The British late war tank casualty survey wo 205/1165 a data dump

[Christianmunich]

The British survey about tank casualties wo 205/1165 is one of the most comprehensive data collections on this topic from the WW2 era. I was planning on writing a major piece on the survey but I am stuck about halfway through so I thought in case I don't finish the work I will at least drop some compiled data that helps people to quickly find entries of interest.

The following are all Sherman 75mms ( As ) and Fireflys ( B ). You can check all entries here: http://ww2talk.com/index.php?tags/wo-2051165/

tanklist2.jpg

The first columns are self-explanatory.

E = The amounts of hits necessary to knock out the vehicle.

This metric wasn't always described directly, if a single hit was scored I assumed this hit to have caused the casualty obviously, beyond that crew bale out and or critical damage as described was taken for this metrix

F = Simply marks vehicles with lacking information for Column E

G = "1" shows this was knocked out before a hit "penetrated" ( survey definition of "penetration" )

H = The number of hits the vehicle sustained and remained running. Only one tank was described that way.

J = The number of hits counted on this vehicle.

K = The number of pens

Hit 1....X = The likely chain of hits. Not easily visible here but I compiled all pens and non pens. AP p... means penetration and "n..." means non pen. If the cell has a green background the impact was a non pen and a foto is in the survey, red for pens with foto. This allows you to quickly search for pictures of non-pens for example.

Non-pens are all projectiles that did not penetrate into the crew compartment. Glancing blows or hits on tracks, gun barrels, hatches et cetera were classed as "scoops". A hit could cut of the gun barrel but would be a "scoop"

After that the number of casualties and position of the crewman. K=KIA W=WIA U=Unhurt

AC = An x indicates the vehicle had burned.


You might notice that most vehicles were ko'ed with a single hit. Here is how I compiled the numbers:

wo1.jpg

~92% of Shermans were ko'ed with a single hit or no hit at all. 2% of the Shermans required more than 2 hits. The majority of Shermans only got hit a single time ~60%. About 10% of the tanks were hit more than 3 times. Not every tank had enough data to establish those data points, such Shermans were ignored.

Here you see the compiled hits areas and if they penned or not.

Hits.jpg

You can nicely see how "non-pens" as metric carry little weight for evaluating protection if glancing blows and hits on the outer boundaries of the tank are included. Take a look how the M5 has a higher non-pen rate:

wo3.jpg

I also compiled the data about crew casualties and it also shows how misrepresented the M4 Sherman is in recent historical works. Despite supposedly having many design features with focus on crew survival like spring-loaded hatches and easy escape paths there is close to no difference in the numbers. The Cromwell actually has superior numbers.

Crew casualty numbers should be evaluated with care. Averages in low sample sizes can easily be influenced by hit areas which had different casualty rate simply based on the position of the crewmen.

wo2.jpg

The numbers and data presented above was my private "research" and I did no double checking like I normally do, so it might contain some errors. The compilation should help you to quickly find tanks that are of interest for your research. Via the ww2talk link you can easily find the specific tank, just click on the unit and then ctrl-f the ID of the tank.

My two cents about the data, this survey highlights what many people saw in the empiric data of combat performance. The Sherman as main tank of an economic powerhouse that had virtually no strictly binding limitations for their tank design was a failure. Sherman tanks due to their extremely inefficient armour design offered no noticeable protection while having limited mobility compared to light vehicles. The data shows quite convincingly that Shermans were nearly always immediately out of combat when hit by the first projectile, even "non-penetrating" hit led to "loss" of the vehicles. The ~10.000 destroyed Shermans come as no surprise but this data connects the dots. The overall sample only mentioned a single tank which crew continued firing the main gun after sustaining a penetration. Crews that had to serve in Shermans were likely impacted by the knowledge that they were supplied with non-protection tanks and often reacted completely rational after an impact. An impact meant the next one will likely destroy the vehicle. Even perfectly positioning the vehicle did not help the crew, the front was as vulnerable as the side. Normally a single hit was enough to take a Sherman out of combat even if the main gun remained combat ready and no penetration occurred. I have also specifically studied the sample in regards to armour protection and found close to no hits that were fully withstood by the Sherman armour. Most non-penetrating hits were hits into running gear or glancing blows. The M4 Sherman offered basically no protection beyond "light tank" capabilities. Recent historians neglect this survey because the contents show a different picture.

Maybe somebody gets use out of this data.

Tüdelü.

edit: As always the data is raw and without a spin. Just the facts.

Addendum: The average hits per vehicle comes at about 1,7. The US sample, which is a bit unclear shows about 2 hits per Sherman. The same survey shows about 3,3 hits per German Tiger/Panther and 2,1 for chassis of Panzer III and IV. The data is more difficult to dissect. Heavier German tanks obviously got hit significantly more often.

[Stiltzkin]

The Cromwell actually has superior numbers.

Hardly surprising.

The Sherman as main tank of an economic powerhouse that had virtually no strictly binding limitations for their tank design was a failure.

I partially disagree, but by 1944's standards it was certainly not the best they could have fielded. Cannot see a substantial difference between Soviet designs though (46% in 41, 55% in 42 and over 90% for 44-45).

~10.000 destroyed Shermans

5,206 irretrievably lost (Jun-Dec 44), all variants, including TDs (M-10), or does this figure refer to all sorts of losses?

[Richard Anderson]

~10.000 destroyed Shermans

5,206 irretrievably lost (Jun-Dec 44), all variants, including TDs (M-10), or does this figure refer to all sorts of losses?

Shhh! If you haven't noticed, this one likes to make up numbers. ~10,000 is what he thinks the "real" numbers are according to his immaculate methodology.

[Mori]

Thanks for the nice work. Going back to raw data is always a worthwhile effort. Although WO 205/1165 has been available on the internet for years, I don't think anyone made the effort of compiling the numbers like you did.

Your conclusions on the 'staying power' of the Sherman seem convincing, but I wouldn't jump to "this tank design was a failure". Assessing tank design is more than ability to sustain more than one direct hit in combat: it's also industrial efficiency, ability to train crews, flexibility of use, dependency etc. All things which are beyond the tactical scale.

If you wish to strengthen your case, you could share the excel. I'm pretty good at troubleshooting analytics on such tool

[Christianmunich]

~10.000 destroyed Shermans

5,206 irretrievably lost (Jun-Dec 44), all variants, including TDs (M-10), or does this figure refer to all sorts of losses?

Those are just the US numbers in the ETO. Overall more than 10.000 Shermans were destroyed in WW2. That Mr Anderson who worked on compiling those numbers does not know this should be surprising but it sadly isn't.

US and UK/Commonwealth forces alone lost more than 7k in the ETO. To that add Mediterannea and other Users and you easily get 10k.

[Christianmunich]

Thanks for the nice work. Going back to raw data is always a worthwhile effort. Although WO 205/1165 has been available on the internet for years, I don't think anyone made the effort of compiling the numbers like you did.

Your conclusions on the 'staying power' of the Sherman seem convincing, but I wouldn't jump to "this tank design was a failure". Assessing tank design is more than ability to sustain more than one direct hit in combat: it's also industrial efficiency, ability to train crews, flexibility of use, dependency etc. All things which are beyond the tactical scale.

If you wish to strengthen your case, you could share the excel. I'm pretty good at troubleshooting analytics on such tool

I agree the sample is extremely understudied while offering a plethora of data. Thanks for your offer but there really isn't much to "troubleshoot" the data is pretty basic, the only thing missing from the spreadsheets are the notes to all the hits, I have also compiled the written statements about the hits but the rest is in the pictures.

[Michael Kenny]

The data shows quite convincingly that Shermans were nearly always immediately out of combat when hit by the first projectile, even "non-penetrating" hit led to "loss" of the vehicles. The ~10.000 destroyed Shermans come as no surprise but this data connects the dots. The overall sample only mentioned a single tank which crew continued firing the main gun after sustaining a penetration. Crews that had to serve in Shermans were likely impacted by the knowledge that they were supplied with non-protection tanks and often reacted completely rational after an impact. An impact meant the next one will likely destroy the vehicle. Even perfectly positioning the vehicle did not help the crew, the front was as vulnerable as the side. Normally a single hit was enough to take a Sherman out of combat even if the main gun remained combat ready and no penetration occurred.

There are a good number of sources that show the opposite. 2 easily found random examples from BLUECOAT during a period you claim to have 'studied extensively'.



2nd Armoured Battalion IRISH GUARDS War Diary.



August 4th
A tp of No. 2 Sqn now faced round NORTH and joined in the battle at long range. Sjt MURRAY’s FIREFLY scored hit at long range on a TIGER, but was itself hit 6 times. The crew did not bale out until the 4th hole was made in the hull

Story Of The 23rd Hussars

August 5th
Sergeant Johnson moved his tank
forward into an exposed position, knowing it to be the only means
of dealing with the enemy tanks. He fired, but a Tiger retaliated,
knocking off his tank's track. Sergeant Johnson got out with his crew
and coolly mended it under intense shelling..............................Sergeant Jackson
had just arrived in a replacement Sherman, which was in a hopeless
mechanical condition, and it broke down on the outskirts. Sergeant
Smith continued alone into the village where he gave valuable sup-
port. .................. He was then called back to the Squadron and on his
way found Sergeant Jackson, who had been surrounded by German infantry
in a narrow lane in his immobile tank, but had refused to abandon his now
almost useless machine despite all attacks and although told to do so by the
Squadron Leader. Both Sergeants and their crews descended from their
tanks and, although under fire, managed to tow Sergeant Jackson's
back to the hill.

[Christianmunich]

The data shows quite convincingly that Shermans were nearly always immediately out of combat when hit by the first projectile, even "non-penetrating" hit led to "loss" of the vehicles. The ~10.000 destroyed Shermans come as no surprise but this data connects the dots. The overall sample only mentioned a single tank which crew continued firing the main gun after sustaining a penetration. Crews that had to serve in Shermans were likely impacted by the knowledge that they were supplied with non-protection tanks and often reacted completely rational after an impact. An impact meant the next one will likely destroy the vehicle. Even perfectly positioning the vehicle did not help the crew, the front was as vulnerable as the side. Normally a single hit was enough to take a Sherman out of combat even if the main gun remained combat ready and no penetration occurred.

There are a good number of sources that show the opposite. 2 easily found random examples from BLUECOAT during a period you claim to have 'studied extensively'.



2nd Armoured Battalion IRISH GUARDS War Diary.



August 4th
A tp of No. 2 Sqn now faced round NORTH and joined in the battle at long range. Sjt MURRAY’s FIREFLY scored hit at long range on a TIGER, but was itself hit 6 times. The crew did not bale out until the 4th hole was made in the hull

Story Of The 23rd Hussars

August 5th
Sergeant Johnson moved his tank
forward into an exposed position, knowing it to be the only means
of dealing with the enemy tanks. He fired, but a Tiger retaliated,
knocking off his tank's track. Sergeant Johnson got out with his crew
and coolly mended it under intense shelling..............................Sergeant Jackson
had just arrived in a replacement Sherman, which was in a hopeless
mechanical condition, and it broke down on the outskirts. Sergeant
Smith continued alone into the village where he gave valuable sup-
port. .................. He was then called back to the Squadron and on his
way found Sergeant Jackson, who had been surrounded by German infantry
in a narrow lane in his immobile tank, but had refused to abandon his now
almost useless machine despite all attacks and although told to do so by the
Squadron Leader. Both Sergeants and their crews descended from their
tanks and, although under fire, managed to tow Sergeant Jackson's
back to the hill.

Those are just cherry-picked anecdotal accounts written by involved personnel without any verification. The British late war sample gives us a good idea about the actual chance of such events happening. If all cases were described like the one given by me on the other thread then we can see about 1 in 100 resumed combat after being penetrated. So 1%. Those are the numbers found by trained researchers who were methodologically compiling tank casualty data. If the numbers reflect the reality we don't know but it should be reasonable to assume. I want to remind everybody that the data is readily available due to the awesome work of the folks at ww2talk, so people can judge for themselves. The sample isn't based on picked anecdotes so they likely represent are a more accurate picture of the fighting in a Sherman.

And yes I have studied bluecoat extensively, not sure how this relates to the data presented in the British survey tho, they have nothing to do with each other and this thread here is a mere data dump of my compiled research about the British survey. We might want to stay on topic.

That post-war "histories" of units strongly differ from actual collected data comes also not as a surprise.

[Michael Kenny]

I also compiled the data about crew casualties and it also shows how misrepresented the M4 Sherman is in recent historical works. Despite supposedly having many design features with focus on crew survival like spring-loaded hatches and easy escape paths there is close to no difference in the numbers.

Obfuscation. What you are trying to re-establish is the myth of the 'Death Trap'. That the crew of a Sherman were more at risk of death and injury than crews of the Unber-Panzers. The numbers show this to be completely untrue and so you default to 'myth-busting' bogus claims that the Sherman had a higher-than-average crew survival rate. It had the normal rate for a normal tank.
At Villers Bocage on June 13 1944 in % terms more Tiger crews (10 in at least 4 fully crewed penetrated tanks)were killed than 4th CLY crews ( 12 killed in 10 fully crewed penetrated tanks). Is that proof that the Tiger was a 'death trap'?

[Michael Kenny]

Those are just cherry-picked anecdotal accounts written by involved personnel without any verification.

Translation: I don't like that data, I will ignore it.

Those are the numbers found by trained researchers who were methodologically compiling tank casualty data.

Doctors who were compiling data on injuries to men.

And yes I have studied bluecoat extensively, not sure how this relates to the data presented in the British survey

It just shows how shallow your reaearch is. The book 'Story Of The 23rd Hussars' featured large in one of your (many) locked threads so clearly the information was already available to you. You chose to ignore it.

And yes I have studied bluecoat extensively, not sure how this relates to the data presented in the British survey tho, they have nothing to do with each other and this thread here is a mere data dump of my compiled research about the British survey. We might want to stay on topic.

If you make a claim and say an action was unusual and I can post evidence to the contrary then it is highly relevant. If you post incorrect assumptions then evidence that refutes those assumptions is 'on topic'.

That post-war "histories" of units strongly differ from actual collected data comes also not as a surprise.

Post-war as in 1945. That is a long time after the events in summer of 1944.........
How do you intend to dismiss the 2 IG War Diary entry about the Sherman with 6 hits ?
Is that no longer 'collected data'?

[Christianmunich]

I also compiled the data about crew casualties and it also shows how misrepresented the M4 Sherman is in recent historical works. Despite supposedly having many design features with focus on crew survival like spring-loaded hatches and easy escape paths there is close to no difference in the numbers.

Obfuscation. What you are trying to re-establish is the myth of the 'Death Trap'. That the crew of a Sherman were more at risk of death and injury than crews of the Unber-Panzers. The numbers show this to be completely untrue and so you default to claims that the Sherman did not have a higher-than-average crew survival rate. That is correct. It had the normal rate for a normal tank.
At Villers Bocage on June 13 1944 in % terms more Tiger crews (12 in at least 4 fully crewed penetrated tanks)were killed than 4th CLY crews ( 12 killed in 10 fully crewed penetrated tanks). Is that proof that the Tiger was a 'death trap'?

Kenny, it is just data if you don't believe it so be it. I don't see how cherry picking anecdotes that support your case helps, this is unscientific. Proper sampling is in order to get proper results. In regards to crew casualties, I said in the other thread "low" crew casualties were also supported by most crews leaving the vehicle after the first hit. This likely increased overall casualties but decreased per tank casualties. If most non-penetrating hits prompt the crew to bale then casualties obviously decrease. But this is hard to establish from the Wo 205/1165 so this is just my speculation.

Is that proof that the Tiger was a 'death trap'?

No, it is not, that is not how this works, sorry. The Wo 205/1165, on the other hand, is pretty strong evidence to how combat of Sherman forces may have looked on a number basis. Like explained in my OP I believe there is a reason this sample is underrepresented in historical works.

Translation: I don't like that data, I will ignore it.

You have just chosen a cherry that suits your argument, of course it gets ignored. It is not "data" it is a cherry-picked anecdote.

Doctors who were compiling data on injuries to men.

Doctors btw are trained in methodological work and are easily capable of collecting data in a scientific way. Obviously more so than an post war unit history written by members of that unit.

It just shows how shallow your reaearch is. The book 'Story Of The 23rd Hussars' featured large in one of your (many) locked threads so clearly the information was already available to you. You chose to ignore it.

What has this to do with the WO 205/1135?!?

If you make a claim and say an action was unusual and I can post evidence to the contrary then it is highly relevant. If you post incorrect assumptions then evidence that refutes those assumptions is 'on topic'.

Kenny, I don't make the claim. I say the British survey shows it. The data clearly shows that people cherry-picking anecdotes from unit diaries distorted the reality. The survey is clear. We see how actual Shermans operated and how they sustained combat. This is the most comprehensive evidence about this in existence. No other survey about this is that detailed. It is the Nr 1 empiric evidence about Sherman casualties. I can only repeat myself, if you don't like the data then so be it. Like 4000 British Shermans were destroyed and 50 war diaries of units exist you can certainly find some more outliner cases that support your point but that is not how sampling works. The cases you describe are reflected in the British survey as well, about 1% of the Shermans resumed firing after penetration. Obviously, the sample is pretty small so percentages should be taken with a grain of salt. But those are the numbers. Going from that you could expect to find 40ish such cases, but they would still be about 1%. That is the math and the issue with picking cherries, you can take outliners and make them look like the norm, that is why we should appreciate the WO 205 as research tool, it cuts the cherry picking and gives actual helpful data. I merely compiled this data and offered my research so other people can find interesting cases easier and draw their on conclusions.

Cheers Christian.

[Michael Kenny]

Doctors btw are trained in methodological work and are easily capable of collecting data in a scientific way. Obviously more so than an post war unit history written by members of that unit.

Doctors who, in many cases, worked from information provided by members of the unit rather than examining the actual tanks.

Screens9oyhtrsohot_ghhg3-vert.jpg

Screensbnbnbhot_3-venbrt.jpg

Screenshot_2-vertrt5566y.jpg

Can we take it that you will now look again at your 'research' and remove all the information that was ' written by members of that unit'.
After all if anything ' written by members of that unit' is not good enough when posted by me then obviously it is not good enough when posted by you.

[Michael Kenny]

I say the British survey shows it.

No it does not. You seem incapable of understanding the difference between the actual document and your interpretation of what it says.

[Mori]

Thanks for your offer but there really isn't much to "troubleshoot" the data is pretty basic, the only thing missing from the spreadsheets are the notes to all the hits, I have also compiled the written statements about the hits but the rest is in the pictures.

I wasn't explicit, let me be so: keeping the spreadsheet for yourself is a sure way to create distrust, even from the most open-minded people.

The fact you don't share it is an issue, and a severe one. Yes, it took you time to compile, but that's no rationale - it also takes time to write long posts on this forum, yet you still do it.

On the other hand, sharing your file should generate a lot of positive comments, not only about your attitude but also about the quality of the work. There is also some probability that comments strengthen your case, e.g., by fixing small errors or pointing alternative analysis.

[Mori]

Can we take it that you will now look again at your 'research' and remove all the information that was ' written by members of that unit'.
After all if anything ' written by members of that unit' is not good enough when posted by me then obviously it is not good enough when posted by you.

At least this information could be added to the database. If it's randomly distributed, it won't change the conclusion.