Military Analysis. Statistics. And Hard Sums.

Adam

Very informative post.

Regarding the tank output,RV Jones used a similar model to predict the production of Wurzburg early warning radar spares from item serial numbers seized on the Bruneval raid. His proposal was to approximate the number of Wurzburg/ night defence boxes available in the Kammhuber line.

 

There is something on this in Hinsley's 'British Intelligence' Vol 2.
Production estimates were the job of MEW (Ministry of Economic Warfare) using intelligence from PR and agent networks. Early in 1942, depletion and weakness of GAF on the Russian front was higher than expected, This was due to an over estimate of production capacity. A new department AI(2a) (Air Intelligence) was set up which provided more accurate figures using data from crashed aircraft, (works numbers, dates of manufacture) and intercepted messages from test and delivery flights.
Aircraft production estimates were 'virtually accurate' Sept-Nov 1942. By January 43 the method was under estimatinng (by around 20%-25%)
U-boat production figures were pretty accurate due to PR, prisoner interrogation and broken Enigma networks giving sea trial data.
Not seen anything on tanks yet, but the books are a colossal amount of facts (around 3000 pages for the first 3 volumes)

 

Ms# P-059 Project # 47 Tank Losses


German Tank repair,losses and replacements. Collated post war from German accounts. How would these tie in with intelligence of the time?

 

I find Hinsley very readable.

All the best

Andreas

 

I've only got Vol 3 Pts 1 & 2 - they are certainly full of facts but they're not a mass of statistics. Most importantly (for those of us that will never 'read' them) they are very well indexed, Vol 3 Pt 1 is a mere 690 pages of which just under a hundred are the index.

 

I've only got Vol 3 Pts 1 & 2 - they are certainly full of facts but they're not a mass of statistics. Most importantly (for those of us that will never 'read' them) they are very well indexed, Vol 3 Pt 1 is a mere 690 pages of which just under a hundred are the index.

I would agree, quite an amazing work. full of facts. Now long out of print, it was five volumes in 6 books at about 9 1/9 inches of shelf space.
Sometimes you see volumes for sale, sometimes silly prices. I think a few volumes are rare. Vol 3 I think is more common, and more useful. I'll post some scans on the thread subject...
No doubt NMP will offer reprints at silly prices one day.
There was a single volume abridged issue, which is probably not so useful as a research tool.

 

Some scans on MEW/AI with some tank production figures. I wondered if the serial numbers were ever coded or obfuscated and it seems that some random gaps were introduced.

 

Hmmm, never read a good biography of him - anyone know of one?



R. V. Jones' book Most Secret War is all I know of, a sort of autobio. Essential reading. He was worth his weight in gold, the science version of the BP boffins.

 

Numbers can tell you a lot; I don't believe in that "lies, damned lies" cliche. (I am a baseball fan, and that sport lends itself well to numerical analysis.) Yet I also recall Niall Ferguson doing a statistical analysis of WWI losses, the point of which was that the German Army came top. The fact that the Germans LOST ANYWAY didn't count with him, apparently.

 

Numbers can tell you a lot; I don't believe in that "lies, damned lies" cliche. (I am a baseball fan, and that sport lends itself well to numerical analysis.) Yet I also recall Niall Ferguson doing a statistical analysis of WWI losses, the point of which was that the German Army came top. The fact that the Germans LOST ANYWAY didn't count with him, apparently.

Alan,

The irony is that the German army came out on top in WW2 as well. Not on the basis of statistical analysis but more from a subjective assessment of their cool uniforms, superior weapons and technical advances.
The scores of SS reenactors ignore the fact that they LOST as well. Go figure!

 

Some fascinating documents too, Von P. I was surprised by the continued presence of the Pz Kw III in frontline service in 1943 and even into 1944.

 

Some fascinating documents too, Von P. I was surprised by the continued presence of the Pz Kw III in frontline service in 1943 and even into 1944.

Well into 1944.
The best way of seeing what was in front line service is to look at the monthly losses.
Pz III losses 1944
March..134 gun ..26 Flam...47 Befl.
April ...50............6 ..........3
May.....1.............1...........4
June...41.............0...........0
July ....7..............0.........15
Aug ....2..............0...........0
Sept ..21.............0.........56
Oct....11.............0...........7
Nov ....0............. 0..........4

 

Official Munitions Productions of the US by Month July 1 1940 to August 31, 1945


Not the coffee table book of war - but -'the battle was won in the quartermasters stores!'







INTERNATIONAL AID STATISTICS WW2


Lend lease etc - Dollar Values!


On a course an American pal - gave one of those 5 minute talks much used by directing staff to boost an individuals confidence in giving lectures. A humorous take - 'when we blast a bridge that cost millions with a few thousand bucks worth of US ordnance the US taxpayer will in turn eventually have to pay for the damage to the bridge!'

 

Some scans on MEW/AI with some tank production figures. I wondered if the serial numbers were ever coded or obfuscated and it seems that some random gaps were introduced.

At least until early 1942 this does not seem to have been the case. I have seen ULTRA messages that clearly identify both tanks and planes according to production serials. The gaps seem to have been introduced when series changed (although I'm not sure on that).

All the best

Andreas

 

At least until early 1942 this does not seem to have been the case.


Yes, page 63 scan above mentions random gaps introduce Spring 1943.

 

Stumbled on a strange little nugget while looking for something else:
The German Tank Problem - Statistical Analysis Techniques, War - Statistical Consultants Ltd.
~A


This goes a bit deeper...

German tank problem - Wikipedia, the free encyclopedia

Nice.

Turing used Bayes Theorem in the naval Enigma problem.

 

LMGTFY

(Let me Giggle that for you)