However, one thing that still remains somewhat hidden in the sheet's data yet is quite important is alpha damage. For example, the ISU is not such a feared unit because its DPS would be outrageously high, but because a single shot can cause huge health damage in the blink of an eye. If you got hit once, you don't risk the second shot even if you know that the reload will take ages (at least from the SOV player's perspective).
Similarly, the Sherman has often been critiqued to be too cost efficient due to the switchable rounds that too easily wipe or deter Axis squads. For units with switchable rounds one could also assume that the AT specialized round then performs much worse compared to a "generalist round".
We could also take the outrage of the Tiger being useless against infantry after the recent patch as an example, but to close this part I'll just quote Sander form a recent King Tiger thread:
Except that it [the KT] does have excellent anti-infantry. It reliably hits enemy squads for ~40% health every 1-2 shots, more when they are clumped, which is significantly better than the Panzer IV because it forces off squads faster and generally has a better chance to wipe squads.
So, let's put this to the test. You could run a ton of simulations with MMX's sheet and get the data from the first shots only, then process those. The issue is that all of these are randomly generated therefore deviate between simulations, and your sample size has to increase with high scatter units. Additionally, nobody knows how many samples you need per scatter area to really give reliable output.
So I sat down and coded (quite a while actually). Instead of picking everything randomly and then averaging over a huge sample size, I tried to generate more reproducible output by calculating almost "all possible" shots that could occur. Instead of working with RNG shots, I created a mesh over the complete scatter area. On every node, a shot is created and the effect on a virtual squad is measured. This leads to 400.000-800.000 simulated shots over the whole scatter area, and for each shot the AoE damage to every single model is collected.
By doing so, we get a very clear picture of how our chances are for certain effects. How many shots cause one hit kills? How many damage the target squad by at least 20% of the total health? How many shots cause 40 damage to how many models?
Or, to put Sander's quote to the test: Can the King Tiger really reliably cause 40% health damage (not nit-picking here, it just provided a good example)?
For the following data was created at a distance of 40 meters on standard formations of Volks and Riflemen (as they would be in open terrain) with the model in the center being targeted. This also has the advantage that we see the differences caused by squad formation, since both squads have 5 members. Natural hits are neglected since the chance to hit naturally is very slim, just scatter shots are considered.
First off, let's go straight for the heavies:
We can see that the chance for OHKs is actually not that high. Even for the almighty ISU the chance to cause a OHK is about 10%, and no vehicle is able to cause two OHKs against a squad in standard formation. It seems that Volks perform worse then Riflemen, which could be due to the different spacing of the squads or the targeted model. The nerfed Tigers and the IS2 all perform abysmally bad, so the we can conclude that all heavies primarily cause health damage with their first shot. But how much?
Looking at squad level, we can have a check. This is the data on shots that caused at least 40% health damage to the total squad ("they 'passed' the check for damage according to the legend):
Again, Riflemen perform better in general, many of the heavies do not seems to be able to deal 40% of damage against a single squad in standard formation. Sadly for Sanders, under the given conditions (which are not the exact in-game situation obviously), the King Tiger can barely do 40% of squad damage. But what if we lower that threshold to 20%?
Suddenly, magic happens. Way more shots now seem to be able to do at least 20% of damage, even the recently nerfed Tigers. The difference between both infantry squads becomes larger, Allied tanks seem to deal at least 20% damage in 4 out of 5 shots, while Axis tanks shooting at Riflemen still have a hard time. But is it possible to get a closer look on actual damage distributions between the models?
Of course it is! In the following graphs we can see how many models have received over 20 or 40 damage, respectively.
Here we can see the that the ISU can be a true beast (it's still firing at range 40 though, this should rarely happen in game). Although it's OHK potential is not excessively large, it's ability to damage models is very high. Guaranteed damage on Volksgrenadiers, at least 3 but often all 5 models loose half their health or more. The King Tiger shines against Riflemen in this statistic, being able to deal decent damage on 1-3 models with one shot and - at least alpha damage wise - performing better than the much discussed Pershing. The Tigers however are still not able to do much.
For smaller damages, we can get a better resolution if we lower the threshold even further.
Now finally, Tigers don't look that bad anymore. At least in the lower damage compartment, they work similarly reliable as other heavies. ISU is almost guaranteed to deal at least 20 damage to 4+ squad members, but in general, all heavies perform well.
Since the post is already getting very long, I'll just leave you with some remaining data for mediums to look at.
A release of this script is planned, but without schedule so far. Big thanks to MMX for allowing me to use his stats data from his scatter simulator for calculations. The script lets you somewhat easily filter vehicles according to categories and their stats (for example, you could have a look at all vehicles with near AoE larger than 1, a penetration > 200, all heavies as above, all vehicles from certain factions or a combination of all of this). Additionally you can pick your target squad, thresholds and distances. Plots will be plotted automatically, but many of the analysis features are yet to be implemented.
Cheers,
Hannibal
medium tank data:
Relic Entertainment
cblanco ★ 
보드카 중대 
VonManteuffel 
Heartless Jäger 
