Modding retreats

[Jagger]

Pocus, are we able to mod attempt to retreat values?

I know that retreat is determined by a ratio comparison between the attackers and defenders. The percentage difference between the larger force and smaller force is used to determine the percentage chance the smaller force will retreat. So if the attacker is 200% larger than the defender, the defender will attempt to retreat 100% of the time before battle occurs. If the attacker is 150% larger, then the defender will attempt to retreat before battle 50% of the time.

What factors are used to determine the ratio difference between the two forces-number of elements or power? Are major defensive/offensive modifiers such as level of entrenchments and crossing major rivers taken into account? Is cohesion a factor considered in determining attempt to retreat? Are the quality of troops and leadership considered? Or is the percentage a straight forward numerical comparison of elements?

The reason I ask is because I just lost Nashville in my PBEM with Kilcavalry. My defense was 3+ divisions (pwr 1200-1300) entrenched to level 5 with good defensive commanders. Grant's attack (pwr 2200+) was directly across a major river, the Cumberland, with the 25% negative offensive modifier for crossing a river. The Union had 96 elements to the CSA 53 elements. The total numerical odds based on elements were 1.8 to 1. All mobile CSA formations attempted to retreat and did retreat. The only CSA troops that fought were the fixed militia which couldn't retreat.

But if the major river crossing modifier was considered, then Grant's force should have lost 25% of its advantage. The entrenchments are another 25% advantage in combat. Those are significant strength adjustments reducing the simple 1.8-1 numerical element advantage if battle had occurred.

I looked at the battle report to see what odds were calculated to determine the retreat percentage. Unfortunately the battle report was truncated and did not provide information on this specific battle.

Regardless I assume the retreat odds were fairly high because everyone retreated that could retreat.

To be honest, I expected the Nashville force to defend Nashville. Considering the power of level 5 entrenchments and defense of a major river, I think they should have fought and considering battle dynamics, very possibly won.

Even if they had lost, my guess is they would have inflicted significant damage on Grants force which would have required time for recovery and slowed further advance. By retreating, Grant's force is in excellent condition and able to continue advancing immediately.

And my force has abandoned one of the best possible defensive positions available in the game without a fight-and is very poorly positioned without a comparable defensive position available. Retreating from an absolutely premier defensive position at a strategic location without even a fight was disasterous.

I have seen three other situations in which I was surprised at retreat in my current PBEM-Stonewall at Harper's Ferry, Longstreet at Manassas and Kilcavaly's force at Alexandria. In each case, they were outnumbered but had outstanding defensive advantages-except Alexandria. In my two situations, I expected and wanted a defense but ended up with retreat instead.

And now I look at my defense of Springfield, Mo which has level 5 entrenchments but a power of only 235. Then I look at the Union force at Rolla with minimum power of 480. That is an advantage of over 2-1 odds in straight forward power. Normally I would hope that Union power advantage would be counter-balanced by level 5 entrenchments. Now I am wondering seriously. Will my Springfield defense simply retreat, just as at Nashville, if the Rolla force advances with 2-1 advantage in power/numbers. I suspect they will.

Are any defensive/offensive modifiers considered when calculating the attempt to retreat percentage before battle? Can we modify any of those factors?

Are we able to modify the ratio difference used to calculate the attempt to retreat?

Thanks in advance!

[Guru80]

I understand how you feel Jagger. I have had the exact same thing happen to me (retreating that is) but the difference in my situation was I had numerical and strength advantages yet I gave up Manasses with entrenchment and 2 to 1 odds (in my favor) yet Jackson comes up in poor weather and my entire force retreated to Harper's Ferry. The only shot fired was at my retreating soldiers backs.

At Bowling Green, entrenched, supiorior numbers, againt he only shot fired was at my soldiers running from the city.

I am the US however and in this particular PBEM game I am just getting crushed. I guess it is the luck of the draw but I at least thought I would put of a fight, not get mowed down running from the field ;-)

[Jagger]

I understand how you feel Jagger. I have had the exact same thing happen to me (retreating that is) but the difference in my situation was I had numerical and strength advantages yet I gave up Manasses with entrenchment and 2 to 1 odds (in my favor) yet Jackson comes up in poor weather and my entire force retreated to Harper's Ferry. The only shot fired was at my retreating soldiers backs.

At Bowling Green, entrenched, supiorior numbers, againt he only shot fired was at my soldiers running from the city.

I am the US however and in this particular PBEM game I am just getting crushed. I guess it is the luck of the draw but I at least thought I would put of a fight, not get mowed down running from the field ;-)

Actually I am surprised your troops retreated before battle if they significantly outnumbered the enemy. That suggests that the ratio comparison is more than a simple comparison of strength.

I am really curious what factors are taken into account.

[GShock]

I am really curious what factors are taken into account.

I had already pointed out there's no masking in this game...you always know everything possible within the fog of war. You speak of retreats but ... are you sure your enemy would always march in plain sight?

What about reserves, what about decoys? Some sort of unit masking would be very nice, especially considering the retreat issues. I've noticed the "mask" ability but that only helps in entrenchment bonus

[Pocus]

The algorithm can indeed be improved (are the variables available for modding: not for now).

What it takes into account:

a) your power value compared to what the enemy have, in involved stacks only (ie if you have one group in passive and the other in defensive, the defensive stack will take the brunt of the assault and not the passive one)

b) the history of losses from previous battle rounds (don't apply in your case). The more you dishes damages compared to what the opponent did, the less prone to retreat you are whatever the ratio.

c) if the region CiC is entrenched, retreat will is 0.85. This part can be improved by taking the average of the entrench value of all stacks, and apply something like x0.9 for each trench level.

d) river crossing is not taken into account, true. can be improved.

e) ROE. You won't retreat if you have the ROE 'defend to the last' button. But we are not in Napoleon's Campaigns

I know it can be frustrating... I can improve the algorithm, knowing that it is only the ROE which will be precise enough to accomodate all situations.

[Jagger]

The algorithm can indeed be improved (are the variables available for modding: not for now).

What it takes into account:

a) your power value compared to what the enemy have, in involved stacks only (ie if you have one group in passive and the other in defensive, the defensive stack will take the brunt of the assault and not the passive one)

b) the history of losses from previous battle rounds (don't apply in your case). The more you dishes damages compared to what the opponent did, the less prone to retreat you are whatever the ratio.

c) if the region CiC is entrenched, retreat will is 0.85. This part can be improved by taking the average of the entrench value of all stacks, and apply something like x0.9 for each trench level.

d) river crossing is not taken into account, true. can be improved.

e) ROE. You won't retreat if you have the ROE 'defend to the last' button. But we are not in Napoleon's Campaigns

I know it can be frustrating... I can improve the algorithm, knowing that it is only the ROE which will be precise enough to accomodate all situations.

I think using .9 for each entrenchment level would definitely help.

Here is some number crunching using .9.

Entrenchments +10 per level of entrenchments to defense or offense power
-Level 2 entrenchments require 2.2-1 power odds to create 100% attempt to retreat decision by defender. 1.2-1 power odds to create attempt to retreat decision by defender

-Level 4 entrenchments require 2.4-1 power odds to create 100% attempt to retreat decision by defender. 1.4-1 power odds to create attempt to retreat decision by defender

-Level 6 entrenchments require 3.2-1 power odds to create 100% attempt to retreat decision by defender. 2.2-1 power odds to create attempt to retreat decision by defender.

-Level 8 entrenchments require 3.6-1 power odds to create 100% attempt to retreat decision by defender. 2.6-1 power odds to create attempt to retreat decision by defender.

If you used +20 instead of +10

Entrenchments +20 per level of entrenchments to defense or offense power

-Level 2 entrenchments require 2.8-1 power odds to create 100% attempt to retreat decision by defender. 1.8-1

-Level 4 entrenchments require 3.6-1 power odds to create 100% attempt to retreat decision by defender. 2.6-1

-Level 6 entrenchments require 4.4-1 power odds to create 100% attempt to retreat decision by defender. 3.4-1

-Level 8 entrenchments require 5.2-1 power odds to create 100% attempt to retreat decision by defender. 4.2-1

IMO, .10 or maybe .15 would work well.

Here are some experimentation with a negative modifier to attacker power when crossing major and minor rivers.


Terrain modifiers:
Major River Crossing -40% attacker power
-Requires greater than 3.6-1 power odds to create 100% attempt to retreat decision by defender. 2.6-1 power odds to create attempt to retreat decision by defender

Minor River Crossing -20% attacker power
-Requires greater than 2.2-1 power odds to create 100% attempt to retreat decision by defender. 1.2-1 power odds to create attempt to retreat decision by defender

Example 1:

Assuming attacker crossing major river (-40%) against defender in level 5 entrenchments (+50% for 5 entrenchment levels).

From defender side: Defender 1000 power x (1+.5) (level 5 entrenchments)= 1500 adjusted power. 1500 power x 2 for automatic attempt to retreat= 3000 power. Attacker needs 3000 power x 1.6 (40% negative major river crossing)= 4800 power to create automatic attempt.

4.8-1 odds required for automatic attempt to retreat by defender. Defender attempt to retreat begins at 3.8-1 odds.

From attacker side: Defender in level 5 entrenchments=1500 power. Attacker at adjusted power 750 or less will auto attempt to retreat. 750 pwr x 1.6= 1200 power. So an attacker of power 1200 crossing major river against a defender of 1000 power defending in level 5 entrenchments will automatically attempt to retreat. An attacker of power 2200 or less will create an attempt to retreat. So odds of 1.2-1 will result in auto attempt to retreat by the attacker.

Odds of 2.2-1 will create an attempt to retreat by the attacker when attacking across a major river against level 5 entrenched defender.

Example 2:

Assuming attacker crossing major river (-40%) against defender in level 5 entrenchments (+100% for 5 entrenchment level).

From defender side: Defender 1000 power x (1+1) (level 5 entrenchments)= 2000 adjusted power. 2000 power x 2 for automatic attempt to retreat= 4000 power. Attacker needs 4000 power x 1.6 (40% negative major river crossing)= 6400 power to create automatic attempt to retreat.

6.4-1 odds required for automatic attempt to retreat by defender. Defender attempt to retreat begins at 5.4-1 odds.

From attacker side: Defender in level 5 entrenchments=2000 power. Attacker at adjusted power 1000 or less will auto attempt to retreat. 1000 pwr x 1.6= 1600 power. So an attacker of power 1600 crossing major river against a defender of 1000 power defending in level 5 entrenchments will automatically attempt to retreat. An attacker of power 2600 or less will create an attempt to retreat.

So odds of 1.6-1 will result in auto attempt to retreat by the attacker. Odds of 2.6 to one will create an attempt to retreat when attacking across a major river against level 5 entrenched defender.

Example 3:

Assuming attacker against defender in level 5 entrenchments (+50% for 5 entrenchment level)

From defender side: Defender 1000 power x (1+.5) (level 5 entrenchments)= 1500 adjusted power. 1500 power x 2 for automatic attempt to retreat= 3000 power. Attacker needs 3000 power power to create automatic attempt to retreat or 3-1 odds. One division will defend against 3 divisions. One division will always attempt to retreat against 2 divisions or more.

From attacker side: 1000 power defender in level 5 entrenchments= 1500 power. Attacker at adjusted power 1500 or less will auto attempt to retreat. So Attacker will always attempt to retreat at .75-1 power odds or 750 power. Attempts to retreat will begin at 1.5-1 odds or 1500 power. 1.5 division will always attack 1 division in level 5 entrenchments without attempt to retreat. .75 division will always attempt to retreat when attacking 1 division in level 5 entrenchments.

Example 4:

Assuming attacker against defender in level 5 entrenchments (+100% for 5 entrenchment level)

From defender side: Defender 1000 power x (1+1) (level 5 entrenchments)= 2000 adjusted power. 2000 power x 2 for automatic attempt to retreat= 4000 power. Attacker needs 4000 power power to create automatic attempt to retreat or 4-1 odds. One division will defend against 3 divisions. One division will always attempt to retreat against 4 divisions or more.

From attacker side: 1000 power defender in level 5 entrenchments= 2000 power. Attacker at adjusted power 1000 or less will auto attempt to retreat. Attacker will always attempt to retreat at 1-1 power odds. Attempts to retreat will begin at 2-1 odds. 2 divisions will always attack 1 division in level 5 entrenchments without attempt to retreat. 1 division will always attempt to retreat when attacking 1 division in level 5 entrenchments.

Anyway some ideas with some numbers to get an idea of impact.