AGEOD Forums

I wonder what is the factor that determines a battle to come to an end.

In SCW im testing combats and the results are battles ending with the loser side units al with 0 of Cohesion.
I dont want hte losing side being so damage in cohesion, i would prefer they retire sooner from combat.

That's a tough one!

Battle can cease either because there is a rout, or a voluntary retreat. In the second case, it is a combination of feeling overwhelmed and some luck (as you roll some percentile dice). In the first case, you rout if cohesion drops a lot and you fail morale checks at various level.
Oh, and there is also autoretreat... After some losses, you will automatically succeed in your retreat roll.

So I would say that leader quality, max and current cohesion of troops and discipline are very important here. Plus you can adjust autoretreat variables.

Pocus wrote:Plus you can adjust autoretreat variables.

Thats interesting. How could i?

Leibstandarte wrote:Thats interesting. How could i?

Look in \Settings\GameLogic.opt

refer to:
http://www.ageod.net/agewiki/Retreat_Chance_formula_and_modding

Read the Fascinating Wiki (RFW)

lodilefty wrote:Look in \Settings\GameLogic.opt

refer to:
http://www.ageod.net/agewiki/Retreat_Chance_formula_and_modding

ok, lodilefty. I suppose you are wondering why the hell i dont read DEFINITELY the wiki

i promisse im going to read it carefully.

more questions:
i have this in RUS settings "cbtCohBonusrAutoRetCoeff= 50 " putting a bigger number tha 50, like 70 give more oportunities to retreat beacuse loss of cohesion. For me is difficult sometimes to understand the FWIKI. (UFW)

Since I'm just looking into that as well, I'm going to necromance this topic!

COHESION-AUTO-RETREAT-COEFFICIENT

Leibstandarte, I can only guess as well. But judging from the formula that I could find in the GameLogic file (autoretreat% = BasePercent * (Coh + bonus) / 100), this coefficient is used

1) to make troops with high fighting power (high cohesion) sustain battles longer. If the cohesion of a force is high, auto-retreat will be triggered when more casualties have been suffered. E.g. if you set the coefficient to 50 and the "cbtBasePercentAutoRet" to 15, a force (?) with very high (average?) cohesion - let's take 100% average cohesion for testing purposes - will auto-retreat when it suffers 0,15* (100+50)/100= 0,225. I.e. it will autoretreat with 22,5% casualties.
2) to make troops with low fighting power (low cohesion) auto-retreat earlier. E.g. if you set the coefficient to 50 and the "cbtBasePercentAutoRet" to 15, a force with very low cohesion - let's take 0% average cohesion for testing purposes - will auto-retreat when it suffers 0,15* (0+50)/100= 0,075. I.e. it will autoretreat with 7,5% casualties.

You can easily deduct the consequences of changes to this factor. Right now it is calibrated around 50% cohesion. The value that you've set in "cbtBasePercentAutoRetIf" applies if a force has 50% average cohesion. If the average cohesion is lower or higher than 50%, units will autoretreat with more (if high cohesion) or fewer casualties (if low cohesion). Increasing the value should make forces tendentially auto-retreat earlier/with fewer casualties than you've set in cbtBasePercentAutoRet. Decreasing it should make forces auto-retreat later/with more casualties than you've set in cbtBasePercentAutoRet.

(Well in fact all this is not really affected by the value of the coefficient, but by the overall formula used. The coefficient only determines at what cohesion status the "cbtBasePercentAutoRet" applies)

However, we don't really know how auto-retreat works, how (average?) cohesion is determined, and how losses are determined. I guess that if the auto-retreat is triggered (only in between combat rounds, or immediatly once the % casualties for auto-retreat are reached?), the C-in-C automatically wants to retreat (regardless of comparison of combat power?). Then the individual stacks might still have to do withdraw-attempt-tests.

BETTER WAYS TO END COMBATS FASTER/MAKE ENGAGEMENTS LESS DECISIVE

Personally, I'd take a look at the command postures and rules of engagement-settings if I wanted to make forces more likely to retreat. Here are the actual settings:


// Offensive Posture - First ROE (All-out attack)
roeRetreatWill0 = 000|000|080|090|100|100
roeRetreatChance0 = 000|000|060|065|070|075

// Offensive Posture - Second ROE (Standard attack)
roeRetreatWill1 = 100|100|100|100|100|100
roeRetreatChance1 = 100|150|150|150|150|150

// Offensive Posture - Third ROE (Conservative attack)
roeRetreatWill2 = 100|100|150|150|150|150
roeRetreatChance2 = 100|150|150|150|150|150

// Offensive Posture - Fourth ROE (Feint attack)
roeRetreatWill3 = 100|200|200|300|300|300
roeRetreatChance3 = 100|150|150|150|150|150

// Defensive Posture - First ROE (Defend at all cost)
roeRetreatWill4 = 000|000|000|000|000|000
roeRetreatChance4 = 000|000|000|000|000|000

// Defensive Posture - Second ROE (Standard defense)
roeRetreatWill5 = 100|100|100|100|100|100
roeRetreatChance5 = 100|150|150|150|150|150

// Defensive Posture - Third ROE (Defend & Retreat)
roeRetreatWill6 = 100|100|150|150|150|150
roeRetreatChance6 = 100|150|150|150|150|150

// Defensive Posture - Fourth ROE (Retreat if engaged)
roeRetreatWill7 = 200|200|200|300|300|300
roeRetreatChance7 = 100|150|150|150|150|150


Prior to each of the six combat rounds per day (hence 6 values per entry: round 1l2l3l4l5l6), the C-in-C (the highest ranking officer of a faction in the region) will decide (based on his, and only his own combat posture/roe) if his side (all stacks involved in the combat) WANT to retreat. The exact formula is obscure, but obviously a comparison of combat power (own troops versus hostile troops) is made, plus the above mentioned modificators for combat postures/roe are applied. For example, a leader with "retreat if engaged" will have his have his base retreat-will (based on the power-comparison) doubled (200%) prior to round 1, and even tripled prior to round 4. Plus the retreat chances (of individual stacks (?) who try to retreat after the C-in-C has decided for his faction to withdraw) are also increased (150% if trying to retreat prior to round 2 or the following rounds). I've made quite a lot of changes here myself in order to make it less risky to split up forces and work with smaller detachements. I've drastically increased retreat wills and retreat chances prior to round 1 (before actual battle, that is) for certain combat postures, for example. A single hussar unit will be anihilated if it had to fight a full round against a big stack.

For some more thoughts (and/or guesses ) about combat-ending-mechanics, take a look here as well: http://www.ageod-forum.com/showthread.php?21275-wip-Collection-of-information-on-retreats-routs-and-pursuits&p=205852&viewfull=1#post205852

-------------------------------------------------------------------------------------------

PS: This extract from a battle log might tell us something on how the base-retrat chance is calculated:


17:08:11 (Reporting) Checking combat stance for Prussia in region 424 Chemnitz
17:08:11 (Reporting) InCS %: 100 New Retreat Will %: 0
17:08:11 (Reporting) Ammo %: 100 New Retreat Will %: 0
17:08:11 (Reporting) Weighted average of Trench levels: 2
17:08:11 (Reporting) Power of opponents compared to us: 1922.62 % Base Retreat Will: 0.00 %
17:08:11 (Reporting) Dice rolled: 14
17:08:11 (Reporting) => We want to stay in battle
17:08:11 (Reporting) Checking combat stance for Austria in region 424 Chemnitz
17:08:11 (Reporting) Preliminary test: No need to retreat from this battle: OppPower: 84 FacPower: 1714 .


Obviously command-penalties (lack of command points and/or being part of an army structure in general?), ammunition-supply-level, entrenchment level (1 or 2 in ROP), and the power-comparison are factored into that calculation. In this case the Austrians were storming Chemnitz, thus the Prussians had a retreat will of 0% even though the power of the Austrians was 1922.62% of their own power. The prussians were trapped in the city and therefore automatically had the "defend at all costs"-command posture.

Another example:

17:16:56 (Reporting) Checking combat stance for Prussia in region 351 Glatz
17:16:56 (Reporting) Preliminary test: No need to retreat from this battle: OppPower: 775 FacPower: 1061 .
17:16:56 (Reporting) Checking combat stance for Austria in region 351 Glatz
17:16:56 (Reporting) InCS %: 0 New Retreat Will %: 100
17:16:56 (Reporting) Ammo %: 100 New Retreat Will %: 100
17:16:56 (Reporting) Weighted average of Trench levels: 0
17:16:56 (Reporting) Power of opponents compared to us: 136.90 % Base Retreat Will: 37.00 %
17:16:56 (Reporting) Dice rolled: 44
17:16:56 (Reporting) => We want to stay in battle

Here we can see that the retreat will of a faction seems to be 0% if its own combat power (hopefully only the units which are actually taking part in the combat are considered here!?) is stronger then the opponents' power. Each % that the opposing forces are superior, however, translates 1:1 to % of retreat will (Power of opponents compared to us: 136.90 % (775 vs. 1061) Base Retreat Will: 37.00 %). This would then be modified by the command-posture, I guess? In this case, it might have been 100% prior to round 1.