View Single Post
Old 18th August 2008, 14:20
Posts: n/a
Victory numbers : some precisions on precision

In my long post N 5 in this thread I mentioned half a dozen various (very low) totals of 1940 French fighter victories given by A. Gillet - written by me in red and bold characters - but I forgot to give the necessary explanation on the amazing "precision" of several of these figures (now I'll use the English way of writing numbers - with a decimal dot) :

354.787 (sic) victories.

On page 111 it is

341.29 (sic), don't ask me why.



These are, according to A. Gillet, the numbers of actual victories NOT just claims filed by the pilots.

For details please see post N 5 above. As you all know - I presume - total numbers of victories are almost always very uncertain, be it for a whole air force and a whole battle or even for WW II, for a German "Geschwader" (100-160 fighters during different periods) even only during the French Campaign or the BoB, for an RAF wing or squadron, a French Groupe de chasse or even for individual aces having claimed, for example, from 8 to 25 victories, later up to... 352.

Why are these totals so uncertain? The most important explanation obviously is already the uncertainty about very numerous victories, or claims, as many people nowadays prefer to call them, which is a strong indication that many people doubt every single claim (or victory) or at least want to see some solid proof. Today some people claim that the average overclaim rate during WW II was 2 : 1. So total claims would have to be reduced by half (it has still to be proven that this applies to the fighter forces of all countries). This uncertainty is caused by multiple sources of errors made in good faith, sometimes (much more rarely) on purpose to claim some victories which never were real. At least at the beginning (during the Phoney War only?) Allied pilots draw wrong conclusions when a Me 109 dived steeply, leaving a trail of black smoke. They didn't know (yet) that very often the 109 had not been hit and the black smoke was produced when the pilot applied full throttle, his fuel then not being quite perfectly burned, leaving a rest of carbon particles etc. In numerous instances an enemy aircraft was shot down by a fighter and fell down or its pilot dived more or less steeply to make a belly-landing but another fighter pilot, not being aware of this, shot it down again. In theory the same AC could be shot down several times, resulting in 2-5 victory claims (mostly 2 in such instances) for one single actual victory.

Of course we should not exaggerate the effects of these errors. They did occur but not as massively as some people think. Nowadays we have got a lot of information from the "receiving side", be it the Allied air forces or the Luftwaffe having suffered these losses, to cross-check the claims and the actual losses... which were registered and this is part of the problem for apparently part of the losses - for whatever reason like air attack (on the ground), destruction of some documents in battle or lack of time, or otherwise - were never put down in writing and officially registered, at least within the Luftwaffe.

It often happened, too, as we all know, that fighter pilots and rear gunners claimed some victories in good faith but were wrong, whatever the reason again. In the book on KG 27 we can read the story of a German rear-gunner who said his bomber was attacked by four Morane fighters. He shot down three, he claimed, but the fourth shot down the German bomber while this super-gunner was replacing his empty ammo drum (75 rounds) and couldn't fire. Nothing is impossible but I have some doubts about these three victories won in a few seconds by one single rear-gunner with his light machine-gun. Besides, this story proves, once more, that contrary to what Brian Cull reported in his book "TDIM", French fighters DID "press home their attacks with every ounce of vigour". This is at least what GERMAN (not British) aircrew remember (those who survived).

So I would say that already if a British squadron, a French GC or even half a GC (escadrille or flight) or a German "Staffel" (squadron of 12 in theory) claimed, for example, a total of 45 victories this total could always correspond 30-55 actual victories or something of this kind. Let us not forget that often an E/A was damaged and was not claimed as destroyed but it crashed shortly before landing back at base or, after landing, it was considered u/s and scrapped or used for cannibalisation, if at all. This was a real victory but the victors didn't know so these reservations work both ways. Even individual pilots' scores -as you know - sometimes were too high by several AC or even several dozens.

To sum up, inevitably precision was very poor when counting victories or claims.

Now Mr. A. Gillet comes and publishes grand totals, for the entire French fighter arm, of, for example, 354.787. Everybody having the slightest idea of errors and precision in science and technology knows that such a figure is just an illusion. Every single air battle could already result in an error of 1 to 5 victories (for example), giving a claim error of - for example - 2 to 6 or 33.33 % (sometimes more, sometimes less). When adding all these claims in order to obtain the grand total for the FC it hardly becomes better for the many errors, or possible errors, hardly compensate for each other. They rather are added to each other. So if we take "354.787" the real figure could be 299 or 506 as well, we just don't know and even the official confirmations awarded from some desk at HQ don't help much. Already giving ONE figure after the decimal point is very illusory - nave. Giving three is sheer madness.

Here we bump into the permanent problem of many purely literary-educated people, who quite simply can't understand such a reasoning. Let us try in a slightly different way : "354.787" is just an illusion because the real figure (hidden by many errors) could be anything between 200 and 500 so it is just useless to give the total with an accuracy of one thousandth of a victory, like here (...7 not ...6 or ...8)! In fact here the illusory relative (proportional) precision is one to 354,787 or better than one third of a millionth, which is possible only for the most sophisticated scientific, possibly technical, measurements including fundamental nuclear research in particle accelerators (CERN etc.) or in space technology and time measurement (which can be extremely accurate thanks to nuclear methods). Electronic devices for ex. are often very precise. A precision of 1 % on the values of their components is already excellent : usually it is rather 10 % as far as I know (perhaps it improved recently), 5 % at best, rarely 1-2 %, for example for resistors (how many ohms?), capacitors (how many micro or picofarads,) etc. 1 % is considered excellent.

One particularly amusing example of illusory precision is Formula 1 car races and all comparable car or motorbike-races : they often give the speed of a vehicle as being something like "302.986 km/h". It is very illusory and nave because the actual length of the road, or whatever, never is known with such a high precision - by far not - and even less the real trajectory, or path, followed by the vehicle so 303 km/h instead would be fair enough, possibly too acurate. Time, on the other hand, can be measured with great precision (nowadays) but they don't go beyond one hundreth of a second for a time per lap of about 85 seconds, giving lap times like 84.97 seconds - a precision MUCH lower than A. Gillet's for actual victories, below one ten thousandth!

Totals with "only" two figures after the decimal comma are very illusory too, just ten times less for they are ten times less (seemingly) accurate.

To make it short, victory totals with any figure right of the decimal dot (or comma) are an illusion for at least all the figures beyond the decimal dot are devoid of any connection to reality. On the other hand they give amateurs a very comforting impression and illusion of being highly scientific and accurate. Nope!

Last edited by Grozibou; 18th August 2008 at 16:23.