Analysis of RFCs by defenders set to Ignore Losses

[Curtis Lemay]

ORIGINAL: golden delicious

EDIT: just finished reading the thread

I expect that the combination of no terrain effects and no combat odds effects tended to cancel each other enough so that no one could easily spot them during general play.

This. My experience of TOAW in the past is that combat results feel right overall, but the detail of complaints about ragged remnants holding on when in fortified status rings true, and there's no denying Bob's test results.

Sounds like a clear cut case.

Thanks, Ben.

[Oberst_Klink]

ORIGINAL: Curtis Lemay

ORIGINAL: Oberst_Klink

The model was changed *to my knowledge* becasue of the Ant-Attack bug...

No. That was a different issue. The reason was due to a player making tests and discovering that terrain didn't affect RFC chances. See this thread:

tm.asp?m=2078034

Thanks for pointing it out Bob. I should have rephrased the answer before posting

Klink, Oberst

[golden delicious]

ORIGINAL: Curtis Lemay

Thanks, Ben.

I thought you'd like it. Now, make sea supply roads work again.

[Veers]

You're the man Bob.

[Curtis Lemay]

ORIGINAL: golden delicious

I thought you'd like it. Now, make sea supply roads work again.

Everything works the same under Old Supply Rules.

[golden delicious]

ORIGINAL: Curtis Lemay

ORIGINAL: golden delicious

I thought you'd like it. Now, make sea supply roads work again.

Everything works the same under Old Supply Rules.

Marvellous.

[sealclubber]

Is there any expectation as to when this will be fixed, if ever?

[golden delicious]

There has been no definitive published comment on progress of development.

[sealclubber]

Okay. I guess I'll be the first to say it - if Matrix games fixes this bug I'll buy TOAW 3.

-- Hollow incentive. Turns out I bought TOAW 3 in 2008 and totally forgot about it and never installed it!

[Menschenfresser]

I've read through this and other threads about the RFC-Combat Odds issue presented and I was hoping to get some practical advice.

If I understand this, as presented by Bob, we're only talking about factoring in the ratio of equipment? So, when changed, if all goes as planned, combat, where the equipment ratio is near 3:1 in favor of the attacker, should still work roughly as it does now.

Something I don't remember comments about were Bob's suggestions for future standards (3:1 normal, 1.5:1 half, 6:1 double, or something to that effect). Wouldn't this further increase defender bonus for a lot of combat where attack-defender are closer to equal? Or am I misinterpreting the proposed fix?

Until such a glorious day, are there suggestions on mitigating the problem? Should one consider using a lower loss setting when smaller sized units (say regimental/division in a corps-based scenario) are left to defend a hex by themselves?

[Curtis Lemay]

ORIGINAL: Menschenfresser

I've read through this and other threads about the RFC-Combat Odds issue presented and I was hoping to get some practical advice.

If I understand this, as presented by Bob, we're only talking about factoring in the ratio of equipment? So, when changed, if all goes as planned, combat, where the equipment ratio is near 3:1 in favor of the attacker, should still work roughly as it does now.

Something I don't remember comments about were Bob's suggestions for future standards (3:1 normal, 1.5:1 half, 6:1 double, or something to that effect). Wouldn't this further increase defender bonus for a lot of combat where attack-defender are closer to equal? Or am I misinterpreting the proposed fix?

I was thinking of "normal" meaning a straight (unmodified) morale check for RFC. So, if you had effective 3:1 assault ratio (ground only, no support) odds after the combat round you would have the defender's unmodified morale percent chance of forcing an RFC, etc (as in 3.2 and earlier). This, of course, would be modulated by terrain/deployment effects (so all infantry vs. infantry in a fortified hex would require 24:1 odds to be "normal"), etc. If you had an effective 6:1 odds, his morale would be halved, etc. If you only had 1.5:1 odds, his morale would be doubled, etc.

That would be the default. Designers would have a game parameter to scale the effect one way or the other (so WWI scenarios could be made much harder to gain ground, etc.)

Until such a glorious day, are there suggestions on mitigating the problem? Should one consider using a lower loss setting when smaller sized units (say regimental/division in a corps-based scenario) are left to defend a hex by themselves?

House rules to that effect, or similar, have been suggested.

[secadegas]

ORIGINAL: Curtis Lemay

I was thinking of "normal" meaning a straight (unmodified) morale check for RFC. So, if you had effective 3:1 assault ratio (ground only, no support) odds after the combat round you would have the defender's unmodified morale percent chance of forcing an RFC, etc (as in 3.2 and earlier). This, of course, would be modulated by terrain/deployment effects (so all infantry vs. infantry in a fortified hex would require 24:1 odds to be "normal"), etc. If you had an effective 6:1 odds, his morale would be halved, etc. If you only had 1.5:1 odds, his morale would be doubled, etc.

That would be the default. Designers would have a game parameter to scale the effect one way or the other (so WWI scenarios could be made much harder to gain ground, etc.)

If this can be put into effect - and although i hear many things lately, i'm here long enough to trust you only - i'd say this would be "the beginning of a beautiful friendship"...

[Menschenfresser]

Thanks for the input.

Here's to the future...

[governato]

I have completed by test on RFC probabilities comparing the standard TOAW version with the new patch, created by user kmitahj.

This new version has an increased probability of RFC, as most players felt that the 3.4 upgrade was too biased in favor of the defender. It also includes the previous 'AA patch'. Before I show the test results, here is how I have set the test scenario. This is very similar to Bob's test described in this thread. Bob was very kind to send me and Oberst Klink a copy. I added to it some artillery for both sides, as I felt that it was closer to a real situation for most scenarios.

As in the original test, there are 13 terrains being tested and three deployments, for a total of 16 tests. The MRPB is 1, so no combat can last more than one round. Each test has 20 trials – for a total of 320 combats. Each combat consists of one attacking unit vs. one defending unit. The attacker is always 9000 infantry squads and 768 122mm
Howitzers. The defender is initially 3000 infantry squads and 256 122mm Howtizers. (x3 odds). Initially the attacker as 100 proficiency and the defender 70. The Attrition divider is 10. Both sides have 100 readiness and supply. The scenario is then run 20 times for a total of 200 trials for each test and a total of 16x10x20 combats. After the combats complete, I total how many defenders RFCd in each test. Simple enough. The goal is to make sure that the new patch behaves as expected.

The first set of tests used the standard TOAW 3.4 version, this is our reference and the results are shown in the image below. Results are consistent with expectations, with the caveat that there is still a bit of statistical noise.

[governato]

The patch followed the simple approach of modifying (tuning down) values of existing T/D factors used inside RFC check procedure. The corresponding Deployment and Terrain factors are lowered by the same, small value. Here are the chances for RFC choosen for this particular patch implementation (in brackets are numbers used in original 3.4 and 3.2 versions):

Fortified deployment: 65 (84,50)
"Fortified line" terrain: 65 (84, -)
Entrenched deployment: 50 (65, 33)
dense urban, badlands terrains: 50 (65, -)
mountainous terrain: 50 (50, -)
Dunes, Urban, Bocage terrains: 33 (50, -)
Defensive deployment: 20 (26, 20)
Forrest, Jungle, Hills, Wadi terrains: 20 (26, -)

The tests reveal that RFCs outcomes increased by about 20% overall with the introduction of the patch, however the increase is more significant for the more difficult terrain (with a DF of x4 or X8, as Fortfied Line, Dense Urban or Fortfified deployment). In this case the number of RFCs events increased by more than 50%.


We agree with Bob that changes should be incremental, so this is probably good enough, even if this patch does not directly solve the problem of making the chance of RFC depend on the combat ratio, which would be optimal. kmitahj commented that a more complex change was beyond what was possible (remember he is editing the executable!).


We hope to make the patch available in the next few days, but in the meantime comments are welcome.



Btw, for those not up to date kmitahj also recently released the AA patch, which is described
and obtainable here:

tm.asp?m=3376085

[Sensei.Tokugawa]

Great input into the community well-being, thanks a lot governato. I am assuming that there's no patch attached yet? I have been using the AAA fix and need to admit it is working albeit with some randomization of the results I would say.

[Oberst_Klink]

ORIGINAL: burroughs

Great input into the community well-being, thanks a lot governato. I am assuming that there's no patch attached yet? I have been using the AAA fix and need to admit it is working albeit with some randomization of the results I would say.

There is, but I wait for the 'official' statement of our Kamerad in Seattle It also worked like ace during my various test of my latest beta-version of Kharkov '43.

Klink, Oberst

[governato]

Oberst! I only ran the tests, 'kmitahj' is the one who made the patch so he should be the one to officially release it.

[Sensei.Tokugawa]

Herr Oberst, If You need a commie tester for Kharkov '43, I am up for that, I have been obsessed with the spring offensive for some reason for quite some time already.

[Oberst_Klink]

Ja, ja, I know. Kapitan Kloss is the creator, Fabio You've done some extraordinary work and the tests so far are positive, nay?

Klink, Oberst