Counting serial numbers
- From: "Geoffrey Sinclair" <gsinclairnb@xxxxxxxxxxxxx>
- Date: Thu, 17 Jan 2008 14:05:36 -0500
72 to 91From the American Statistical Association Volume 42, 1947, pages
An Empirical approach to economic intelligence in World War II by
Ruggles and Brodie.
It details the activities of a US intelligence team working in the UK
using the serial numbers of various captured equipment, or on
captured documents to figure out German war production.
Since the RAF did aircraft they largely concentrated on ground
So they started by cracking the two letter code telling when a tyre
was made, this was in fact a pre war code so consumers would
not know the tyre age, only industry insiders would know. They
had a sample of 2,000 tyres in October 1943 and a methodical
effort raised that to 11,000 by July 1944.
In addition the tyres were marked with their mould number, as
well as the percentage of natural rubber in the tyre. By noting
the dates and knowing the production capacity of a mould they
could estimate production by time by manufacturer and also
Germany's rubber consumption.
Next came tanks, as people are aware each tank itself had a
chassis number, with each manufacturer having a block of
numbers. What complicated things was if a model change
occurred the manufacturer would start the new chassis
number at the next 100. So for example 10651 is the last
of the old model 10700 is the first of the new model. This
was shown by the fact numbers ending on 00 to 25 were
over represented and numbers 76 to 100 under represented.
Yes there is an error in the paper, it should be 1 to 25 and
76 to 100, or 0 to 25 and 76 to 99.
After 1940 each tank manufacturer was given a three letter
code to replace their brand name. However each manufacturer
used a distinctive faceplate to mount the code. Krupp's
Magdeburg plant simply filed the company name and trademark
off and then stamped their letter code. As a further check each
tank had to be accepted by the Wehrmacht, and each inspector
had his own code, and they stayed at a given plant, and they
stamped their code onto accepted tanks. Sometimes the tank
log books even identified the inspector by name.
As a result the investigating team concluded the Germans produced
14,000 tanks 1939 to 1942, versus the then existing estimate of
40,000. The new 1942 figure was 3,400 versus 18,000. Chamberlain
and Doyle note production as 787 in 1939, 1,593 in 1940, 3,285 in
1941 and 4,436 in 1942, total 10,101. (counting flame and command
vehicles). Jentz does not have production figures for January to August
1939 but does say the totals were 1,543 in 1940, 3,257 in 1941 and
4,227 in 1942 (and 321 September to December 1939), so 9,027
plus whatever the 1939 production figure was. The USSBS goes
for 247 September to December 1939, 1,459 in 1940, 3,245 in
1941 and 4,137 in 1942.
Other markings also helped, it turns out gearboxes were numbered
in either one or two simple series. As a result in January 1945 the
allies had an idea of November 1944 Panther production. The
dates stamped on components helped tell when the tank was built
and give an aide of industry stocks, parts sitting on the shelf for
months were clearly in good supply.
The article claims the allies received a Panther from the USSR
and one from Sicily, both arrived in England before D-Day.
Analysis gave an estimate of 270 Panthers built in February 1944.
(Jentz says 256 Speer says 276) This indicated there would be
more Panthers present in France than previous intelligence
estimates had predicted.
When it was realised how much useful information could be
extracted from the markings forms were made up asking for
the 40 to 50 most important markings for a given piece of
The team released a report on half tracks, another on trucks and,
because they were not a priority air target, only some basic work
on captured guns and ammunition markings.
V-1s and V-2s came in for special attention. Things like the
inspection markings showing who was manufacturing the weapons
and showing the effort was so dispersed bombing the factories
was not likely to be of much use. V1 production was significantly
over estimated, the Germans had manufactured far too many
components compared with the ability to make the sheet metal
bodies. After an early underestimate of V2 production due to
the numbers expended in tests production was estimated at
around 15% more than actual, mid September to mid November
1944 then almost exactly correct from then until mid February 1945.
The article notes the results obtained were usually good estimates
of German production, using the Speer figures released post war.
The big failure was assault gun production, which was grossly
underestimated, due to a lack of data.
For example their estimates of tank production were 169 in
June 1940, 244 in June 1941 and 327 in August 1942, the
Speer figures are given as 122, 271 and 342 respectively. The
figures from Jentz are 111, 256 and 317.
These estimates can be compared to the previous allied estimates
of 1,000, 1,550 and 1,550 respectively.
Remove the nb for email.
- Prev by Date: Re: Churchill's reference to 'perverted science'
- Next by Date: Re: Why so few U-boats at the start of the war?
- Previous by thread: Unguided, surface to surface rockets in WW2: effective?
- Next by thread: Did Operation Bajadere ever happen?