Soviets post your Horrible loses here

[Stelteck]

The problem is that attrition is not a very fun gameplay mecanism, while combat is.

Also, if combat looses do not matter so much, then as a consequence a lot of the thing you manage in game have no real effect, such 90% of support regiments such as artillery.

[morvael]

Please remember that some of the losses in battle are deferred until the logistics phase. Previously (before 1.09.01, and to an extent before 1.11.00) the game was showing you "virtual" losses, by counting all the men in destroyed elements and half the men in damaged elements as lost, while actually they were not all lost then and there (and even later). Not all men in destroyed elements are killed (some are disabled and/or captured), and not all men in damaged elements are disabled (some elements may be lost later, and then some of their assigned men will become disabled). Now battle reports show only true, immediate losses, while the impact of battle is carried over to the logistics phase, and the losses viewed from turn-to-turn perspective are correct. Some things like damaged MG or rifle squads or panzerfausts are not showing in the LOST/DAM results of a battle at all (because they are not men, not guns, and not AFVs). The number of men that you see LOST are men captured/killed during battle, and DAM men are men disabled during battle (from destroyed elements). Losses from damaged elements will be resolved only in the logistics phase. So to each battle you have to add dozens of men and equipment that will be lost later.

1.11.00 #8 Improved battle report details, it will now show correct number of men in DESTROYED group (it should be equal to LOST men - sum of men actually killed and captured in combat), and DAMAGED group (it should be equal to DAM men - men actually disabled in combat). DISRUPTED group remains unchanged (total number of men in disrupted elements).

[morvael]

ORIGINAL: Stelteck

The problem is that attrition is not a very fun gameplay mecanism, while combat is.

Also, if combat looses do not matter so much, then as a consequence a lot of the thing you manage in game have no real effect, such 90% of support regiments such as artillery.

Combat losses matter a lot, the problem is some of these are resolved some time after combat (call it lingering combat damage).
Support regiments and artillery are super important, because superior firepower results in more destroyed, damaged, and disrupted elements of the enemy in combat. Disrupted elements do not count for the purposes of final CV (which can give you desired result - retreat or hold), and increase fatigue after combat (which will again increase losses later when moving, fighting or resolving logistics phase attrition).

[Stelteck]

ORIGINAL: morvael

Combat losses matter a lot, the problem is some of these are resolved some time after combat (call it lingering combat damage).
Support regiments and artillery are super important, because superior firepower results in more destroyed, damaged, and disrupted elements of the enemy in combat. Disrupted elements do not count for the purposes of final CV (which can give you desired result - retreat or hold), and increase fatigue after combat (which will again increase losses later when moving, fighting or resolving logistics phase attrition).

I would really love this.
But, like a lots of thing in the game, i do not really have data to demonstrate it.
Maybe you have.
The efficiency of each unit and support units is hard to analyze.

I do some limited testing with AT artillery brigate against panzer division of 250 tanks.
Without AT brigade a soviet division will kill/damage maybe 1/2 tanks.
With it, maybe 5/6.
What is the CV reduction. 3% ?

[morvael]

Have you checked how many elements were disrupted in detailed view (unfortunately you can only compare this across all elements)? They caused a temporary drop in CV for the purposes of that combat resolution as well. It could be crucial. The same applies to air support.
IMHO there are a lot of disruptions, CV drops a lot from the possible value (unfortunately again, unless you play in Alt CV mode there is no relation between pre-combat CV and combat CV, so you can't see the drop).

[morvael]

If we assume the ratios estimated by Dupuy (20% KIA, 80% WIA of which 15% are wounded seriously), and map KIA to LOST, while seriously WIA to DAM, you would have to triple numbers shown on battle report to get "real world" numbers (and even that estimate does not take into account that more men will be LOST/DAM in the logistics phase as a result of this combat).

[TheLysander]

[RforRush]

Are there any fully finished human vs human games on v1.11 to compare losses?

[EwaldvonKleist]

A few questions to HLYA and the general audience that came up while catching up with all the posts of the last days/weeks:
1) Are the early war absolute losses generally too low?
2) Are the early war loss ratios right?
3) How differ losses when attacking a regiment vs. when attacking a division?
4) Are the late war absolute losses right?
5) Are the late war ratios right?

This are open questions.
What I believe is that late war German counterattack ratios are not right. I am by no means an expert for WW2 combat, but my impression is that the German Wehrmacht was able to inflict massive losses with favourable ratios in counterattacks against overextended and not entrenched soviet units, at all scales. From bataillion level counterattacks to huge corps counterattacks. But in WitE counterattacking is ususally not a good idea except for morale building, because the ratios are 1:1, except for AFVs. Here an example from Stefs AAR:

[RforRush]

Early war loss ratio is right only for AI vs AI games (only for killed, AI can't make pows). Late probably too. Absolute losses are obviously not right.
Anyway, in my opinion Axis losses are low for the entire game, except maybe 1945. Or, more correctly, germans losses are too low. Others are fine.
I will post stats comparing real 'historical' losses ratio with game's one.

[morvael]

But when you consider that people play with hindsight and no political/psychological restrictions, no "not one step back" orders, how do you think the game can replicate absolute losses? This is impossible. I wish there was a script where the AI would play entire war in historical manner, and loses could be tuned to that scripted game.

[RforRush]

I do not know how to replicate them, I'm talking about losses ratio which is completely wrong and 2 times bigger than 'historical' ratio. This is in March 1944 in my AI vs AI test game, so maybe it will change later.

[EwaldvonKleist]

RforRush, you mean that the ratio is in favour for the Axis, am I correct in my understanding?
The problem is that the overall losses in the war include POWs, but WitE German players don't let Stalingrad pockets happen.
So the ratio is mostly a result of early war Soviet pockets and late war frontal assaults, and the loss ratio of Soviet frontal assaults on German positions should be quite big.

[Stelteck]

If i'am not wrong, more than half of german panzers were out of order at the end of 1941 while in game it is quite difficult to drop panzer numbers from 5000 to 4000.

German lost a lots of tanks while not being surrounded.

In game manpower losses looks not so bad and could agree with morvael analysis. Analysis that is coherent and credible to me.

But i wonder if equipment looses, from guns to tanks are not really underestimated.

[RforRush]

ORIGINAL: EwaldvonKleist
RforRush, you mean that the ratio is in favour for the Axis, am I correct in my understanding?
The problem is that the overall losses in the war include POWs, but WitE German players don't let Stalingrad pockets happen.
So the ratio is mostly a result of early war Soviet pockets and late war frontal assaults, and the loss ratio of Soviet frontal assaults on German positions should be quite big.

Yes, ratio favors Axis greatly. What exactly I mean by ratio is only killed (in battle, from wounds, sick, accidents, missing (but assumed dead)) ratio, without POWs. 'Historical' killed ratio (without POWs died in captivity) is about 1 : 1.63 (using Krivosheev data for USSR and Axis Allies, Rudiger Overmans data for Germany). In the game killed ratio is about 1 : 3 and more.
Permanent casualties 'historical' ratio is assumed between 1:1 - 1:1.4. This can't be replicated nor in player vs player game, nor in AI vs AI. Thus the only ratio we can compare with reality is number killed.

What I noticed in AI vs AI game it's both sides are making a lot failed attacks, but for Axis forces failed attack result in hundreds losses while for Soviets it's usually thousands.
I am not sure we can compare AI vs AI games with reality considering that Soviet AI is making more stupid attacks than Axis (basically Axis making most stupid attacks in 1941 - start of 1942, from 1942-1943 it's constant Soviet AI attacking). We'll see when test game will be finished (AI turn is taking centuries from 1943 and on [:@]).

ORIGINAL: EwaldvonKleist
ratio of Soviet frontal assaults on German positions should be quite big.

Yet Axis frontal attacks don't produce same loss ratio, which is not realistic.

[RKhan]

I'm not sure this is a good or bad example:



Over 6:1 ratio of forces, almost 4:1 odds yet loses of 1:44.

[morvael]

240:4696 is 1:20, and this was a fighting withdrawal under heavy air superiority. I'm sure the loss of 10% of AFV is also bad for the Germans.

Besides, losses from before September'41 do not count at all. Soviets have a special "bonus" for this early phase of war, whey they suffer extra losses.

[SparkleyTits]

I am just looking at that battle myself now Khan!

I almost allowed you to cut off the divisions below that hex with only 1 panzer there so I unfortunately had to move one division back so you could get that battle and push me back there

[RKhan]

240:4696 is 1:20, and this was a fighting withdrawal under heavy air superiority. I'm sure the loss of 10% of AFV is also bad for the Germans.

Besides, losses from before September'41 do not count at all. Soviets have a special "bonus" for this early phase of war, whey they suffer extra losses.

I'm not nearly as vexed by these things as others, so long as the game feels right at a macro level, you can hardly expect everything to be modelled precisely. I'm not even sure there is a consensus on what the model ought to be, or that we can understand something so complex as a major battle in every detail.

R

[RKhan]

I am just looking at that battle myself now Khan!

I almost allowed you to cut off the divisions below that hex with only 1 panzer there so I unfortunately had to move one division back so you could get that battle and push me back there

I'm sure you have noticed by now that I like to cut off the panzer spearheads... [:)]