Matrix Games Forums

Forums  Register  Login  Photo Gallery  Member List  Search  Calendars  FAQ 

My Profile  Inbox  Address Book  My Subscription  My Forums  Log Out

BTR AXIS AI Cheats

 
View related threads: (in this forum | in all forums)

Logged in as: Guest
Users viewing this topic: none
  Printable Version
All Forums >> [Current Games From Matrix.] >> [World War II] >> Gary Grigsby's Eagle Day to Bombing the Reich >> BTR AXIS AI Cheats Page: [1]
Login
Message << Older Topic   Newer Topic >>
BTR AXIS AI Cheats - 8/2/2019 3:00:58 AM   
ReginaldBednar2

 

Posts: 9
Joined: 8/2/2019
Status: offline
I have played 450 of 700 turns for the full campaign as Western Allies. The last date played is November 9, 1944.
The scoring is as follows:
Air Superiority or whatever it really is: 2
Industry Damage: 28
Terror: 25
To Win: 85

My goal is to elucidate the cheats used by the Axis AI, and it will be clear why there are
no counter-measures for any of them!

Item 1
The number of P-51B replacements is grossly undervalued. There where around 8,000 build over 18 built over
18 months, around 14.6 per day. The Pacific theatre apparently got P51C and P-51D models. So the true
The actual replacement rate was around 14 per day, not the game 8 per day. In my game I ran out of P-51Bs, and still was not able to convert P-38s to P-51s.
Apparently most of the P-38s in Europe where converted to P-51s.

Item 2
Weather forecasting in the game is terrible. If weather recon planes are not to be modeled, a simpler estimated probability of good weather for a given target
would be more useful. The weather plays a big part in the strategy of targets selected, since the industry of the Reich
is constantly getting repaired, and if no bombing missions are flown one day the industry damage typically drops by 1 point!

Item 3
The Russian over-running of Eastern Theatre industry is not historical, and lags the actual events. This seems very strange since the Western Theatre offensives
are almost precise duplicates of the historical ones. The key offensive in the one in Poland Silesia, which is the equivalent of the Eastern Theatre Ruhr. It should fall in late January 1945
or early February 1945. In effect the air war is over at that time since this area was so
vital to the Reich. This is true no matter how many super-jets are fielded.

Item 4
When the bombing raids are very successful with lots of Luftwaffe fighter shots down, the games
goes into a super FLAK versus fighter mode, where around 20 P-51s are shot down flying back to England!
This is very embarrassing for such a detailed game. Can you imagine training the 88s to get even 1 maneuverable P-51?
Clearly getting rid of the P-51s is a "play-balance" mechanism.

Item 5
The PK for fighter combat is way too high, especially for the Luftwaffe pilots (training issue will be discussed shortly)
One website has a USAAF set of plots showing from January 1943 to June 1944 the Luftwaffe lost 13,678 planes in combat and
another 9872 planes in "non-combat", perhaps including those shot down in training flights. This totals 23,550 AC.
Currently in the game, 28,126 Luftwaffe fighters have been shot down. This helps tremendously the AI tactic of jumping bombers which
are stragglers, and helps poor pilots get more kills than they otherwise would have. Also, FLAK is way too lethal. I usually
only strafe airbases one or two times a month, because more attackers over-all are shot down
than planes on the ground! The AI is not as likely to stack super-FLAK groups together either
like a player would.

Item 6
For some reason statistically, one type of Allied fighter or bomber is picked on for most
of the losses for a raid (we are talking about 1000 aircraft formations). The Luftwaffe shot down 100 B-17s one turn which were
will escorted, and another time 30 P-51Ds were shot down. Other times 30 P-38s will be shot down. Another time 30 P-47s will
be shot down. The key is the other fighters don't take hardly any losses at all! If the losses were all coming from one
group, that would be a different story. If fact, one less group next time is not a big
loss in overall capability, but taking losses across several groups is devastating.

Item 7
This has been complained about by others. The Allied fighter flies on it own path home,
dives over a site with balloons, and is shot down by FLAK or higher flying Luftwaffe aircraft.
This item does not bother me as much as Luftwaffe aircraft following Allied fighters home
and shooting them down one by one like bombers.

Item 8
The Luftwaffe has a super supply of pilots with good PK. In my game, 10,203 Luftwaffe pilots
were killed, and another 13,075 were wounded (28,126 aircraft lost). I understand in the actual
campaign the Germans ran out of pilots who could survive a few combat missions. So this effect is very
disappointing. The Germans are shooting more fighter down now than at any point in the
campaign with FW-190D9 and ME-109K fighters. Ironically fewer bombers are being shot down, and the
Luftwaffe lets the target be obliterated before attacking.

Item 9
The recon data has two serious flaws. The first is that the types of aircraft on each
airbase is not specified, just an exact count of the total planes. The second is that after 1 or two days
the recon information drops the flak information.

Item 10
Allied fighters and bombers which are damaged, even very badly damaged, have to fly all the
way back to there originating base. Many aircraft are down between an available airfield and the
home airfield. The Luftwaffe is much closer to their home bases, so a lot of their aircraft are
saved, while many Allied aircraft crash. Also, many pilots would presumably not be killed.


Please fill free to commit about this (player or programmer).
Post #: 1
RE: BTR AXIS AI Cheats - 8/2/2019 12:22:49 PM   
Arkham

 

Posts: 85
Joined: 9/7/2006
Status: offline
#10 i agree with. It drives me nuts to see crashes for BC aircraft over england. "Well chaps, we returned to our starting location of Kent, we have 3 engines out and half a tail. Lets fly all the way to Hartlepool!"

(in reply to ReginaldBednar2)
Post #: 2
RE: BTR AXIS AI Cheats - 8/2/2019 3:26:59 PM   
simovitch


Posts: 4492
Joined: 2/14/2006
Status: offline
Which version are you playing? It looks like you are either playing the Jet-Age scenario or pre 1.04 patch. the reason I'm asking is we did fix some things with 1.03 and 1.04 but there is still some work that can be done.
Item 1. P38's will upgrade to P51B's in 44, but you have to have P51B's in the repl pool to do anything. I'll let the data team (i.e. warshipbuilder) chime in on the replacement rate, which is easy to fix.

Item 2. We are working on a way to provide some forcasting ability. I'm thinking of some type of message when a major weather event will occur 3-5 days out, like a "Russian High" or an "Arctic Low" etc. It will take a bit of fancy coding to get it right so it may be a while yet.

Item 3. These event timings are easy to change in the code. Currently there are 2 east front events, 1st in late August '44 (Romania collapses) then in late February '45 (east front collapses). I don't want to impact the balance too much because Allied victory is heavily dependent on the timing of the collapsing fronts.

Item 4. There are no "play balance mechanisms" built into the code that I see, but I want to check if there is any targeting preferences. An 88mm HE shell can spread a lot nasty stuff in an air burst, but I agree that speed of the AC should be a defensive benefit. Flak gets moved around more now so there should be less hidden pockets of AA guns lying about.

Item 5. We can look into this. PK="percent kill"? I'd like to hear what others think about the PK's in the game. I personally think that morale is too high in this game and should be brought down as the war wears on. With 1.04 the AI will move flak around more effectively and will concentrate light AA in airbases more often .

Item 6. We can check into this.

Item 7. Yes, this is currently being looked into.

Item 8. We can check into this.

Item 9. I don't see these as serious flaws. Do you really want to see old information that may be inaccurate? The AI does move stuff around a lot each turn.

Item 10. Maybe we can add a random probability that a "crash" over friendly territory becomes a "forced landing" with a delay returning to service for pilot and aircraft? Have to be for both sides though...


< Message edited by simovitch -- 8/2/2019 4:47:31 PM >


_____________________________

simovitch


(in reply to Arkham)
Post #: 3
RE: BTR AXIS AI Cheats - 8/3/2019 3:34:59 AM   
ReginaldBednar2

 

Posts: 9
Joined: 8/2/2019
Status: offline
Thanks for such a quick response! I think I have version 1.03, with the readme file dated November 26, 2017.

In regard to play balance with item #4, I think this applies for most of the other items as well. That is, they were added to try to get a better play balance. Note that the Axis are not seriously effected by weather (radar), the Luftwaffe production does not affect the Eastern retreat, there is no Allied FLAK, and the Germans don't care (in the game) where Allied aircraft are based.

PK stands for "probability of kill".

In regard to your comments about item #5, the PK computation must be very complicated (morale, experience, maneuverability, climb rate, guns, altitude difference, etc.), so there could be many locations in the code where play balance was manipulated. Its seems to me version 1.03 was where the PK shot up, including FLAK.

In regard to item #9 (recon info going stale), I think old FLAK information and knowing where the best Luftwaffe fighter are (where) for the dangerous base strafing runs would be much better (and realistic). Perhaps this is a play balance killer, due to the huge number of Allied recon planes available. My preference would be to skip plotting recon flights and have some kind of intelligence based on target priorities (jets, FW-190D9s, ME-109K4s, night fighters).

For Item 10, the code has a delay associated with changing aircraft types. This is elegant. Perhaps all that is needed is a delay to get the damaged aircraft back to the home base. In my opinion this problem is so noticeable because FLAK PK is so high at times. Perhaps just adding it to the replacement pool with a delay is even easier.

I have no idea if I have a solid chance of winning the game at this point. The Allies in November 1944 knew the air campaign was an overwhelming success (that is there was still a significant air force left and bombing did do significant damage). The strategic air campaign by itself was not the most important factor in the Allied victory.

Thanks

(in reply to simovitch)
Post #: 4
RE: BTR AXIS AI Cheats - 8/3/2019 4:04:52 AM   
warshipbuilder


Posts: 878
Joined: 2/23/2013
From: C-eh-n-eh-d-eh
Status: offline
Re: Item #2
quote:

if no bombing missions are flown one day the industry damage typically drops by 1 point!


I refer you to page 80 in the manual.

Re: Item #1 Nooooo they did not build 8000 P-51B's, D's yes, B's no. That said, the production rate in the game seems low. At the moment it is 6 per day in 1943 and 8 per day in 1944.I need to find out exactly how many the USAAF in the ETO actually got their hands on before I will a change. If somebody out there knows that,PLEASE LET ME KNOW! Also cite your source so I can verify the information.



< Message edited by warshipbuilder -- 8/3/2019 5:17:14 AM >


_____________________________

warshipbuilder

Any ship can be a minesweeper, once.
Editor WitE2Pedia
ED/BTR Ressurection Project

(in reply to Arkham)
Post #: 5
RE: BTR AXIS AI Cheats - 8/5/2019 3:03:52 AM   
ReginaldBednar2

 

Posts: 9
Joined: 8/2/2019
Status: offline
Yes, I stand corrected. In the game the P-51C is also the P-51B.
There were 2000 P51Bs built and 1750 P51Cs, of which the British got
636 for the Mustang IIIs. This means 3114 P-51B/Cs should have gone to Europe.

I looked in my game version, and the Allied player gets 1326 P51B replacements, while
the British get 726 Mustang III replacements.
This means the game is short 1788 replacements for P-51Bs.
This is more than enough to keep up with losses.
Currently I had to take many P-51B units and convert them to Mustang IIIs or IVs, since
I am completely out of replacements. As you mentioned, the game is also short on P-51D
replacements. I read in a book that most of the P-38 squadrons in Europe where converted to
P-51Ds. The book "Air Force Combat Units of World War II", edited by Maurer Maurer,
shows when each Fighter Group converted from P-38s to P-51s. It is very tedious to look
this up, but it may be of help as well.

Thanks for catching my mistake.

(in reply to warshipbuilder)
Post #: 6
RE: BTR AXIS AI Cheats - 8/8/2019 2:59:40 AM   
warshipbuilder


Posts: 878
Joined: 2/23/2013
From: C-eh-n-eh-d-eh
Status: offline
I had a look at a few aircraft production numbers for context in comparison to the P-51B. None of them matched reality. So that will be my next project once I am done with the one I have.

_____________________________

warshipbuilder

Any ship can be a minesweeper, once.
Editor WitE2Pedia
ED/BTR Ressurection Project

(in reply to ReginaldBednar2)
Post #: 7
RE: BTR AXIS AI Cheats - 8/9/2019 8:28:37 PM   
ReginaldBednar2

 

Posts: 9
Joined: 8/2/2019
Status: offline
The P-51 data below comes from the book "Air Force Combat Units of World War II", edited by Maurer Maurer, Chartwell Books.

It will be assumed that each P-51 Group contains 72 a/c at full strength (same as game).

The following Groups converted from P-47 to P-51 in the month specified:
VIII Fighter Command, 65 Wing: 4 FG (Apr 44), 355 FG (Apr 44), 356 FG (Nov 44), and 361 FG (May 44) = 288 P-51s
VIII Fighter Command, 66 Wing: 78 FG (Dec 44), 353 FG (Oct 44), and 361 FG (May 44) = 216 P-51s
VIII Fighter Command, 67 Wing: 252 FG (Apr 44), 356 FG (Nov 44), 359 FG (Apr 44), and 364 FG (Sum44)= 288 P-51s
15 AF, 306 Fighter Wing: 325 FG (May 44), 332 FG (Apr 44) = 144 P-51s

The following Groups converted from P-38s to P-51s in the month specified:
VIII Fighter Command, 65 Wing: 479 FG (Dec 44) = 72 P-51s
VIII Fighter Command, 66 Wing: 55 FG (Jul 44), 479 FG (Dec 44) = 144 P-51s
VIII Fighter Command, 67 Wing: 20 FG (Jul 44) = 72 P-51s

The following Group converted from Spitfires to P-51s:
15 AF, 306 Fighter Wing: 31 FG (Apr 44) = 72 P-51s

From April 1944 to December 1944 (9 months or 279 days), 1296 P-51s were added, and assuming 100% replacements,
2592 P-51s over 279 days is 9.29 a/c per day. Another reference has P-51 FGs containing 111 to 126 a/c (Feb 1945 TOE). Assuming
111 P-51s per FG, that would be 1998 a/c with 100% replacements, 3996 P-51s or 14.32 P-51s per day. Taking the average, 12 P-51s per day may be the best compromise with the available data. Note that this starts before April 1944, so the planes will be ready in time in Europe. This would amount to 12 P-51Ds per day (up from 8). Note that this does not factor in other groups using the P-51.

The P-51Bs are more nebulous. Production started in June 1943 and the game is short a little less than 2000 P-51B/Cs total. This would be equivalent to 6 P-51Bs per day starting in April 1944 until the end of December 1944. Unfortunately I could not find when production stopped. P-51D production started in March 1944. The P51B/Cs where used in Europe until the end of the war (VE Day).

(in reply to warshipbuilder)
Post #: 8
RE: BTR AXIS AI Cheats - 8/9/2019 9:08:24 PM   
simovitch


Posts: 4492
Joined: 2/14/2006
Status: offline
Make sure we account for what was sent to lend lease to England and all the other allies.
My vote is that the player needs to mind his casualties and we don't mess with what Hard Sarge and his group came up with. But that's just me.

_____________________________

simovitch


(in reply to ReginaldBednar2)
Post #: 9
RE: BTR AXIS AI Cheats - 8/9/2019 11:38:21 PM   
warshipbuilder


Posts: 878
Joined: 2/23/2013
From: C-eh-n-eh-d-eh
Status: offline
quote:

we don't mess with what Hard Sarge and his group came up with.


In order to get this project off the ground I had to agree, with Matrix, to only fix things that were broken. Such as the turn back bug. Others have to be kept within the spirit of the game as it was set out. We just can't morph it in to another entity. Possible changes such as a 24-48 hour weather forecast just make sense and are not a big deal. Making changes to the a/c production schedule is a whole other matter. Unfortunately Hard Sarge has passed on, so I can't consult with him on how he arrived at the numbers he did. Maybe he just didn't have the correct figures to work with. Yes I will look at the production numbers at some point and see how they work out as compared to reality. It could be a case of the Allied player gets 4000 a/c as of now, but we make changes, but he still gets 4000 a/c just different amounts of a/c types. No matter, that is a ways off, as I have other things on the go.

I would like to thank simovitch and our programmer Alberto, for all the work they have done to date and fixes they have made that have helped to improve the game. Without them we would still be spinning our tires at 1.02. So thanks guys.

_____________________________

warshipbuilder

Any ship can be a minesweeper, once.
Editor WitE2Pedia
ED/BTR Ressurection Project

(in reply to simovitch)
Post #: 10
RE: BTR AXIS AI Cheats - 8/9/2019 11:56:52 PM   
ReginaldBednar2

 

Posts: 9
Joined: 8/2/2019
Status: offline
I'm very sorry Hard Sarge passed away; I really enjoyed his comments. Thanks for discussing the P-51 replacement issue.

(in reply to warshipbuilder)
Post #: 11
RE: BTR AXIS AI Cheats - 8/12/2019 4:58:11 AM   
Pete McCullagh

 

Posts: 71
Joined: 6/7/2019
Status: offline
Just some thoughts,
Re 2) weather. Allies always had a pretty good forcast of weather coming from Atlantic (literally 100s of ships and ac), from the east, not so much. Cloud cover does not change from5% to 90% unless there is a major front) (or vice versa). Changes should be +- 10-15% absent a front. Systems typically take 2-4 days to move thru. Yes there are always exceptions and winter is worse than summer. Future changes to weather mods should keep this in mind. I have no idea as to the complexity of the current model or what is possible, just wishing for as good a model as is practical.

(in reply to ReginaldBednar2)
Post #: 12
RE: BTR AXIS AI Cheats - 8/12/2019 5:13:28 AM   
Pete McCullagh

 

Posts: 71
Joined: 6/7/2019
Status: offline
More thoughts re10) I posted somewhere else something similar, but here goes. Any ac that crashes on landing will require significant repair, not 3-4days, more like 3-4 weeks. So here is the suggestion- all crash on runway and say 50% of crash over friendly territory go into limbo (repair/transport) for 30 days, then return to pool. Squadron/group draw from pool as in destroyed ac. The net effect will slow upgrades due to smaller pools. This should reflect history a little better than current model. At least this is what 23 years in aircraft maintenance has told me.

(in reply to Pete McCullagh)
Post #: 13
RE: BTR AXIS AI Cheats - 8/15/2019 3:02:24 AM   
ReginaldBednar2

 

Posts: 9
Joined: 8/2/2019
Status: offline

quote:

ORIGINAL: Pete McCullagh

Just some thoughts,
Re 2) weather. Allies always had a pretty good forcast of weather coming from Atlantic (literally 100s of ships and ac), from the east, not so much. Cloud cover does not change from5% to 90% unless there is a major front) (or vice versa). Changes should be +- 10-15% absent a front. Systems typically take 2-4 days to move thru. Yes there are always exceptions and winter is worse than summer. Future changes to weather mods should keep this in mind. I have no idea as to the complexity of the current model or what is possible, just wishing for as good a model as is practical.


The Allies must may sent recon planes to check on the weather as well, which is not factored
into the game. I am not looking for super-realism here, just some hint as to which areas of
France, Germany, Italy, and Austria have the best weather (which may be still horrible).
This could be just a weather button with targets colored green, yellow, and red. As usual,
the data could sometimes still be wrong!

(in reply to Pete McCullagh)
Post #: 14
RE: BTR AXIS AI Cheats - 8/15/2019 3:18:42 AM   
ReginaldBednar2

 

Posts: 9
Joined: 8/2/2019
Status: offline

quote:

ORIGINAL: Pete McCullagh

More thoughts re10) I posted somewhere else something similar, but here goes. Any ac that crashes on landing will require significant repair, not 3-4days, more like 3-4 weeks. So here is the suggestion- all crash on runway and say 50% of crash over friendly territory go into limbo (repair/transport) for 30 days, then return to pool. Squadron/group draw from pool as in destroyed ac. The net effect will slow upgrades due to smaller pools. This should reflect history a little better than current model. At least this is what 23 years in aircraft maintenance has told me.



The key here is what does "crash" really mean. Obviously when the plane is destroyed
the damage was so great the plane had to be written off. A crashed plane presumably
took more damage than an un-crashed plane. There are no Allied spare parts modeled.
A plane could crash because one wheel was shot up, or the landing gear. Hopefully there is
some type of global damage number for the plane, which could then be translated into "X"
days to repair. Note that saving the crew was more important than saving the bomber.
Finally, the crew could bale out if the plane rather than be more like a Kamikazi mission.

(in reply to Pete McCullagh)
Post #: 15
RE: BTR AXIS AI Cheats - 8/15/2019 2:47:03 PM   
Pete McCullagh

 

Posts: 71
Joined: 6/7/2019
Status: offline
With regard to weather recon: these missions went In a day to several hours prior to the main raid. Their data was always out of date by the time it got to the mission commanders. Our view of the map at 4:30 is probably better than what they had. However they had fairly good long range forcasting say 3days to a week. We get nada!
With regard to crashing over friendly territory, returning a portion of those to service would simulate landings at emergency fields.

(in reply to ReginaldBednar2)
Post #: 16
RE: BTR AXIS AI Cheats - 8/15/2019 3:48:02 PM   
Arkham

 

Posts: 85
Joined: 9/7/2006
Status: offline
How is it that German NI's know which city to orbit when I am trying to fake them out? If i make the path look like its going to say, Hamburg and then turn back west before getting there to hit Bremerhaven, the AI somehow knows this and circles Bremerhaven, not Hamburg 10 minute before my bombers get there.

(in reply to Pete McCullagh)
Post #: 17
RE: BTR AXIS AI Cheats - 8/16/2019 2:38:01 AM   
ReginaldBednar2

 

Posts: 9
Joined: 8/2/2019
Status: offline

quote:

ORIGINAL: Arkham

How is it that German NI's know which city to orbit when I am trying to fake them out? If i make the path look like its going to say, Hamburg and then turn back west before getting there to hit Bremerhaven, the AI somehow knows this and circles Bremerhaven, not Hamburg 10 minute before my bombers get there.


I noticed the German NI's are waiting at the return portion of the bombing run before the
first bombs are dropped! I think the AI uses the plots for paths to search. It seems more
ideal to use the actual ground radar plots to insert the NI's into the most dense bomber
stream parts.

(in reply to Arkham)
Post #: 18
RE: BTR AXIS AI Cheats - 8/17/2019 1:01:38 AM   
simovitch


Posts: 4492
Joined: 2/14/2006
Status: offline
The AI knows enough when and where to set "picket" patrols at likely places and intercepts raids fairly efficiently. It does not have more information than a human would but it can run calculations better with what info it has. I encourage folks to play the RAF in BoB or the LW in BTR and see if you can force a sudden death on the attacker.

_____________________________

simovitch


(in reply to ReginaldBednar2)
Post #: 19
RE: BTR AXIS AI Cheats - 8/20/2019 1:17:10 AM   
ReginaldBednar2

 

Posts: 9
Joined: 8/2/2019
Status: offline

quote:

ORIGINAL: simovitch

The AI knows enough when and where to set "picket" patrols at likely places and intercepts raids fairly efficiently. It does not have more information than a human would but it can run calculations better with what info it has. I encourage folks to play the RAF in BoB or the LW in BTR and see if you can force a sudden death on the attacker.


It seems the Luftwaffe NI always move in groups, with each group moving along a line.
Several groups can move along the same line. The lines appear to be very close to the
actual bomber streams. I don't remember pickets not on these stream lines.

(in reply to simovitch)
Post #: 20
Page:   [1]
All Forums >> [Current Games From Matrix.] >> [World War II] >> Gary Grigsby's Eagle Day to Bombing the Reich >> BTR AXIS AI Cheats Page: [1]
Jump to:





New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts


Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

0.141