"Successful hits v successful knockouts" Topic

I am currently looking at tank (or anti tank) v tank firing in the western desert. Are there any stats anywhere regarding the percentage of direct hits which successfuly knockout or disable the target first time where the shell penetration and the target armour are roughly comparable? ie: 2pdr v PzII or III or Italian 47mm/Short 75mm v A10/A13/Crusader.
What percentage of the time would a first strike successfuly disable the target?
I accept this may either be too difficult to equate given the variables or perhaps there just aren't the hard stats available. Failing that I will be interested in your gut opinions. Thanks.
Gerard.

Well the British cannons were using solid shot AP, which caused less internal damage but was more likely to penetrate (compare the 37mm PAK to the 40mm 2pdr; 35mm penetration versus 66mm at 100 meters). I suppose the knock out depends on the crew. Any penetration might convince some crews to abandon the vehicle, whereas some might only be forced out by a more significant hit.

Judging by what other games tend to use, I believe a 1 in 3 chance of knockout is reasonable.

EDIT: Perhaps of interest, pouring over John Salt's tome of these statistics, "Applied Operations Research, Examples from Defense Assessment" (Plenum Press, 1988) suggests a graph by which the probability of a hard kill is based on the degree to which the munition's penetration performance exceeds the targets armour, to wit; 10% P(k) for 1x overmatch to 65% P(k) for 1.5x and 80% P(k) for 2x overmatch. These are obviously judgements made about modern penetrators and don't take the quality of arms and armour of WW2 into consideration.

I think 1 in 3 is not a bad ratio – I was reading about a KV-1 tank that was hit six times with two penetrations which led to the surviving crew bailing out

As noted, a lot depends on how quickly the crew bails – most tanks are not totally destroyed, they are abandoned when the crew can't effectively fight any longer

Tiger 1 during a counterattack operation in 43 took 22 hits literally falling to pieces on the crew. They all survived.

Not the western desert, but there is of course the well documented instance of the Char B "Eure" taking 140 non penetrating hits and continuing to fight on. At least in the early period, i.e. solid shot in use, it would appear to be a matter of crew morale. In the case of the Eure it must have sounded like a rain of hail and yet the crew fought on and did not abandon their tank.

One scrap of evidence from Bovington reads as such:

Although considerably greater penetration is obtained if the projectile remains whole, if the projectile has sufficient kinetic energy or the armour is thin, a hole may be smashed through. The breaking up is then not entirely unsatisfactory, for the fragments and the pieces of armour driven out form missiles very dangerous to the crew. The effect of fragments of plate has recently been emphasized by reports from Libya that the entire crew of Italian tanks are usually killed by a hit by the 2 pr. shot owing to the inferior quality of the armor which flakes very badly. This did not occur with our tanks.

Interesting points to ponder. 1 in 3 is an interesting ratio. I wonder how well it equates to most popular rulesets?

It all depends upon the tank really. I don't know if there was ever any statistical study of tanks being hit and finding out which hit caused the tank to bail or blew it up. I doubt that one could even go back to do such a study as the tanks and crews are obviously now either dead or gone. It is all speculation based upon anecdotal information combined with statistical documentation which has survived – such as armor penetration of various guns against various armor thicknesses and slope.

Anecdotally – there are too many stories of tanks being hit many, many times without the crew bailing or a shot penetrating – some from the very beginning of the war – a certain French Char B1bis? The KV-2 at the crossroads holding the bridge which was penetrated repeatedly but still kept up the fight until a handgrenade was stuck through the '88 hole finally killing off the rest of the crew, to the Tiger hit so many times it was falling apart.

If you really want to look at stats for vehicle hit/penetration/kill/bail I'd look to the loads of data complied by the U.S. Army on penetration by various guns versus various armors at various ranges – and then factor in the "morale" factor of the crew. Easy to generalize, but harder to give any statistics which mean anything in this area I fear.

Don Featherstone, in his "Tank Battles in Miniature," says:

"One of the first clear technical lessons learned in Africa was that tanks went on fire when hit, not so much because their petrol was ignited, but because the charges of the ammunition inside the tank were fired by hot splinters which pierced the shell cases…. When machine-gun ammunition began going off inside a turret there was all hell to pay and it generally took care of those inside if they were still alive."

Achtung Minen has already mentioned the curve from Shephard, Hartley, Heysman, Thorpe & Bathe, which is the only place I have seen an attempt made at any general estimate of these things. For what it's worth, a very good approximation to their curve can be achieved with a simple scheme using 1d6, as follows.

Take the effective armour thickness (in any units you like) as a defence value, and multiply by five the expected penetrative power (I'd suggest taking the W/R limit if your penetration figures specify it) of the projectile (expressed in the same units) to take as a strike value.

Express the relation between strike and defence values as an odds ratio, in the traditional boardgame style, rounding off in favour of the target. Then use the following table to find the roll-to-kill needed for the appropriate odds:

5:1 1
6:1 1 or 2
7:1 1, 2 or 3
8:1 1, 2, 3 or 4
10:1 1, 2, 3, 4 or 5

This gives a shot with a 20% overmatch of the plate a one-third chance of scoring a kill, for those people who like one-third as a number.

As an example; take an 88/56 firing at a Sherman at 500 metres and hitting the turret front. I take the 88's penetration value at this range to be 123mm (from Bird & Livingston) and the Sherman's turret front to be 76mm thick at 30 degrees, so equivalent to 95mm (using the armoour basis curve from WO 185/118).

Defence value = 95
Strike value = 123 x 5 = 615
Odds ratio = 6:1

Therefore a kill on a roll of 1 or 2.

Personally I think it unlikely that any tank is likely to survive more than a couple of clean penetrations by anything larger than 57mm calibre, but this is mere prejudice, there is very little data that I am aware of on the subject.

It is obviously very difficult to collect this data from the field. It is the custom of tank and anti-tank gunners, with good reason, once they have hit a target, to go on hitting it until they are sure it is dead -- either on fire or the crew seen to bail out. Given the difficulties of identifying exactly what you are shooting at, too, it is not unheard of for gunners to engage AFVs already put out of action -- as well as suspicious-looking items of farm machinery and piles of rubbish. So, obviously, the number of penetrations a tank has accumulated by the time an OR inspection team comes round to look at it may include many hits after the killing one.

From a theoretical point of view, it is obviously possible to conduct shotline analysis, either using a program such as TANK KILL or a simple 2-D square-counting approach as described in Morris Driels' book on weaponeering. The first requires not only access to a suitable program, but also a detailed 3-D model of the internal volumes of the target AFV; the second requires a 2-D picture of the vulnerable internal bits of the tank, and some judgement as to the probability of lethality if they are hit.

Firing trials results would probably give the best idea, if you can find them. These still require judgement as to how likely a given amount of damage would be to impose a "kill". Before the simplified (I believe US-originated) system of K-kill, F-kill, M-kill and P-kill was adopted, Chertsey used its own damage code for firing trials, which used a splendid little graphic somewhat reminiscent in appearance of a prawn cracker. The figure had a small inner circle, an outer ring, and five lines radiating from the centre. The inner circle stood for firepower, the outer ring for mobility, and each of the radiating lines for a crew member. The circle or ring would be half blocked-in or fully blocked-in to represent complete or partial loss of firepower or mobility, and each crew line partly or wholly crossed through to represent that crewman being assessed as wounded to killed. Finally, the whole figure would be surrounded by a jagged line if the target was on fire.

All the best,

John.

NO Sherman would survive a hit from an 88 at 500 yards range. The chances of a kill should be 100%. The shot would penetrate, and any surviving crew would bail out as fast as they could.

If the gun can penetrate 123mm of armor, and the tank has 95mm of armor -- the round will penetrate the armor if it hits it. By definition. End of story.

Lots of data to be going at there, thanks John. I agree with the fact that a clean strike by an 88mm against anything with Sherman type armour will normally penetrate and do exactly what it says on the tin.

"a clean strike by an 88mm"

There is the rub of course, 'clean strikes' at exactly 90 degrees to the target are the exception rather than the rule under combat conditions, which is why there is some variation. Flank shots are also rather more prevalent irl than on the wargames table.

wrt multiple shots, my dim recollection is that the rule of thumb for the number of shots from an AT gun required to destroy a tank was 5-10. This obviously includes misses, non penetrating hits, and overkill as well as single shot kills.

There are very few "single shot kills" in real life in WWII that are "documented" and those that are documented are the ones where the enemy tank blew up after the first shot. As stated above – AT gunners fired until the tank burned or the crew bailed – as it was the only way to be sure – as they didn't have the ability to nuke them from orbit yet.

Indeed, there are examples salted throughout various books concerning shermans surviving hits from german guns due to poor striking angles. Unfortunately the second usually was enough or the crew baled from the first near death experience.

Another interesting factoid concerns why Russian crew loved the Valentine so much. Since British armour contained 3% nickel there was no spalling and the crew had a much better survival rate. Unlike the 1% or so used in Soviet armour which is not enough to cause this effect.

Which leads to another argument; That late war German plate was of poor quality due to production pressure. Foundries did not have time to face harden the outer surface leaving the softer steel to absord the kinetic energy of a round. It lead to internal damage even from non penetrators because the plate fractured and spalled or joints hit by rounds shattered.

I have never found a study or reports on this but close examination of many preserved German AFV's.

Jubilation T Cornpone wrote:

Lots of data to be going at there, thanks John.

If you mean "lots of data to hunt down", yep, but I don't know how much is there to be found in the first place. The older I get, the more depressed I become about how very little we really know, in numerical terms, about how even the simple bits of the direct-fire battle work.

However, this discussion has stimulated me to dig through the pencilled scrawl in some old PRO notebooks, and I now have a couple more meagre snippets to offer.

The first is, curiously, yet another vote for the "about a third" position, and from an odd source. WO 291/2320, "The Tactical Use of V-Agents", dated 1960, mention is made of the possibility of using nerve gas in anti-tank projectiles:

"In the US it has been found that V-agents, used in conjunction with armour piercing projectiles, are very effective against tank crews. To introduce the agent into the tank it is necessary to achieve complete penetration of the armour. If this is achieved with a normal AP shell, the chance of kill is of the order of 30 per cent, so that V-agents might in this case give a much improved result."

My gut feeling when I read that was, and still is, that 30% is far too low, but, as I said, this is mere prejudice. Some evidence for higher P(kill) given penetration numbers comes from other sources, though.

WO 291/530, "Analysis of Results Obtained in Battle with 6pdr 7cwt Guns", dated 15 Oct 1943, is based on 33 engagements. The results are clearly biased, because "Only successful enagements have been reported. The results show what the gun can do, not what it usually does." This is reinforced by the fact that the number of hits obtained for shots fired is "comparable with good results under training conditions". Of the 17 enagements for which hits and misses were recorded, 30 out of 41 rounds hit (73%); of these, in 10 engagements below 800 yards 18 rounds scored 15 hits (83%), and in 7 engagements over 800 yards 23 rounds scored 15 hits (65%). That's good shooting.

Anyway, to get back to the point, the number of penetrating hits was recorded for 16 engagements, of which 9 were head-on (recall Martin's point about side shots being commoner than in wargames). Of 32 tanks knocked out, 13 were set on fire, and the number of penetrations required to start a fire is given as:

After 1 penetration -- 6 certain, 2 doubtful
After 2 penetrations -- 1
After 3 penetrations -- 2

For 2 fires the number of penetrations was not known, and 2 tanks smoked but did not burn, probably after 2 penetrations.

The report also makes the point that "a fire is more likely to result from an "easy" penetration than from a "difficult" one", so perhaps justifying increasing P(kill) with increasing overmatch.

Now, on to Shermans and 88s --

Don Lowry wrote:

NO Sherman would survive a hit from an 88 at 500 yards range. The chances of a kill should be 100%. The shot would penetrate, and any surviving crew would bail out as fast as they could.

If the gun can penetrate 123mm of armor, and the tank has 95mm of armor -- the round will penetrate the armor if it hits it. By definition. End of story.

Not "by definition"; the definition of the W/R limit (which is probably more or less what you're getting for most late-war penetration figures) is the point at which you expect to see 50% clear projectile wins (projectile all through). What the shape of the curve is for the probability of penetration above that point, I don't know, and I doubt that there is any acceptable generalisation. Furthermore, there is no guaranty that a penetration will always result in a kill.

However, I do tend to agree with Don that the 33% seems a bit low in this case (and remember that the Shephard, Hartley, Heysman, Thorpe & Bathe figures are for a book of OR student exercises, not for actual analytical work).

So, last doc and then I promise I'll stop boring you, let us consider WO 291/2385, "Anlaysis of 75mm Sherman Tank Casualties Suffered Between 6 Jun and 10 Jul 1944". This presents a sample of 45 tank casualties from 11 Armd Div and 2 Cdn Tk Bde, excluding DDs. Of these 40 were lost to AP projectiles, 4 mined, and 1 to an unknown cause. 37 of them brewed-up; 33 of the AP victims, 3 of the mined ones, and the unknown one (I am surprised by the high proportion of brew-ups on mines). The AP hits came from 75mm and 88mm guns. On a side note, the habit of reporting everything as an 88 is commented on: "Estimates by fighting soldiers were found to be unreliable since many reported they had been knocked out by 88mm, when in fact it had been 75mm shot, while the reverse mistake has not yet been discovered."

Total hits recorded on the 40 AP casualties were 53 x 75mm and 12 x 88mm for a total of 65. Of these hits all penetrated except for 3 75mm hits. The number of hits required to knock-out was:

1 hit -- 25 tanks
2 hits -- 11 tanks
3 hits -- 2 tanks
4 hits -- 1 tank
8 hits -- 1 tank

The proportion of hits that knock out was 62%, and the average number of hits to knock-out was 1.63.

The distribution of the hits was 19 on the front, 36 on the side, and 10 on the rear (remember Martin's point again).

Unfortunately this doesn't tell us all we'd like to know, as we don't know what precise kinds of 75mm or 88mm we are dealing with here, nor the ammunition natures they are using; 88/56 APCBC is obviously a very different proposition from 88/71 APCR.

Oh, and another thing: I seem to have suffered some kind of bog-eyed brainfart in looking up the penetration figure for the 88/56 in Bird & Livingston. The figure I gave is the one they give for the 75/48. The 88/56 firing APCBC at 500m they credit with 151mm, so the kill number for a turret hit I should have given according to my simple scheme was 1, 2 or 3.

Finally -- I really will stop after this -- I calculated the expected number of kills using this simple scheme from the number and distribution of hits reported here. Assuming all the 88s were 88/56s, all the 75s were 75/48s, all the ammunition was APCBC and all the hits were at 750 metres, and assuming that the proportion of 88mm hits did not vary by target aspect, I arrived at a total expected figure of 42.33 kills, that is, a 65% success rate, or about 1.5 hits per kill. Whatever misgivings one might have about such a crude and questionably-based scheme, and the assumptions just listed, this looks to me like extremely good numerical agreement. One should probably not read too much into it, though, as the results are pretty much dominated by the large proportion of side and rear hits, which have a very high chance of success.

All the best,

John.

Just because a knocked-out tank was hit twice does not mean that it was the second shot that knocked it out. It could have been knocked out by the first shot, but the firing crew didn't know it, or wasn't sure, so pumped another round into it just to be sure. Or the second shot could even have come from a different gun -- maybe even several minutes later -- the crew of which didn't know that someone else had already knocked it out.

Your data on brew ups concerns Shermans in Jun/Jul '44. Probably most if not all of these had dry ammo storage. That model Sherman was notorious for burning even when not penetrated.

Don Lowry wrote:

Just because a knocked-out tank was hit twice does not mean that it was the second shot that knocked it out. It could have been knocked out by the first shot, but the firing crew didn't know it, or wasn't sure, so pumped another round into it just to be sure. Or the second shot could even have come from a different gun -- maybe even several minutes later -- the crew of which didn't know that someone else had already knocked it out.

Indeed so -- I made precisely this point near the start of the thread, as part of my general embittered whinge about how very hard it is to get worthwhile data on this sort of thing. I think it vanishingly unlikely that the "outlier" Sherman in the sample that took 8 hits was really fit to fight after the seventh; it seems far more likely that an enthusiastic German gunner decided to give it a proper working over, whether because it was slow to burn or because it was his idea of fun doesn't much matter.

Still, if one looks at the numbers, it seems to me suspicious that the number of tanks going out to the first hit is 62.5% of the sample -- really rather close to the overall proportion of hits expected to kill. It is of course always possible to imagine that the results are accounted for by 100% post-penetration lethality and a 33% or so chance of needlessly repeating a successful shot.

Of course there is also a bias the other way, in that this report is based only on vehicles that were knocked-out. What we really need are figures based on tanks with penetrating hits that stayed in action, but I have never seen such a thing.

All the best,

John.

There could also have been some tanks that were hit and the crews bailed, but later the tanks were recovered before someone came round to tally the results.

Unfortunately for us rules-writers, the participants in a battle have more pressing things to do and think about than keeping score.

Isn't it the case that gunners are taught to keep firing at the target tank until it brews? Otherwise they don't know whether it has truly been KO'd.