Matilda vs. PzIVF HC: bug?

[ww2steel]

I just noted today that the PzIVF (7.5cm L/24, HC, Dec'41-Dec'42) when you point it at a Matilda shows the chance of a kill as "None". This seemed odd, since the round is capable of penetrating 82mm @ 0°. The 75mm armor, with angles as low as 0° should be easily penetrated.

You don't really need to do the math to figure out that it should penetrate easily, but:

Assuming a standard deviation of a little under 10%, with 90% quality armor it should have a 98.7% chance of FULL penetration. It should have, literally, about a 99.998% chance of at least a partial penetration.

Actual game test results do not come up quite this high, but close.

Why does the game engine display "None" to the user?

Mike

[dieseltaylor]

Possibly because the choice of shot is not yours and it is showing normal AP. Now if you were to cheat and arm it only with HC I doubt the game engine would notice : )

Incidentally HC was very inaccurate past 400 metres and is not something modelled in the game. German tanks crews would carry a few for dire emergency.

[JoMac]

I think the chance kill is assuming the majority of your AP ammo, and not special ammo like HC, HVAP, APCR, APDS, etc.

[Elmar Bijlsma]

It calculates kill chance from the main AT round (being the AP round) not the 'special' HC round. The HC round can indeed defeat it but it's standard, plentiful AP round can't.

If a PzIV(short) hasn't got any AP it will consider the HC it's main AP round and it should display the correct odds to kill, I believe.

[dieseltaylor]

Really? I look forward to the results of the test : )

[ww2steel]

Actually, at the date I tested (Feb'42) HC WAS the common AT round. Roughly 5 standard AP are carried and between 0 and 16 HC are in the racks. I agree, CMAK must be measuring the standard AP. The 7.5cm L/11 IG that does not carry AP shows that it can penetrate.

Upper hull results (75mm, 0°, 90% quality): 208 hits tests

Full Pen (>67.5mm pen achieved): 152, 73%

Partial Pen (>58mm): 53, 26%

Internal Armor Flaking(<58mm): 3, 1.4%

No Significant Damage: 0

I think I underestimated the standard deviation a little (I don't know how much CM increases the SD for low quality armor, I'm just guessing based on a Rexford post).

The ranges I have done so far: (number of tanks disabled - Imm/gun/destroyed/bailing)

100m PzIVF=22%/Mat=71%

300m PzIVF=12%/Mat=67%

I will eventually take this out to 2000m and repost. This is for unaltered ammo loadouts (maybe about 4% don't carry and HC and generally are a writeoff), regular crews, 100 tanks per side-per test. Clearly the PzIV has an advantage... too bad they even pop smoke and reverse occasionally. CM does simulate the extremely poor accuracy, but I agree that it is maybe not harsh enough for some ammo (and too harsh for others).

I will try deleting all HC and seeing how it displays. I will also look at it in CMBB which I am extremely familiar and see if it does the same, I don't think CMBB does this. I am nearly certain that CMBB looks at the best penetrating (maybe not most lethal) ammo on board and gives you the kill chance for that ammo.

More after I get some sleep!

Mike

[ww2steel]

The results:

The computer displays the kill probability by looking at ammo in this order: T, AP, HC, HE

If it has no T then it looks at AP, etc. So, if it has T on board, it displays that. If it has AP on board, apparently (at least on the PzIVF) it will not display HC, even though it is generally better. Deleting all of the AP gets it to show HC kill stats.

CMBB does this as well, I checked.

The matchup between the PzIVF and Matilda II in 1942:

(shows any tactical kill- aban, KO, bailing, gun damage, immobile)

100m PzIVF=22%/Mat=71%

300m PzIVF=12%/Mat=67%

500m PzIVF=5%/Mat=59%; 47% first round hit for PzIV, Mat usually pops smoke >=500m

750m PzIVF=0%/Mat=67%; Matil fires infrequently

1000m PzIVF=0%/Mat=67%

1500m PzIVF=0%/Mat=67%; 5% PzIV 1st round hit, Matil does not fire at >= 1500m

2000m PzIVF=0%/Mat=67%; 1% PzIV 1st round hit

These are results of 700 pairings (1400 tanks), frontal engagement, played 2 player, no orders given, seperate pairings cannot see eachother

[dieseltaylor]

WW2steel. I deduce your tests are head on which makes testing easy but does not really allow for more natural battlefield events. I am concious that the Matilda side and rear armour is substantially better than the MkIV.

Also your figures quoted above are a little bit lacking in refinement though you probably have the exact data not included above. Specifically including all tactical damage is not really that helpful as a gun damaged tank or an immobilised one does not get you any points in CM1. Gun damaged tanks tend to make themselves scarce to avoid being knocked out so you are unlikely to get the points.

This is probably not the case but if all, or the majority of the MkIV's were KO'd but a substantial proportion of the Matilda's were still live then the bare statistics would be misleading.

The longer range shooting I take means that after expending all their ammo the MkIV's got those results. However it is not clear how much HC you were giving to the MkIV's for these results in your test but I assume you are buying plattons as you mentioned in your first testing post. At least with the Matilda's you know there are roughly 80 rounds of AP. : )

I appreciate the work you do - but give me a set of figures and I will look for wrinkles in the methodology : )

The first round hit rate is astonishingly good at 500m. Bit worrying really! Incidentally both HC and APDS are over credited with accuracy over RL.

[ww2steel]

Diesel, if you've seen my website you know I have much more data than shown. The Matils were almost always KO'ed because (as you certainly know) an HC penetration does many many times more damage than an equivalent penetration by 1942 2pdr shot. These tests are only for one turn/minute, because generally in the game after one minute of tank on tank combat the situation has changed as such that any two tanks that have not killed or damaged the other by the end of the minute then they probably wont.

As far as counting tactical kills- a gun damaged tank is mostly worthless except for exposing enemy AT guns that shoot at it and maybe pursuing infantry with a bow MG. A tracked tank is a pillbox, and much more easily dealt with- especially if a green crew abandons the tank. I like CM because it reflects real life as well as possible, I could care less about points. If you are a points player, it should be agreeed that a tactical kill should at least count to some percentage as a KO. (A LARGE % IMO)

I never mentioned platoons, I took 100 of each vehicle as loaded in Feb'42. The loadouts are in the first line my second post. Besides, who cares how many rounds the Matildas carry (it's actually 93 if I remember correctly) when 90 of them are burning in the racks on 50% of them. (I see you are in London, gotta give you a little ribbing here.)

I don't think I have any 'wrinkles' in the methods. I was testing HC versus the Matilda's frontal armor to inspect why it was presenting 'None' instead of 'OK'. Why would I consider side armor in these tests? Also, you didn't have to deduce that it was a frontal engagement, it says it right in the text. If you have a better way to 'field test' 1400 tanks in a statistically significant way that can eliminate all user bias, I'm all ears.

See ya,

Mike

[ww2steel]

This is from the old version of my program. (New one is down for work right now). Please excuse the poor graph resolution (10°, 100m), the new version goes down to 1° and 1m. Hope the big pic isn't too annoying. This gives a pretty good representation of the 360° vulnerability of the two tanks to eachother. Click the pic for fullsize. Website's in the signature.

[Joachim]

PzIVfs vs Shermans in late NA battles. PzIVs in flanking position. PzIVfs suffer from cower panic - despite lots of HC. So it is not just Matildas. It is a HC/AP bug.

But I ran only one test scen....

Gruß

Joachim

[dieseltaylor]

I am amazed at the lethality given it is only a minute of firing. HC loses effectiveness quite quickly with angles in RL and I believe this is reflected in the game. Would putting the Matildas at 30 degrees off make a significant difference?

I understand this is nothing to do with your original testing but it is intellectually an interesting step from your lethality figures to what lowers it - or indeed increases it.

[ww2steel]

Yes, (if my chart is showing up right on the last page) at 30 to 50°, if the turret is straight ahead, the HC round 0f 1942 has less than 90% the power required to penetrate the Matilda. Below 90% very few penetrations of any kind occur. The 90% armor quality increases the standard deviation, so penetrations can occur at a slightly lower achieved penetration than normal.

The graphs show the front of the vehicle towards the top of the page, showing the vulnerability af the turret, upper hull, and lower hull. Red is 110% power to armor thickness, orange is >100%, yellow is >90%. Red almost always penetrates, yellow rarely penetrates, and outside yellow almost never penetrates at all. Overall takes the weakest panel at any given angle and displays the range at which it can be penetrated.

If you have a Matilda, turn the hull about 40° from the PzIV and hope you don't take a round in the turret face. If you do, at least there's a chance it won't do any damage. This is good practice for most tanks, though the angle may vary slightly.