Jump to content

More evidence the AI incorrectly fires tungsten...

Recommended Posts

I just ran another big test. This time between a Panther and the T-44. I first noticed this little quirk when graphing the T-34/85 against a Panther- that the AI chooses to fire APCR (tungsten) EVEN WHEN IT DOES WORSE THAN STANDARD AP. The running theory is that the AI simply chooses the best penetration numbers, and fires it, irregardless that the AP would easily penetrate and do more damage.

First, before anyone starts hollering about statistical significance this graph used 1575 pairs of these two tanks, at ranges from 100m to 2000m.

With the T-44 there is ABSOLUTELY NO RANGE WHERE IT IS ADVANTAGEOUS FOR IT TO FIRE APCR, BUT IT DOES ANYWAY. I had 5 rounds of APCR in each T-44, which they always fired at ranges up to 1500m. I took all of them away for the next series of tests and T-44 casualties DROPPED and Panther casualties INCREASED at all ranges below 1500m inclusive. It was worst at 100m where it made the Panther 12% more likely to defeat a T-44 that carried a handful of APCR. The average difference was 7%.

Intuitively it should either have no effect (not being fired) or should improve performance if it is selected to be fired. If it performs worse the AI tank commander should be leaving it in the rack… but he is not.

I can post the graph if anyone really really wants to see it, as I have the others, but it's a pain.

Is this a bug?


Link to comment
Share on other sites

I wouldn't call it a bug, I'd just say that the ammo selection routines aren't perfect--it's not like it's firing cannister or something. Also, I'd give three reasons why it might not even be unrealistic:

1) The AI (or real-life) crews might not have the benefit of your 1500 tests--even if the Soviet Army had done comparable tests and knew enough to advise its crews to fire AP instead of tungsten, it is easy to believe that such findings might not have been properly disseminated. If I recall correctly, the average Soviet tank destroyed less than one German tank, so many crews probably didn't have massive amounts of personal experience to draw on about the effectiveness of different types of ammo vs. different targets.

2) The crew might be grabbing the most accessible round in the rack that they think might do the job--this could arise with green units, etc. that didn't arrange their ammo properly, an incorrect pre-battle briefing which caused a suboptimal ammo load plan, or naturally in battle as the crew fires rounds and depletes their neat pre-battle load plan.

3) Also, I'm not sure if I've followed all of your tests exhaustively, but is this true against targets other than the panther? If not, the target might have been misindentified, thus leading to an incorrect ammo choice?

Link to comment
Share on other sites

Even if it were determined to be a bug, there are some bigger issues out there that are screaming for revision (behind armor Soviet 45mm effect anyone, Soviet armor cower routines? Just ask JasonC, he has a comprehensive list of things that could use work, or are actual bugs/mistakes.), but I don't think there is any chance of the BFC crew doing another patch.

I did find a quote from 1999 saying that any bugs or errors found in their products would be patched,

We are hardcore, dedicated researchers with plenty of development and research experience under our belts (personally I graduated with a BS in History). But of course we aren't perfect. What we miss our hardcore, dedicated testers are likely to pick up on (note that most are VERY upset with CC3...) What we all miss will be patched whenever someone can document a mistake.

quote source thread

but I have read more recent postings saying that they wouldn't continue updating old platforms ad infinitum.

Maybe there will be an update, but I wouldn't count on it. Honestly, I can't blame them, I have no idea how many man hours it takes to put a patch together. There is no profit in continually patching games that are 3-6 years old, and BFC is, after all, a business, and unless they charge for the patch, it would be work that generates no revenue.

Link to comment
Share on other sites

Yeah, I expect no patch, I just chalk it up to: Ukrainian farm boy sees big tank and fires 'super armor smasher' round. Seems plausible to me... even though I am just explaining away a bug.

Shoot, I've even told Battlefield before that I'd put $200 towards adding a Maus to the game. They need to use the multiturret 'technology' from CMAK and make an update to CMBB- like a special weapons mod for the old soviet heavies and the huge german designs. IMO the Maus has as much right to be in the game as the IS3.


Link to comment
Share on other sites

76 - sorry I missed your post (only 1 contact in right now).

All good points. And true, I've destroyed more Panther tanks in my tests than were actually produced. Cool thoughts about the incorrect prebrief, crappy ammo loadout or arrangement, etc. Add that to my 'Ukrainian farm boy'. I have done this against other targets than the Panther (for instance the Tiger), but nowhere near as exhaustively.

I am doing an unofficial 'Panther Ops Manual' that literally covers everything a Panther can and can't do in the game. Everything from speeds/bogging % in all terrains, artillery vulnerability, vul. to air attack, AT guns, win/loss graphs to other vehicles in different scenarios, a listing of all of the changes done to the tank that are reflected by game graphics (and a list of some that are not) and probably much more I'm forgetting right now. (I'm a professional pilot.) I thought it would be cool to come out with manuals similar to the completely exhaustive (and once a year for our FAA tests exhausting) aircraft manuals.

I know I probably won't make anything for all the time I'm putting in, but the Panther is my favorite tank, so I decided to tackle that first.


Link to comment
Share on other sites

Actually I study the east front almost exclusively. The IS3 is hands down the biggest threat to the Panther. The Panther cannot kill the IS3 from the front, and even with shots from the side (the IS starting at a 90° angle) there is no range that the Panther even acheives parity (a 50%+ victory ratio). I forget which ranges were optimal, but I think it was around 500m.

Off hand no other vehicles have the upper hand on the Panther. The T-44 was close, the IS2 I think kicked its butt at less than 500m, but the Panther really is great in armor battles. That's pretty much what it was designed for, because its HE round pretty well sucks compared to other tanks of its weight of the day.

For some reason I love the short barreled SU122, that's probably the next manual I will write. (Probably wont sell any, but Im just doing it for fun anyway.) This Panther one even spending 8 - 10 hours a day on it will probably still take me a few more months.

I've got a new HUGE range that can properly accomodate 100 vehicles per side. I now run 200 tests at each range (I use 7 of them now), so I have it down to only taking about 45 minutes to test two vehicle pairs. Not perfectly statistically significant, but the graphs look nice and clean, and if somethign looks out of wack I just run a few more tests to make sure it's real data and not a fluke.

Bed time, gotta work for a few days.

See you guys on Friday probably!


Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Unfortunately, your content contains terms that we do not allow. Please edit your content to remove the highlighted words below.
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Create New...