Mailbag 05-19-09

Print

User Rating: 4 / 5

Star ActiveStar ActiveStar ActiveStar ActiveStar Inactive
 

Got a question for the MailBag? Send it to This email address is being protected from spambots. You need JavaScript enabled to view it.

Q: We don’t know if static FBs will ever be replaced but it must be planned for don’t you think?
 
A: We'll have mobile (not fixed) FB's as soon as all the code to make that possible is done. We cannot predict when exactly that will be. However most already know that it's in our plans (because discussed it previously) and that's all that can be said until it's much closer to being an in game reality.


Q: What is the deal with AI, are they different on the different countries?
 
A: All AI uses weapons and ammunition drawn from the same data files that players use.

AI weapons:-

MG.34 (7.62mm)
PaK.36 (37mm)
FlaK.30 (20mm)
FlaK.28 (40mm)
FlaK.36 (88mm)

Bren.303 (7.62mm)
QF 2pdr (40mm)
Oerlikon AA (20mm)
Bofors (40mm)
QF 3.4" (92mm)

Fn.36 (7.5mm)
Sa.Mle.34 (25mm)
Ca.Mle.38 (25mm)
Bofors (40mm)
Sa.Mle.17 (75mm)

We don't want the AI to be any more red vs blue than the rest of the weapons in the game

Q: How can CRS claim that they are right when I just shot a sherman with a pak 40 and it turned around and killed me after 1 hit it!?
 
A: Here's how it works. We build a tank and we give it armour that is X mm thick where the data we gather says it should be. This is the thickness of the target when you hit it with a round from a gun, that it has to penetrate if you're talking about ATGs versus tanks. The thickness we use is based on historical data.

Then we build an anti tank gun. We give it the performance characteristics the data we gather on it says it should have. This is also taken from historical data. We tell that gun to call an ammunition file we also model. We give that ammunition the penetration performance the data we gather on it says it should have. Again, this is data gathered from historical test data.

Then we test fire the ammunition from that gun at a flat plate armour target, until we get the results we are looking for. We do this using the same game engine and files you play with. After we think we have it correctly set up in data, we validate that data with 100 test shots at each range we test for. This is normally 100m, 250m, 500m, 750m, 1000m, 1250m, 1500m, 1750m, 2000m and 2500m. That's a minimum of 1000 test firing samples, after the gun and ammunition are completed modeling. So in fact they get fired a lot more than 1000 times during this process. We actually do this in the game engine using the actual weapons files you use in the game. Just like you doing a controlled test and firing that gun 1,000 times. The only differance is we use a flat plate target, so the impact angle never changes for any shot and we don't have any missed shots either.

When the ammunition, and the gun it is fired from, are performing to a degree that the average of each 100 test firings (at each range point) is within +/- 2% of the data we have on that weapon, for every point in the 10 range tests (100m -> 2500m) we set up ... we check it in and you get it in the game when it is released in a patch that includes those weapons or data. If we don't get the results the data says we should we keep working on it until it does perform as the data says it should.

So ... we end up with a PaK.40 ATG that firing APHE ammunition at 2600 ft./sec. can penetrate 105mm of flat RHA plate at 0 degrees and 1500m range. Averaged over 100 test shots, or perhaps quite a few more by the time testing is fully completed. That's what your PaK.40 can do in the game, on average. We don't test over a login to a remote server just so that we don't get those anomalies. All testing is done locally in RAM.

Sure, players report "I shot at a Sherman and my friend says I hit it" (only the shooter sees the exact point of impact, the observer does not) but I gotta tell you, if that player hasn't done at least the minimum of 100 test shots (all hits no misses) at precisely an identical angle and range for each one of those shots (no variation) and then averaged the results for a mean avearge indicator .... well I'm going to trust our method more.

We don't create the weapons/ammunition files firing against tank models, because the requirement we get identical angles and impact criteria 1000 times in a row is not possible with that method.

So that's how an ATG is built. Or the gun on a tank for that matter. It's a pretty solid way to make sure the GUN and AMMUNITION do what they are supposed to. Once you start shooting at tanks (varied and composite angles of impact will result) you are employing the gun and ammunition we developed by the process described above ... in a shooting situation that might see you fail to achieve the same result, but that won't be because the gun does not do what it is supposed to according to the data we have on it. It will be be because of other factors.