after the LR ships fire all of their vollies. It's possible that the IF ships are firing w/ the LR ships, but I doubt that. single volley IF ships seem to hold onto their missiles until an enemy target is w/ some kind of closer range (so far seems to be point blank range, but needs further testing on sensor and DF weapon affects). I also need to test how much damage is being done by the autofired missiles w/ IF and LR, vs. only LR ships.

[visage]

Originally posted by Blaze
1 spinal hard beam w/ a range of 8500 Gm. the IF ship didn't fire until just about point blank range (the same range if there was no DF weapon) about how many squares is 4250 Gm?

That was a hardbeam modded to have shorter range?

Squares seem to be around 3750 or 4k WRUs (Weapon Range Units).

I am curious to take a look at your savegames, and the mods you use.

I really haven't been using mods for these tests, apart from modding the files to get specific stats from which to perform the test. The TechTables I was modding to generate the situation described was the one from my autobuild mod.

It's possible that the IF ships are firing w/ the LR ships, but I doubt that.

Devise a test to distinguish between the two case?

[Blaze]

my last sentance was the test. I'll elaborate more. A LR TF w/ 2 ships. one LR w/ missiles (2 vollies), and one recon w/ only sensors. look at the damage displayed when the missiles hit. then go back to an earlier save game, and make the same TF, but add an IF ship w/ 2 vollies of missiles, and compare the damage displayed when the missiles hit. Of course a few runs will need to be done. The PD mount reset on me for the hard beam. I made a new design w/ spinal and range of 17000 Gm. I've run 3 tests, and still not seeing the 1/2 DF range firing. I'll try w/ a longer range in my next tests.

[Timushack]

First I'd like to thank Visage for a very informative post. Great job of reverse engineering there! Kudos

I'm putting together some spreadsheets that will produce normalized damage per space at various ranges. It will compute average base damage and apply all accuracy, multi-fire, drop-off calculations, space, etc. Each spreadsheet will analyze different combinations of modifications and/or mounts, with regard to range effects.

I want to try maximizing damage per space at ranges that just exceed alien LR and see if I can get my ships to stop just outside of enemy range. Does anybody know the dimensions of a grid square?

I noticed the mob file has a column named SizeFudge. I didn't see any mention of it in the postings. Did this field come from the new patch? (I'm new to the game) or is it simply not used?

The next task will probably be to do the same thing for damage amount (so it's penetration values can be weighed). Don't have a static ip but I'll be happy to email the spreadsheets when I'm done with them.

[visage]

Originally posted by Timushack
Does anybody know the dimensions of a grid square?

My best guess is that a square is approx 3750-4000 Weapon Range Units.

Drive speeds use a different scale; I seem to remember finding a square to be approx 15k Drive Range Units.

I noticed the mob file has a column named SizeFudge. I didn't see any mention of it in the postings. Did this field come from the new patch? (I'm new to the game) or is it simply not used?

I believe it is unused. I've not actually tried changing it radically and looking for something different happening. There are quite a few fields that seem to be vestigial.

[Timushack]

Ok, I've got over 20 spreadsheets dealing with most combinations of mods over 8 adjustable distance bands (didn't do AP or Env yet). I'm getting some interesting results. Continous seems to be far better than I orignally thought at mid to long range, but you all probably knew that. :-)

I want to repeat this process for each mount type, at least if it is non-linear, and that depends on how the factors are applied. If, for instance, the Distance multiplier for a spinal mount of 2 is applied directly to all distance fields in the mob file, then damage dropoff and max accurracy would not occur until a 9000 range. If this is the way it works, different mounts would not cause linear changes to all weapon efficiency at a given range. Has anybody collected some empirical data on this? I'm assuming all mount factors are applied to base stats before any calculations occur. :)

[visage]

Originally posted by Timushack
If, for instance, the Distance multiplier for a spinal mount of 2 is applied directly to all distance fields in the mob file, then damage dropoff and max accurracy would not occur until a 9000 range. If this is the way it works, different mounts would not cause linear changes to all weapon efficiency at a given range. Has anybody collected some empirical data on this?

Yup. No dropoff occurs before 9k in that case.

[visage]

Finally had the opportunity to do some LAN testing today.

Fighter lasers modded to always do 5 damage and never miss. Fighter fusion to the same, but AP of 0.5. Fighter neutron modded to same as lasers, but MultFire *=5.

Firing on a target with medium duranium armor (6 deflection): laser interceptors did no damage (no surprise there), fighter fusion did 5 damage (whoa! AP affects deflection), fighter neutron did 0 damage (no surprise there).


Looking that the AP/Deflection result, I started to wonder if AP affected chances to hit armor (since I hadn't actually tested that since the patch). Modded a DF laser to do always do 2 damage, but with an AP of 0.1. It took about 20 shots before I hit armor, it was doing 2 damage per shot before that. (20 is a bit of a stretch, makes me want to test this more thoroughly, especially since that armor-hitting shot might not have actually been an armor-hit; the sound seemed right, but I didn't see a yellow 2).

[visage]

Ok, I think I now know how AP works. I think it's pretty wacky, but:

AP is the probability that you do any of the armor step of damage resolution. If you've got an AP value of 0.7, there's a 30 percent chance you ignore deflection entirely and pass by armor points directly to internals.

Did some more tests, the clincher of which was a 3-damage 0.5-AP weapon firing on 15-deflection armor. Half the time, it showed 3 red. The other half the time, it didn't do any damage. No yellows.

[whyBish]

Wow! That changes the way A.P. looks, Makes it a whole lot better on smaller mounts (due to not needing to fight deflection) and against smaller targets (due to not having to take out all the armor). Perhaps I will need to mod A.P. back in ;)

[Ron_Lugge]

Originally posted by visage
Ok, I think I now know how AP works. I think it's pretty wacky, but:

AP is the probability that you do any of the armor step of damage resolution. If you've got an AP value of 0.7, there's a 30 percent chance you ignore deflection entirely and pass by armor points directly to internals.

Did some more tests, the clincher of which was a 3-damage 0.5-AP weapon firing on 15-deflection armor. Half the time, it showed 3 red. The other half the time, it didn't do any damage. No yellows.

OK, I'm going to go cross-post this in OS.

[Phezzan]

Excellent AP test.

Anyone know if Planetary beam bases choose mounts the same way SR & LR do? - Do they choose mods as well or ignore them entirely?

I think my Ext Range mod FuBared planet defense (either that or PDef is sorely lacking)

Thanks
Phez

[Minor Annoyance]

Planetary beam bases always use ultra spinal mount even if you haven't researched them. I don't know about mods.

[Phezzan]

I ask because I'm using a mod - and I have rearranged my mount order.

if it just uses the bottom mount - then I need to mod the beam bases or add an unattainable Ult Spinal below my Standard Mount.

Phez

[Blaze]

ok, I went w/ a different test, my original testing should be good enough. I had in a LR TF, on LR ship, one IF ship, and a recon ship. The LR ship had 2 volleys, and the IF ship had one. The recon ship also had my longest DF spinal mount too, and a fighter. I attacked a planet and controlled the attack. the LR ships autofired their 2 vollies, and the IF ships held onto the one volley until an incomming missile volley, or fighters came in point blank range. This to me is fairly good proof that IF/Carrier ships behave differently than LR ships. It might be that the IF ships appeared to act like LR ships, because they had DF weapons? In addition, gateway103 has done some extensive testing, and found that IF ships fire at max DF range. He tested w/ and w/o spoters. visage, just so we can confirm this, did you use spotters at all? How did you do your testing? GW doesn't like posting here, so atm I'm a go between.

P.S.
I have a theory that in watched battles, ships will still have the same basic behavior as in controlled battles, but the AI can give commands at almost instantaneous speed, and also has acess to controls that were exacto knifed during the great swiss cheesing of moo3. Just a theory though. I say this because I don't see any instances where ships don't fire in watched battles vs controlled battles.

[visage]

Originally posted by Blaze
ok, I went w/ a different test, my original testing should be good enough. I had in a LR TF, on LR ship, one IF ship, and a recon ship. The LR ship had 2 volleys, and the IF ship had one. The recon ship also had my longest DF spinal mount too, and a fighter. I attacked a planet and controlled the attack. the LR ships autofired their 2 vollies, and the IF ships held onto the one volley until an incomming missile volley, or fighters came in point blank range. This to me is fairly good proof that IF/Carrier ships behave differently than LR ships. It might be that the IF ships appeared to act like LR ships, because they had DF weapons? In addition, gateway103 has done some extensive testing, and found that IF ships fire at max DF range.

I just attempted to repeat the above results and failed.

First test: three-ship LR TF. All ships had over-ranged (17k) single Spinal lasers and light lasers. The LR ship also had one PD nuke, no reloads. The Recon ship had one PD nuke, three reloads. The IF ship had 4 PD nukes, no reloads.

Sending against an enemy homeworld where there were three hawk defenses. Result: upon gaining visual detection, which was outside of the range of the lasers, all six missiles were launched. The three Recon reloads followed as expected.

Second Test: removed all the DF weaponry from the IF ship, but otherwise identical. Results: identical.

Third Test: I gave the IF ship a single light MD (removing all the lasers). Otherwise identical. Visual contact was made at long range, but within the laser range. Results: identical.

So, no matter what I do, I'm not getting the behavior you describe of IF ships holding their fire when they're in an LR TF.

Every test I've performed has findicated that TF designation determines the behavior of the TF and that the ship designations inside it are irrelevant.

[Blaze]

I'll try to confirm your results, but note one thing, you should try replicating my test. Secondly, it might be possible that the # of visual missiles might be off (just a theory). Here's the basic design of the LR TF I used.

LR: 2 missile volleys, engine speed 1.
Recon:spinal mount of longest beam weapon, basic sensors, 1 laser interceptor.
IF: 1 volley of missiles, engine speed 1.

Though the engine speed shouldn't matter, I include it for completeness. I have also noticed that in MP games, my IF missiles in this config get lost, but that is probably due to the infamus sync bug, and you also counted the missiles.

[visage]

Originally posted by Blaze
I'll try to confirm your results, but note one thing, you should try replicating my test. Secondly, it might be possible that the # of visual missiles might be off (just a theory). Here's the basic design of the LR TF I used.

LR: 2 missile volleys, engine speed 1.
Recon:spinal mount of longest beam weapon, basic sensors, 1 laser interceptor.
IF: 1 volley of missiles, engine speed 1.

Though the engine speed shouldn't matter, I include it for completeness. I have also noticed that in MP games, my IF missiles in this config get lost, but that is probably due to the infamus sync bug, and you also counted the missiles.

Done. Two missiles in the first volley, not one. Only two missile volleys fired. Repeated the experiment; after the second volley I tried to get it to flush any volleys it would be hoarding according to your theory that the IF ship hadn't fired: gave it direct attack orders, but it didn't have any missile volleys left to fire. The IF ship had fired in the initial volley.

Btw, why use that configuration...? it seems that if you want to show that IF holds its fire you want to have IF with two missile loads and the LR with one. That way you should theoretically get three volleys total under your theory.

[Blaze]

ok, so far i've confrimed all of your tests. I originally used 1 missile volley on the IF ship, because to the missle fire counting bug. I know distinctly that I saw the IF ships holding onto their missiles in one of my tests, I have a few hunches as to why, I'll do some testing. So far, I'd like to say I really appreaciate the effort you are doing, and have cought me on a few mistakes.

[zanzibar]

here (http://www.ina-community.com/forums/showthread.php?s=&threadid=305313) is how ECM and ECCM work. Completely tested and proven :)

[Ron_Lugge]

Originally posted by zanzibar
here (http://www.ina-community.com/forums/showthread.php?s=&threadid=305313) is how ECM and ECCM work. Completely tested and proven :)

Uh-huh. And sensors/cloaks fit in HOW? And where's your proof?

:mad:60:mad:mad:second:mad:limit:mad:

[zanzibar]

Originally posted by Ron_Lugge
Uh-huh. And sensors/cloaks fit in HOW? And where's your proof?

:mad:60:mad:mad:second:mad:limit:mad:

My proof is, play the game for yourself, and observe the results. And I haven't played around with sensors/cloaks.

[visage]

Originally posted by zanzibar
My proof is, play the game for yourself, and observe the results. And I haven't played around with sensors/cloaks.

That's hardly a reasonable presentation of "completely tested and proven."

[Ron_Lugge]

Originally posted by visage
That's hardly a reasonable presentation of "completely tested and proven."

Please engage in a proper demonstration of "Tested and proven" in the ECM threads that are popping up. PLEASE.

[Blaze]

well I can't replicate any of my inital results on ships behavior. either moo3 went buggy (all of my tests were done in the same instance of moo3), I just didn't realize how I had designed my ships, or my notes were just flat out wrong. Well anyways, I've updated my guide to reflect visages testing results. Thanks visage.
P.S. Visage, have you done any further testing on whether IF TF's fire at max DF range, or 1/2 max DF range?

[visage]

Originally posted by Blaze
P.S. Visage, have you done any further testing on whether IF TF's fire at max DF range, or 1/2 max DF range?

Nope, I've not really returned to this... been instead working on the next edition of my autobuild mod (the major change being that I get the AI to select point defense better).

If I try to test that, I'll post it here; similarly, let me know if you figure out the answer.

Oh, on a related note: I was trying out how well it'd work to give LR TFs Outpost combat ai, and so far it looks good. Unfortunately, I don't think that's possible from a modding standpoint. Now, if we convinced QS to slip the change into 1.2.4... :)

[zanzibar]

Originally posted by visage
That's hardly a reasonable presentation of "completely tested and proven."

Oh, so 50 MP games, with using ECM on my ships is hardly proof now?? Whatever... :rolleyes: I give up... *goes back to playing moo 3*

[Blaze]

Originally posted by zanzibar
Oh, so 50 MP games, with using ECM on my ships is hardly proof now?? Whatever... :rolleyes: I give up... *goes back to playing moo 3*
I think I know why you are getting the results you post here. Correct me if I'm wrong btw. In mp games w/ somewhat slow connections, there is a sync lag bug. This will make missiles/fighters follow targets for long periods of time, before attacking. This will also cause DF weapons to not fire for long periods of time, followed by short periods of time where firing will commence. Tell me, have done much testing in SP?

[Ron_Lugge]

Originally posted by zanzibar
Oh, so 50 MP games, with using ECM on my ships is hardly proof now?? Whatever... :rolleyes: I give up... *goes back to playing moo 3*

You are providing qualitive, isntead of quantative values. Thats the problem.

[visage]

Originally posted by Blaze
I think I know why you are getting the results you post here. Correct me if I'm wrong btw. In mp games w/ somewhat slow connections, there is a sync lag bug. This will make missiles/fighters follow targets for long periods of time, before attacking. This will also cause DF weapons to not fire for long periods of time, followed by short periods of time where firing will commence. Tell me, have done much testing in SP?

Not to mention "completely tested and proven" with no actual evidence given. :)

"I always do X, and see behavior Y" does not mean that Y is caused by X.

Explaining what attempts you've made to demonstrate that Y and X are related are handy to actually convince people you're doing something other than blowing smoke.

[Timushack]

I just finished the first cut of my Moo3 weapon analyzer. Its mostly a Visual basic application embedded into a Excel spreadsheet as a macro. Most of the weapon formula's came from this thread so I thought I'd post it here first. You can see the code and tweak as you like, or just use it as an applicaiton.

You can plug in any set of any combination of Moo3 DF weapons, mounts, and mods, set range "from-to", and increment on chart, set targeted ship armor and shield mods. The press the button and get an instant spreedsheet and line chart plotting the results!

This will give you visibity of all weapon effectiveness vs space, min or max damage, accuracy. Only thing I need is place to upload it. Got any suggestions?

[BlackHat]

Originally posted by Timushack
I just finished the first cut of my Moo3 weapon analyzer. Its mostly a Visual basic application embedded into a Excel spreadsheet as a macro.

<SNIP>

Only thing I need is place to upload it. Got any suggestions?

Moo3mods.com might be a good place, although technicaly it is not a mod.

Maybe someone else will volenteer to host it. How big is it?

Later,
BlackHat

[Timushack]

Its 275kb if I remove all the spreadsheets. They can be regenerated by the program.

[Timushack]

The MOO3 Weapon Analyzer is now posted on moo3mods.com.

Please give it a try. I hope I did your research justice, Visage. I'd also like to thank all the other contributers here who made this possible. I'm pretty sure PD still needs a tweak (not well defined since patch) but I think the rest is accurately reflecting your formulas. Let me know if you see any bugs.

[BlackHat]

Originally posted by Timushack
The MOO3 Weapon Analyzer is now posted on moo3mods.com.

Please give it a try. I hope I did your research justice, Visage. I'd also like to thank all the other contributers here who made this possible. I'm pretty sure PD still needs a tweak (not well defined since patch) but I think the rest is accurately reflecting your formulas. Let me know if you see any bugs.

Got a repeatable crash and a suggestion.

So I run it, do a CNTL-Z to start the macro, add some weapons, do the analysis.

At this point with the Weapons list macro still on the screen (with the analysis page still in the background) try and remove one of the weapons from list. CRASH. I get some sort of error. (I dont remember what it is off hand, something about cant link to the object or something.)

[JUST A MINUTE]

Ok I get a RUN TIME ERROR. (Object invoked has disconnected from its clients)

To avoid this bug I have to Close the Weapon Macro window, switch back to the Introduction page then do CNTL-Z again to bring the Macro window back up. Now I can delete weapons.


Suggestion (actualy 3)

1) I really dont wnat to have to bounce back and forth out of the program to a separate table to figgure out what tech level the various Armors, shields, and thickness are. (ADD them to the MOUNT page beside the tech.)

2) On the Enemy Defences tab on the Macro, make another optional window so something that says, "TECH LEVEL" and has a scroller from 1 to 50. Thus I can enter the Tech Level of the Enemy and the Macro can select the suitable values for Armor, thickness, shields, shield size)

3) Like 2 above but for the Weapons select page. This way I can enter MY tech level and Mount size, Then the Macro can select all the weapons and upgrades I would know and put them in the list. Again this means I dont have to keep bouncing out of the program to another list to see what tech level each weapon and/or its upgrades are.

ie. I enter tech 15 and Spinal Mount. The Macro puts:
1. Laser w/M1, M2, AP, AF, Imp, Cont
2. Mass Driver w/M1, M2, AP, AF, Imp
3. Fusion Beam w/M1, M2, Cont, Env
<SNIP>
8. Neutron Beam w/No Upgrades

Later,
BlackHat

[Timushack]

I played with it for a while but I havn't been able to reproduce the crash. I suspect that the save spreadsheet may have gotten corrupt somehow, or maybe there is a version conflict, or maybe I'm just not doing the same thing. I built it using Excel 97. Microsoft is very intolerant about binding spreadsheet cells to controls. Just making a typo in the Rowsource property will crash excel to desktop. Try deleting the "SAVE" sheet if it continues to give you problems. The sheet will be rebuilt then next time you hit ctrl-z although your weapon configs will be gone.

Incidentally you can modify the save spreadsheet manually, but be forwarned there is a "number of rows" field in cell M1 that contains the last row used for the weapons list. I often change the mount by going directly to the spreadsheet. Just make sure they are spelled exactly the same way as the listbox shows them or they won't be recognized.

As for the tech level, I never had the need for it since I was analyzing specific weapons (the ones I had researched) versus the techs the enemy had. It would be an easy change to add the tech level of the weapon to the list box. The tech level of the mods, however, means storing and accessing the tech tree data. Currently all mod information is gleaned from the "original" spreadsheet.

[Croesus]

Originally posted by Timushack
I just finished the first cut of my Moo3 weapon analyzer. Its mostly a Visual basic application embedded into a Excel spreadsheet as a macro. Most of the weapon formula's came from this thread so I thought I'd post it here first. You can see the code and tweak as you like, or just use it as an applicaiton.

You can plug in any set of any combination of Moo3 DF weapons, mounts, and mods, set range "from-to", and increment on chart, set targeted ship armor and shield mods. The press the button and get an instant spreedsheet and line chart plotting the results!

This will give you visibity of all weapon effectiveness vs space, min or max damage, accuracy. Only thing I need is place to upload it. Got any suggestions?
I LOVE Hosting!!! I am puttingit up on my page too, please. I have proven able to host peoples stuff that I like. Even ask the people that downloaded the patch from my webpage.

[BlackHat]

Originally posted by Timushack
I played with it for a while but I havn't been able to reproduce the crash. I suspect that the save spreadsheet may have gotten corrupt somehow, or maybe there is a version conflict, or maybe I'm just not doing the same thing. I built it using Excel 97. Microsoft is very intolerant about binding spreadsheet cells to controls. Just making a typo in the Rowsource property will crash excel to desktop. Try deleting the "SAVE" sheet if it continues to give you problems. The sheet will be rebuilt then next time you hit ctrl-z although your weapon configs will be gone.

Hmm, well I am using Excel2002. It must be a version conflict, since I thought I explained clearly how to reproduce it.
After Analizing weapons, I cant delete the save tab without closing the macro first. (and at that point why bother, I can just click Instructions instead and relaunch the macro and delete weapons then)

however, If after analyzing I do delete the save tab, when I launch the macro (cntl-z) and select a few weapons, when I try to analyze weapons again, I get a nother RunTime Error. (381) Could not get list property. Invalid property aray index.


As for the tech level, I never had the need for it since I was analyzing specific weapons (the ones I had researched) versus the techs the enemy had. It would be an easy change to add the tech level of the weapon to the list box. The tech level of the mods, however, means storing and accessing the tech tree data. Currently all mod information is gleaned from the "original" spreadsheet.

Hmm, about that upgrade mod data. You should probably update it. For example your Analyser thinks LFG has M1, M2, and Improved.
(It Does NOT) because the blocks for those fields have data in them.

Also, if you check the Typos thread, I already updated this table with the correct upgrades and tech levels of the upgrades (which this table has incorrect)

Thus you would not need to access the tech tree data. The correct data would be in this spreadsheet already. I have already done this (substituting the correct upgrade mod data) on my copy of your program.

Later.
BlackHat

[vorenhutz]

@Timushack,
Nice work. I can stop using my horribly kludgy spreadsheet now :)
Any plans to include missiles / fighters?
I think the options to chart minimum damage and maximum damage are swapped ('max damage' is less than 'min damage'). Also the tooltips for the chart options could explain the assumptions you use to calculate the effects of armor + shields.

[Timushack]

Also, if you check the Typos thread, I already updated this table with the correct upgrades and tech levels of the upgrades (which this table has incorrect)
I thought I grabbed the weapons table right from the mob file. perhaps it is a pre-patch version. If you like, send me your version to timushack@aol.com and I will see if it can be integrated into the next release.

Wish I had Excel 2002 on this box. Maybe I'll take the file to work and try to crash it there. Is anybody else experiencing crashes?

[Timushack]

I think the options to chart minimum damage and maximum damage are swapped ('max damage' is less than 'min damage').
I can't get Max Damage less then min damage. Perhaps it looks that way due to Excel dynamic scaling. Can you give me the weapon configuration that did that? It could be a bug.

As for fighters and missles, I'd like to do fighters but I havn't been able to glean how they work yet. Are they considered PD mounts? What is the fire rate? I think I may be able to incorporate them as another mount type in the form.

-----------------------------------------------------------------------
Also has anybody done testing on the PD mount fire rate? The spreadsheet is doing nothing special with it. I always use light mounts.

[BlackHat]

Originally posted by Timushack
I thought I grabbed the weapons table right from the mob file. perhaps it is a pre-patch version. If you like, send me your version to timushack@aol.com and I will see if it can be integrated into the next release.

Wish I had Excel 2002 on this box. Maybe I'll take the file to work and try to crash it there. Is anybody else experiencing crashes?

No you probably do have the latest version from QSI. THe problem is the data from the upgrade mods collums are incorrrect.

I will send you a copy of the file I have and 2 photos showing the origional incorrect data, and the data I corrected.

Later,
BlackHat

[vorenhutz]

Spinal IPC vs Guardian (Damper, UHA Adamantium armor).
Firstly, I think this weapon shouldn't even be able to scratch the guardian.
Maximum damage per hit shows 0 at all ranges (correctly I think).
Minimum damage looks as if it's not taking armor into account.
(Point blank damage is about 9, long range damage about 4).

[vorenhutz]

Also, I think the range scale should go from the minimum point blank range for all weapons considered, up to the maximum extreme range, and just have the user select the number of points to calculate. At least, this should be the default behaviour (user may want to zoom in).
Other efficiency metrics would be nice, too.

cost efficiency
space*range^n (n=0...2) [this acknowledges the added value of long range weapons getting the first shot in combat]
space-adjusted cost (which I don't fully understand, but is detailed somewhere in this thread)

[BlackHat]

Originally posted by vorenhutz
Spinal IPC vs Guardian (Damper, UHA Adamantium armor).
Firstly, I think this weapon shouldn't even be able to scratch the guardian.
Maximum damage per hit shows 0 at all ranges (correctly I think).
Minimum damage looks as if it's not taking armor into account.
(Point blank damage is about 9, long range damage about 4).

I tested this and got the exact same results.
Spinal IPC vs Damper Small, Adamantium UHA

Max damage chart was Zero at all ranges. Min damage chart was 9.25 at 4500 range.

Later,
BlackHat

[Timushack]

Spinal IPC vs Guardian (Damper, UHA Adamantium armor).
Nice find. There is indeed a bug in the minimum damage calculation. It is not deducting for armor deflection. I have fixed it here but will test it a little more tommorrow before posting the update.

[Minor Annoyance]

Originally posted by Timushack
...As for fighters and missles, I'd like to do fighters but I havn't been able to glean how they work yet. Are they considered PD mounts? What is the fire rate? I think I may be able to incorporate them as another mount type in the form.
...

fighterweapons.txt

[Ron_Lugge]

http://www.ataricommunity.com/forums/showthread.php?s=&postid=4084564#post4084564

Should be the linked to post, or the one right after it. It seems to show how missile accuracy works, and I'm wondering if anyone here is familiar with it (E. G. confirm/deny/add to)

[DavidByron]

I'm not sure if stickied threads are subject to the 120 day purge but just in case... since no one has posted on this one for about 110 days... and since it's about the most vital and least reproducible thread on the board...

"bump"

[Croesus]

I second the motion to bump this thread. Now I'll go back to lurking. :lurk: :lurk: :lurk:

[Longspur]

It'd be a shame to lose this valuable thread.

[Archonon]

stay

[Flinx]

:up:

[Incubi]

ooops

[visage]

FYI: I discovered (thanks to people saying "you're wrong, visage!" :) ) that I'd never updated the second post to accommodate the changes in the last code patch... thus while the main post was accurate AFAIK, the second post was saying things like "remember, armor piercing doesn't work"...

I've fixed the errors that I'm aware of in the second post.

[visage]

Visibility finally added. Whee!

[Bhruic]

Sensors and ECCM multiplicatively reduce the OffTgtRg stat. ECM and Cloak multiplicatively increase the DefTgtRg stat. (Values in TechTables which increase OffTgtRg or decrease DefTgtRg are ignored.)

I think you mean OffSptRg and DefSptRg in the bracketed section. Otherwise it makes very little sense. :)

Bh

[visage]

Originally posted by Bhruic
I think you mean OffSptRg and DefSptRg in the bracketed section. Otherwise it makes very little sense. :)

Bh

You said that OffTgtRg>1 is treated as 1, as is DefTgtRg<1, no?

[Bhruic]

Oops, I read that sentence incorrectly.

edit: Although, technically, your statement isn't totally accurate. If a ship had 2 ECCMs onboard, one that did OffTgtRg *= 0.25 and one that did OffTgtRg *= 2.0, then the value that would be passed to combat would be 0.5 - so the 2.0 wouldn't be ignored, because the total value is still less than 1.0. But that's nitpicking. ;)

Bh

[eastwind]

Anyone know where I can get Timushack's weapon analysis spreadsheet? moo3mods.com says "Site Closed".

[Bard of Prey]

Originally posted by eastwind
Anyone know where I can get Timushack's weapon analysis spreadsheet? moo3mods.com says "Site Closed".

Since the death of Moo3Mods.com, the best place to look for any kind of mods (or utilities) is probably www.moo3.at. If the Guardian doesn't have it, then all I can suggest is either contacting the author, or getting somebody who has it to e-mail it to you. (I don't have it btw... :) ).

[blubberbob]

A google search on timushack results in this page:
http://joet.meekers.org/MoO3/MoO3.html
which contains a link to a weapon analyzer spreadsheet

[eastwind]

Thanks! Fwiw I couldn't find it on the guardian.

[moo fan]

quoting from the current (9/06/2004 ) version of the space combat mechanics guide:

"The actual damage done is the Base Damage reduced by a random factor. To determine the random factor, take the smaller of the Base Damage and the damage delta (DmgDelta) stat. Multiply this by a random number from between 0 and 1. Subtract the result from the Base Damage to determine how much damage is actually dealt.

.....

Mounts change all of the stats mentioned above except for the damage delta stat. Thus, mounts with larger multiples on the damage stats (e.g. the Spinal series) suffer much smaller proportional reduction in damage dealt from the random factor. (Obviously, mass-classed weapons don't benefit from this effect.)"



I did a bit of testing and was surprised to see a different behaviour.

What I was trying to do was finding out whether or not damage delta was crippling beam weapon power at far range and if I should therefore rely on mass weapons instead.

I used modified laser cannons with 50k range, reduced size and cost, near damage 10, far damage 1, delta 9 ( but also other variatons: 20, 1, 19; 4,0,4) and put a good amount of those in two ships. One ship used them in standard mounts, the other in modified light mounts.


When I modified light mounts to have the same range but half the near and far damage (and space) of standard mounts, everything seemed OK:
The ship with standard lasers would do about as much damage at extreme range as the ship having twice as many light mount lasers.


But then I modified light mounts to have 2*damage and the result was unexpected: instead of doing at least twice the damage(due to suffering far less from DmgDelta) of the ship having twice as many standard mounts, the light mount ship's damage was only 10-20 % higher.

When I increased the light mounts' multiplier to x4, the increase was about 50 %.


I repeated the tests several times (and during each test several hundred shots were fired) so I'm quite sure it's not just some statistic anomaly.


EDIT:
D'OH! I just realized, that 10-20% increase at x2 is exactly the outcome to be expected.

The +50 % at x4 doesn't quite fit, but when I did other tests at different ranges, they seemed to stick to the rules OK.


If these tests show anything besides that I've got too much time, then it's that beam weapons suffer quite a lot from DmgDelta at extreme range, even when used in the larger mounts.

[Gen.Killer]

<Edited by Xentax: Someone needs to cut back on the caffeine. And this is a Certified Dr Pepper addict talking, here...>

[psilon avatar]

OK, I read through some posts & I recall reading somewhere that Ptd weapons are x2 RoF. AF upgrade gives x3 Rof, so AF + PtD gives x6 RoF. I read where Visage says PtD doesn't get x2. Does PtD get x2 or not, & does combining with AF really get x6?

[visage]

Originally posted by psilon avatar
OK, I read through some posts & I recall reading somewhere that Ptd weapons are x2 RoF. AF upgrade gives x3 Rof, so AF + PtD gives x6 RoF. I read where Visage says PtD doesn't get x2. Does PtD get x2 or not, & does combining with AF really get x6?


Assuming my memories are correct on this subject (it's not something I've thought about in... a great while):

MultFire is a legal modifier for a weapon mod, but not a legal modifier for a weapon mount. Thus, the original Point Defense Mount was bugged in that it depended upon MultFire to work. At some point I managed to transmit this information to someone in QS, and they revised the PD mount so that it did *not* use MultFire.

I seem to recall that the revised PD mount was generally inferior to the Light mount, though.

In short, assuming I remember everything correctly: AF provides MultFire of 3 correctly (which is not the same thing as x3 ROF); PD mount cannot and no longer attempts to.

[EmP]

Somewhere was disscusion is armor piercing modification good or actualy waste of space. I wrote program that calculates average hits to destroy unshielded target, all I need now is suggestion where to upload it or mail it to someone who will upload it. Oh yes, it calculated that 25% AP is in best case (ultra heavy adamantium on leviathan, armor : hp = 256 : 46) is only 35% more effective than those with 0% AP. Edit: file uploaded at http://www.freewebs.com/mooii/ap.cpp

[Tampa_Gamer]

I am using Strawberry Mod 1.1, which contains the visibility patch, autobuild patch, etc. However, it appears that the autobuild prefers Light mounts instead of PD. Is the consensus that the benefits for the Light mount (dmg, range, etc.) outweigh those of the PD mount (2x fire) since they both use the same space? I had that was the consensus last year, but just wanted to confirm.

[Moridin00]

Somewhere was disscusion is armor piercing modification good or actualy waste of space. I wrote program that calculates average hits to destroy unshielded target, all I need now is suggestion where to upload it or mail it to someone who will upload it. Oh yes, it calculated that 25% AP is in best case (ultra heavy adamantium on leviathan, armor : hp = 256 : 46) is only 35% more effective than those with 0% AP.

Edit: file uploaded at http://www.freewebs.com/mooii/ap.cpp
Well thats lovely, but basically all ships have shields on them (esp. leviathans) and you'll need to factor in the damage-reduction of shields too.

Note that you also get shield-piercing weapons. :downcast:

[EmP]

Well thats lovely, but basically all ships have shields on them (esp. leviathans) and you'll need to factor in the damage-reduction of shields too.

Note that you also get shield-piercing weapons. :downcast:

My program doesn't count alot of other factors (like randomized damage and deflection treshold) in order to simplify sitation. Simulating real situation is very complex and it would take multiple repetitions of same case to gain some average information. I'll make that kind of simulator when I'll have enough time.

[solops]

Somewhere, long ago and far, far away I read a discussion about the combat bonus'/penalties for having or not having ships present in each of the task force rings. These penalties are defined in the TaskForceRules.txt file.

Will someone please direct me to or reproduce here that discussion and data?

[shevlon]

is there like uh a translation of the combat mechanics into english???