Mikeydz
-
Posts
375 -
Joined
-
Last visited
Posts posted by Mikeydz
-
-
*sniff*sniff*.... I'm so happy.... gimme a tissue...
-
Got the schedule, but mail to you Treeburst is bouncing...
The following addresses had permanent fatal errors
<mikeman@cablelynx.com>
(reason: 553 5.1.8 <mikeydz@houston.rr.com>... Domain of sender address mikeydz@houston.rr.com does not exist)
Transcript of session follows
... while talking to mail.cablelynx.com.
>>> MAIL From:<mikeydz@houston.rr.com>
<<< 553 5.1.8 <mikeydz@houston.rr.com>... Domain of sender address mikeydz@houston.rr.com does not exist
501 5.6.0 Data format error
-
The problem I think is that the AI can't account for the fact that you have 9 guns vs 1, which more than likely would result in a dead Stug. It's more than likely resolving 9 individual 1 on 1 scenario.
The real problem with this type of complaint is that unless you provide some type of save-game file, it's impossible to dig in and say if there is a flaw in your setup, or if there is a bug. If you read the Redwolf Tank thread from a couple of months back you'll see how important that file is.
-
Forgot to address this in my last post since this has been bugging me ever since this phrase appeared. You refer to the Tac-AI modeling screw-ups. Concerning the ISU-PzIV matchup, you refer to the retreating action as the screw-up in question. But as been pointed out, from a vechile survivability standpoint, a retreat (wether before, after, or while firing, seems to be the best choice It rarely if ever got knocked out when it chose to retreat, IIRC. You may be of the opinion that retreating is the wrong choice, but that doesn't make it a screw-up. In fact, I doubt that there is any code in the Tac-AI that intentionally causes a unit to perform an action that it knows will result in it's death.Originally posted by Mikeydz:</font><blockquote>quote:</font><hr />Originally posted by redwolf:
In my opinion, the game should model "very soft" factors like crew screwups for everybody or nobody.
Steve will need to verify this, but I would be shocked if it would be programed to do that. We player get our units in enough trouble without the AI lending a helping hand.
So I belive the characterization that the Tac-AI "models screwups" is in error.[/QB]</font>
-
Forgot to address this in my last post since this has been bugging me ever since this phrase appeared. You refer to the Tac-AI modeling screw-ups. Concerning the ISU-PzIV matchup, you refer to the retreating action as the screw-up in question. But as been pointed out, from a vechile survivability standpoint, a retreat (wether before, after, or while firing, seems to be the best choice It rarely if ever got knocked out when it chose to retreat, IIRC. You may be of the opinion that retreating is the wrong choice, but that doesn't make it a screw-up. In fact, I doubt that there is any code in the Tac-AI that intentionally causes a unit to perform an action that it knows will result in it's death.Originally posted by redwolf:In my opinion, the game should model "very soft" factors like crew screwups for everybody or nobody. [/QB]
Steve will need to verify this, but I would be shocked if it would be programed to do that. We player get our units in enough trouble without the AI lending a helping hand.
So I belive the characterization that the Tac-AI is in error.
-
Actually, I did realize this. In my opinion the Tac-AI has decided it really likes this match-up on the part of the PzIV. It also might be deciding that the current location (behind the tall pines) is better than any other spot that it could reach quickly. I just took a real cursory glace again at your save file and didn't see any other positions close by that would protect the PzIV better. So with this in mind, the Tac-AI looks like it's deciding that a good match-up combined with a decent position rules out a need for a retreat.Originally posted by redwolf:Your description of the four states of ISU behaviour is correct, but you fail to realize that the Mk IV only has one. In all the tests I ran I never had the Mk IV driver take off while the gunner was trying to target.[/QB]
BTW, I just created a test where the PzIV does exhibit the the retreat action. It's pretty rough, so I don't have any rough %s of how often wit retreats, but let's just say our seemingly "fearless" PzIVs don't like SU-85s or SU-100s.
-
I think one of the problems with Redwolf is evident in the quotes above.. Why does the Tac-AI "always" have to wait for the 1st shot. Time and time again, Redwolf makes these broad claims. It's almost like he lives in a black and white world where somethihg should always happen, or should never happen. But CM is awash in grey. It's been pretty well established IMHO, that the ISU isn't going to fare well on average in his scenario file, but in CM, there are 4 different observed results. Retreat with no fire, fire then retreat, fire while retreating, and stand and fight. And with the random factors, you as a player really don't know quite what's going to happen. Now redwolf and BFC can debate what the "optimal" tactic would be all day long, but if I had to choose between Redwolfs set-in-stone programed response vs. the Tac-Ai variable response, I'll choose the Tac-AI every time.No. The TacAI should always wait for the first shot to come off and then start reversing, if at all. That is in the best interest of the player.Again, prove it. And even if you are correct (which you are not), what about ammo management?
-
You haven't even seen the scenario yet, have you?</font>Originally posted by wwb_99:</font><blockquote>quote:</font><hr />Originally posted by redwolf:
</font><blockquote>quote:</font><hr />Originally posted by wwb_99:
Offhand, I would say MikeyDZ's tests are the only ones that hold water. He ran it thru enough to get a somewhat signficant figure. Aside from that all the tests are immaterial.
-
Time and experience will be the best bet for getting a hang of things.Even on a small map, so much can happen in that one minute turn that you think you missed something. But what's been said so far is all good advice. Keep at it.
-
Having taken a look at this from many angles..I would have to say no.Originally posted by The_Capt:</font><blockquote>quote:</font><hr />Originally posted by redwolf:
In CMBB, is the ISU-122, all aspects specific to this vehicle and all general armor combat factors combined, modeled so that it should better not take on a Pz IV lang from 500-600m?
Yes or no?
</font>
-
Originally posted by Battlefront.com:
Redwolf,
</font><blockquote>quote:</font><hr /> I don't see why this simple question requires a 100 line answer:
So, Redwolf... c'mon... can't you just admit that you didn't know squat about the ISU-122 before you started this thread and were arguing in great part out of ignorance? You thought it was some sort of über Tank Destroyer, which didn't have anything to fear from a puny "mid war" German piece of crap, right?
Steve</font>
-
Yes, the way the ISU is modeled in CMBB, it should not take on a PzIV at 500-600m.Originally posted by redwolf:I don't see why this simple question requires a 100 line answer:
In CMBB, is the ISU-122, all aspects specific to this vehicle and all general armor combat factors combined, modeled so that it should better not take on a Pz IV lang from 500-600m?
Yes or no?
-
:redwolf:
Just to point out that way back on the first page (near the bottom), you said... "It is totally obvious that the ISU should not retreat in this situation. In my first movie it didn't get a shot off at all, in the second it shot while retreating, getting lower hit probablity from being on the move." So your original gripe was that it retreated period, not that it didn't fire at least1 shot before retreating. I'll grant that you have modified your position on what the "bug" is.Folks, the bug is retreat before fire. Please do not post appearences of retreat for reload as examples of inappropriate retreats. I think we all agree that the 122mm armed Russians can rightly retreat for reload.
I'll try and hold back the tears from you outright rejecting my test data. My test wasn't intended to directly address the issue of should the ISU fire off a round before retreating, but to address your contention (which you still seem to hold) that the ISU is the superior tank under this set of circumstances. My tests (which prevented any chance of a retreat for the ISU) was only meant to show that at 500m, head to head, the ISU is not even close to superior. I hoped that if you saw the abysmal performance head to head against the PzIV, you would realize that the issue that you should be arguing against isn't that the Tac-AI is buggy, but that you believe the ISU is modeled incorrectly. (of which I also disagree with you on, but that's another issue).As for the test case with 102 tests, I am afraid I can only accept runs of my test scenario as sample data on how likely it is that the ISU or the MK IV are killed.
That is because it is different. Since only you an I have seen my test scenario, I'll explain it for all the other followers of this thread that might be interested. I set up the ISUs to the rear and sides completely by water. That way the could not retreat behind any type of cover. I also left out the elevation from your save gave, so there was not hull down position. I had to remove the hulldown position to remove the ability to retreat behind the elevated terrain causing the hulldown bonus. The only other terrain I placed was a row of tall pine directly in front of the PzIVs and a row of scattered trees about halfway between the 2 tanks, to lower the "hit probs" to about where they were in redwolfs save file. So with no options to retreat, it forces the ISU to slug it out. And in most cases, if the ISU didn't get a 1st shot kill, the superior ROF from the PzIV wins out. In otherword, the ISU died ALOT.1) the isolated test never shots (shows?) the ISU retreat before firing, so the situation *is* different somehow...3) don't forget all are hulldown there
Steve never claimed that your scenario was an incredible set of circumstances which caused this fluke retreating action. In my opinon, he's argued that in any situation that the ISU meets the PzIV at 500m, if there is a avenue of retreat available to it, the Tac-AI would probably favor that option, though not to the total exclusion of fighting it out once in a while. And even in your save file, you didn't have a favorable hit prob with the ISU. Your pictures (page 4, bottom 1/5 of the page) give a hit prob of 29% for the PzIV vs 27% for the ISU. A virtual dead heat. The ISU does enjoy a better kill chance though, assuming it lives long enough to get a hit.3) Steve keeps claiming this very situation in my scenario is so special that it places the ISU in some magically difficult situation, despite favourable kill and hit probablity display.
The tall pines are the same as the trees right in front of the PzIv in your save file. The hit proababilities are comparable in my test file to your save file. My test hit probs averages are a shade lower, but the hit prob figures are a rough guide. Now if I left out the trees and allowed the higher hit probabilities, then it would help the ISU a bit more in the head to head fight because if would get a few more 1st shot hits (and probable kills) to make up for it's lack of ROF in comparrison to the PzIV. But once again, my goal was to simulate your confrontation as close as possible, while removing any escape routes for the ISU.4) the patch of woods in front of the Mk IV makes the hit probablity low, so effectively the hit probablity is the same as in a much longer range duel. If you just use the same range without woods you get higher hit probability. If you use longer range, you might not see the TacAI magic ability to realize of optimal engagement range[ December 02, 2002, 08:50 PM: Message edited by: Mikeydz ]
-
In my tests (assuming there is no huge fundamental flaw in my setup) the ISU gets killed at nearly an 80% rate. I don't know what variables the Tac-AI used to figure out if it wanted to retreat or not, but if it computes anything close to an 80% death rate, that it would order any vehicle to retreat. Sure, this matchup isn't quite a Stuart vs Tiger fight in CMBO, but it doesn't look good to me. Perhaps most other match-ups produce (pulling a number from thin air) something between a 40-60% rate, which the Tac-AI decides is a good gamble, so it doesn't order a retreat for either vehicle, so you get a slug-fest. This is all conjecture of course, since I don't know how the Tac-AI makes the decisions it make.Originally posted by Walpurgis Night:I don't see this as hitting the core of what doesn't *feel* correct about the ISU tendency to retreat so quickly. Let's just say these tanks are roughly even, or even favor the IVH a bit as Steve suggests. Fine. The problem is that the ISU retreats *feel* disproportionate to how other tank models in CMBB behave when faced by tanks of roughly equal in strength. Stupidly or not, they tend to stand and fight more.[/QB]
I think the heart of the matter of this whole thread is this. There is no "bug" in the Tac-AI in this case.The Tac-AI "knows" that at 500m the odds are not good for the ISU, based on the model BTS has implemented for this vehicle. So the only possible flaw would be in the model of the performance or combat effectiveness of the ISU at 500m. What redwolf needs to do to "win" this debate is then convince BTS using historical data that the ISU was alot better at hanging in a close range duel that the current model supposes. Like I said before, I'm no grog, so I don't have a good feel for this, so I leave it up to the "experts" to figure out those details.
-
Just finished a 104 sample test of a Pz IVH vs ISU-122 one-on-one battle at 500m, with terrain conditions to simulate similar hit probablilities from redwolf's test file. (approx 25-30%). Terrain desinged on the ISU-122 side to prevent possible avenues of retreat.Originally posted by redwolf:The Mk IVs are the weaker AFV here, and they stay, and the stronger AFV flees before the first shot, in about 1/3rd of the cases.
[/QB]
Overall results were 82 dead ISU-122 vs 23 dead PzIVH. (Note 1 battle resulted in a simultanius kill, so numbers don't add to 104)
I'm not an armor grog, so I don't know if theses results are representative with that would happen in real life if you put these 2 tanks up head to head 104 times, but it's obvious why the TacAI is retreating the ISU-122. It get's spanked.
If anyone wanted a copy of my test scenario, to check my set-up, let me know.
-
-
I tried it when CMBO first came out and gave up after 20 minutes. SO much thanks should go to him for providing the database.Originally posted by Kanonier Reichmann:Chris H. must have spent literally hundreds of hours to compile all that information.
Regards
Jim R.
-
http://users.erols.com/chare/cm/
Look at near the bottom of the page under the Combat Mission: Barbarossa to Berlin heading for CM2 XLS Tables.
Full stats on all the infantry, armor units, aircraft, ect in the 3 Excel files in that zip file. Print those bad boys and you'll be good to go.
-
Just for grins I ran a shot seris of tests. (Not my normal exhaustive tests, but hell, it's a holiday). German vs. Soviet, June 41. Both ai-choose at medium quality. 1000 pt. 5 Runs
----------Total---C--G--R--V
Run 1
German--34-----0--2--30-2
Soviet----41-----4--33-4--0
Run 2
German--37-----0--5--30-2
Soviet----41-----7--30-4--0
Run 3
German--30-----0--3--25-2
Soviet----35-----0--7--23-5
Run 4
German--35-----0--4--27-4
Soviet----42-----6--31--5--0
Run 5
German--27-----0--5--21-1
Soviet----39-----0--4--33-2
Total = Total # units
C = # conscript
G = # green
R = # regular
V = # veteran
What looks like happens is when you set a medium quality, the computer chooses to either buy a "green" based force or a "regular" based force. This is randomly chosen based on nationality and date. In 41, it looks like a coin flip on if thier troops are going to be mostly green or regular in a certain battle. Germans look to to normally be a Regular force in 41 (5 runs is to small to rule out a Green force for germany in 41, so I'm sure it happens, just much rarer.)
The experience modifier Michael Emrys mentioned i s on pg 219 of the manual, 1st paragraph. My guess on what happens once the cpu chooses a Green or Regular force is that everything it buys is Green or Regular, and applies said modifier to generate the numbers I got in my runs.
[ November 28, 2002, 01:29 PM: Message edited by: Mikeydz ]
-
Hi easytargetOriginally posted by Strontium Dog:</font><blockquote>quote:</font><hr />Originally posted by easytarget:
strange that i can get transparent buildings and transparent smoke
just no fog
I don't mean to insult your intelligence but you do have the weather effects switched on don't you? Shift + 'w'
Just thought I'd check
</font> -
Have you tried having her play the demo? Could solve all your problems...Originally posted by Brodieman:G'day guys, i've been playing the demos non-stop for a month now and have ordered my bundle version of the games (damn 4 weeks delivery curse Australia's geography!) My question is, Should i tell my wife to pack her bags now? or should i sneak the games behind her back and continue the current forbidden trist? Exactly how addictive are these games?? more than the demo? If so my marriage is in serious trouble!
Brodieman
-
As far as I know, yep. they see it in all it's glory. The ATI fog issue is a long known problem, but unless ATI releases a driver that supports this type of fog coding, there's nothing BF can do other than hacking changing thier fog programing, which isn't gonna happen until the engine rewrite.Originally posted by easytarget:interesting,
thx for the info - i looked in the setup for the vid card, there's definitely no setting for turning on a fog table - sigh
so, do nvidia card users see the fog?
-
CMBO is an awsome game in it's own right. The graphics are a little bit dated, but with all the different mods available for it, that isn't really an issue. Just note the mods only improve the in game graphics, sound or interface appearance. They do not replicate any of the improvements or tweaks made to the feature list added in CMBB. The best starting point that I know of off the bat is www.combatmission.com . There are quite a few other sites worthy of checking out for CM mods, but that one is the best overall starting point.Originally posted by Fire-at-Will:Right then, still not sure which one to go for. I'd prefer the Western Front as opposed to the Eastern Front but does the improvement in graphical quality of CMBB and the other enhancements make it that superior or is CMBO still a great game in its own right even if it looks a little dated (is this the case??)
Also I've noticed that there seems to be plenty of user made add-ons for CMBO, do these enhance the game so much as to make it comparable to CMBB. Is there a list anywhere of definitive "must have" files for CMBO. Can anyone help me out further here ??
Regards to all.
-
ftp://ftp.battlefront.com/pub/demos/cmbo/cmbodemo102.exe
http://www.battlefront.com/products/cmbb/demo.html
Best way to make you decicion is to try the demos of eith CMBO, CMBB, or both, and see if you have an interest in the game (YOU WILL) Then either buy only the one you are interested in, or buy the bundle pack and save some green. (about $17 if you factor buying them seperatly w/shipping costs in the US)
ROW III - The Boots & Tracks CMBB Tourney (Part 2)
in Combat Mission: Barbarossa to Berlin
Posted
To drunk to do any turns right now, you (hic) guys will just have (hic) to wait....