A few words on interpretation of accuracy figures.

[Peasant]

When reading from firing tables the probability of hitting a "tank sized" target with certain % confidence at a given range, it is important to remember that size of tanks around WW2 did vary significantly. For example, here is the german 2.5 x 2m dispersion rectangle overlayed over the M4 Sherman and M3 Stuart tanks. We can see that "probability of the shot striking within 2.5 x 2m rectangle" describes well the % of direct hits on M4 Sherman silhouette, but is a poor predictor for accuracy against M3 Stuart and similar small vehicles.

[Mobius]

When reading from firing tables the probability of hitting a "tank sized" target with certain % confidence at a given range, it is important to remember that size of tanks around WW2 did vary significantly. For example, here is the german 2.5 x 2m dispersion rectangle overlayed over the M4 Sherman and M3 Stuart tanks. We can see that "probability of the shot striking within 2.5 x 2m rectangle" describes well the % of direct hits on M4 Sherman silhouette, but is a poor predictor for accuracy against M3 Stuart and similar small vehicles.

Yes, for small tanks it is a poor predictor. I would center the aim on center of mass.

[Peasant]

When reading from firing tables the probability of hitting a "tank sized" target with certain % confidence at a given range, it is important to remember that size of tanks around WW2 did vary significantly. For example, here is the german 2.5 x 2m dispersion rectangle overlayed over the M4 Sherman and M3 Stuart tanks. We can see that "probability of the shot striking within 2.5 x 2m rectangle" describes well the % of direct hits on M4 Sherman silhouette, but is a poor predictor for accuracy against M3 Stuart and similar small vehicles.

Yes, for small tanks it is a poor predictor. I would center the aim on center of mass.

Thing is, in many situations the bottom portion of the tank is hidden from view by shrubbery, debris, uneven terrain ecc. So it makes more sense to assume that a shot aimed at the visible "center of mass" would be higher than the geometric CoM.

[Mobius]

[Mobius]

You are right in a practical sense. I read that in a modern Swedish article they considered a target over 1000 meters to be obscured below 1 meter in height. This may be self serving per their S-tank.

[Yoozername]

When reading from firing tables the probability of hitting a "tank sized" target with certain % confidence at a given range, it is important to remember that size of tanks around WW2 did vary significantly. For example, here is the german 2.5 x 2m dispersion rectangle overlayed over the M4 Sherman and M3 Stuart tanks. We can see that "probability of the shot striking within 2.5 x 2m rectangle" describes well the % of direct hits on M4 Sherman silhouette, but is a poor predictor for accuracy against M3 Stuart and similar small vehicles.

Yes, for small tanks it is a poor predictor. I would center the aim on center of mass.

I don't think that is the intent of the 2x2.5 m target. It is just something to aim at while zeroing their weapons. Also, checking the dispersion also. I would expect a 7,5 cm L70 weapon, firing at a known range (1000 meters), to have dispersion much smaller than the 2x2.5 meter target.

viewtopic.php?t=228421&start=15

according to Miles Krogfuss, the 75mm L70 KWK42 had the following dispersion:

"Using German data that I have, here are some of my Pzgr.39 figures to meditate about concerning 50% Breite/Hohe scatter:
100 meters 0/.1,
500 meters .1/.2,
1000 meters .2/.315,
1500 meters .45/.55,
2000 meters .6/.8,
2500 meters .9/.11,
3000 meters 1.1/1.5."

If you mean 'accuracy', then that is nothing more than hitting what you are aiming at. Hitting the side of a barn would be accurate, if that is what you are aiming at.

[Yoozername]

This page from the 'hetzer' manual shows how the 2x2.5 M target is marked with both a optics aiming point, and desired gun impact area. The impact area is 'center of height/width', I believe. But with most targets being shorter than wide, and dispersion being greater in height, than width; well, you can do the math.

[Yoozername]

From mobius' website...

6 One way to manually compute the German dispersion and accuracy numbers for a target size of 2.50m x 2.0m. use the following formula:
Converting the 50% zone dispersion to terms of standard deviation. Looking up the German values for 50% dispersion zones rounded to the nearest tenth at say 1500m we find: lateral zone = 0.5m and vertical zone = 0.7m. Divide each by 0.67449 to convert to standard deviations. (See Rheinmetall conversions.)


Lateral : 0.5m/0.67449 = 0.7413m
vertical: 0.7m/0.67449 = 1.0378m
Compare target size to standard deviation (s.d.) to get a ratio of how much of the target is covered:
Lateral: 2.5m/0.7413m = 3.372 standard deviations. This contains 99.91% lateral probability.
Vertical: 2.0m/1.0378m = 1.927 standard deviations. This contains 94.6% vertical probability.
Multiply these together to get the total percentage to hit.
99.91% x 94.6% = 94.51% ~ 95% as listed at 1500m in the 88mm/L71 firing table.

http://www.panzer-war.com/page32.html

[Yoozername]

I think the green lines are more representative of what each AFv represents as a target. My take is that the M4 is actually a larger target than most tanks in WWII. In any case, adjusting to shoot for each AFV 'center of mass' (actually height), gives a better idea of 'accuracy'.

For comparison...note T34/76 is about 2 M height, T34/85 is taller

[ThatZenoGuy]

Yeah the Sherman has a famously large target profile with minimal 'hard points' so to speak. The Pz4 at least has the overengineered machinegun port and drivers port which may or may not provide additional protection, as well as the rather thick cupola taking up a fair amount of the profile. T-34 is just a mess of angles and round surfaces which while thin, could be a pain in the ass to deal with.

Sherman is almost as tall as a King Tiger IIRC.

[neoconshooter]

There are two problems not addressed. First, even a small miss judgement of the actual range to the target will guarantee a miss at medium to longer ranges! Secondly finding the range on the vast majority of tanks used in WW-II by the poorer Nations was done using a "Stradiomatic" type range finder, which was both slow and not accurate, IF you demanded a hit. IE, it took several to over ten seconds to find a range using that system, for most, the vast majority, of tank and gun crews. And even when they did, there was a significant chance that the gunner made a mistake! Dispersion of the shots exacerbates the problem in that if the aim and dispersion were both high, it missed, or bounced off of the roof. If the two cancelled our, then the gunner thinks he has got a hit, but a large portion of subsequent shots at secondary targets miss because the gunner uses the "Correct" sight picture to aim at the rest of the tanks to be targeted.
Range finders fixed this problem to a greater or lessor extent in that they find the range accurately, but take too long to use. A trade off many tankers were less likely to use.
Any thoughts on the ideas expressed here?

[Mobius]

[Yoozername]

There are two problems not addressed. First, even a small miss judgement of the actual range to the target will guarantee a miss at medium to longer ranges!

This is not true. German data actually shows the range where a shot will still hit. As the range increases, the zone narrows, of course, but not as you describe. A rough example is range estimation just needs to be good enough for a high velocity gun +/- 200 meters @ 750 meters. Sensing the tracer, which you did not mention, is how the loop is closed. It is always better to be lower than high. Ricochets are possible. Think of the dispersion as a oval that moves up and down with range.

[Michael Kenny]

Side by side T34/M4

[Yoozername]

accuracy.png

Very interesting data, but it brings up more questions/assumptions. Since it is range test data, under non-combat conditions, that is a factor. What size targets are they using? Panther? It seems the 400-500 meter moving target has much more data points.