CMBN weapons effect tests 2020

[StieliAlpha]

23 hours ago, Drifter Man said:

This was perhaps the most unexpected result for me. It looks like something is wrong with this gun. At least it doesn't matter much because this weapon is rare.

I don't have any real-world testing data and the answer would depend on the setup of the real-world test. I imagine that it most cases it involves someone shooting from a known distance at a target of a given size. At the end the hits of the target are counted. The problem with recreating this setup is that I can't define a target of some fixed size in the game and then count hits. Troops take cover and stand up again. Jeeps are too big. So I don't think it would be possible to do such a cross reference - there would always be doubt that the in-game test does not accurately represents the real-world test.

Yeah, you are probably right.

I read a little further yesterday. Especially an article of one guy, testing different ammo charges for his K98. He set up a very controlled environment, with the rifle supported on a sandbag and fixed, to ensure repeatable tests. He measured the deviation between hits and was talking about cm‘s distance between shots on 100m and 300m ranges, etc. That‘s certainly far beyond anything that CM can provide.

By reading the article, another question popped up: Do you know from your testing, if any wind effects have been considered in the CM model?

Edited August 15, 2020 by StieliAlpha

[Drifter Man]

8 hours ago, StieliAlpha said:

Yeah, you are probably right.

I read a little further yesterday. Especially an article of one guy, testing different ammo charges for his K98. He set up a very controlled environment, with the rifle supported on a sandbag and fixed, to ensure repeatable tests. He measured the deviation between hits and was talking about cm‘s distance between shots on 100m and 300m ranges, etc. That‘s certainly far beyond anything that CM can provide.

By reading the article, another question popped up: Do you know from your testing, if any wind effects have been considered in the CM model?

Yes, I'm testing not just the weapon like that guy, but the whole man + weapon "system", which is an abstraction of the game - all pixel soldiers with certain experience are mostly likely equal in CM, but, for example, how do I decide on the experience of the person who aims and shoots the gun in a real-world test?

About wind, interesting point but I don't know. All tests were run with no wind. In addition, the sample size was still too small to capture subtle differences of a few % that I would expect from wind.

[KGBoy]

Just wanted to bring this thread back for all newbies. Amazing work Drifter Man! Thanks.

[Drifter Man]

Thanks @KGBoy, good idea. I intend to keep the download link alive for as long as it is relevant (Engine 5.0, hmm...?)

[Bulletpoint]

Coming back to these test results, I am puzzled about some things I overlooked the first time.

Rifles

The American Garand rifle has a max range of 475m, while the German Karbine 98K has a max range of only 350m. Is there any historical reason for this?

Also when comparing these two rifles, it looks like the German rifle is more accurate than the Garand at range (fewer shots per kill), but if it's more accurate, wouldn't it also be effective at longer ranges?

Accuracy falloff

Looking at accuracy tables, it seems the number of bullets needed per kill increases linearly at range: Double the range, double the number of shots needed to get a hit. But isn't the accuracy falloff logaritmic in real life? So that as distance increases, accuracy decreases exponentially more?

[Bulletpoint]

Also, it would be really interesting if the data for "firepower" (kills per minute) could be plotted with line graphs, just like the accuracy graphs. The data is already included, but only as bar graphs.

[Centurian52]

58 minutes ago, Bulletpoint said:

Rifles

The American Garand rifle has a max range of 475m, while the German Karbine 98K has a max range of only 350m. Is there any historical reason for this?

Also when comparing these two rifles, it looks like the German rifle is more accurate than the Garand at range (fewer shots per kill), but if it's more accurate, wouldn't it also be effective at longer ranges?

That's interesting. I would expect their range to be identical. The raw ballistic maximum range for both rifles will be thousands of meters, far beyond the range that the human eye is actually capable of resolving a human-sized target. So the limiting factor on effective range for both rifles is human eyesight, not the technical characteristics of the rifles themselves. And since it's the same mk1 human eyeball behind both rifles, the range should be the same. And I believe that range should be around 500 meters. I don't think the Kar98k should be notably more accurate than the Garand (it's pretty accurate, but so is the Garand and every other rifle in this time period). Perhaps the American riflemen are focusing on rate of fire a bit more than accuracy since they have a semi-automatic rifle?

There may be a doctrinal reason why the German riflemen aren't shooting as far. The primary weapon of the German infantry squad is the MG, not the rifles. The rifles are primarily there to support the MG. So they may be relying entirely on the MG for long range fire, and only using their rifles when the enemy gets a bit closer.

1 hour ago, Bulletpoint said:

Looking at accuracy tables, it seems the number of bullets needed per kill increases linearly at range: Double the range, double the number of shots needed to get a hit. But isn't the accuracy falloff logaritmic in real life? So that as distance increases, accuracy decreases exponentially more?

I would think that the accuracy falloff over range should be logarithmic, since people will take more careful aim at smaller/more distant targets. But I'm not sure. I do know that the accuracy falloff should be less than what you would get if you just modeled accuracy as a fixed cone extending from the shooter, in which case if you doubled the distance it would take four times as many shots to achieve a hit (since we're shooting in a 3 dimensional world, so accuracy would falloff by distance squared rather than just distance). But since a linear accuracy falloff is less than that it might be right for all I know.