Question #2: Were there any trends in Ostheer's tactical PoW hauls?
To clarify: in this section and the preceding I take the German data we have at face value, reserving analytical conclusions for Question 3.
Ostheer reported capturing 291,934 men in November 1941. https://web.archive.org/web/20160614163 ... h_gen.html
As there were no operational encirclements during November, and as reservist interceptions at this time were highly unlikely, all of these can be considered "tactical" PoW.
Subtracting November's haul from the tactical total previously estimated at 339k PoW leaves only ~47k tactical PoW during the rest of 1941.
...implying a monthly tactical PoW rate 30 times higher
in November than the rest of 1941.
For the post-Vyazma period beginning on October 19, Army Group Center shows higher tactical PoW's at a more granular level. Its data show at least as a daily PoW's in the second half of November, compared to the 25 days following October 19th. https://wwii.germandocsinrussia.org/ru/ ... ect/zoom/6
Note that while Ostheer's overall PoW's fell in the last decade of November, AGC's did not. Of course in that period AGC was still advancing (slowly) while AGS's defensive/retreat began in late November and AGN was stuck.
Aside from the data peak in late October and November, there are signs of a smaller peak earlier: AGC reported 92k PoW during the defensive battles ("Abwehrschlacht") on its front between August 6 and September 27. https://wwii.germandocsinrussia.org/ru/ ... ect/zoom/6
Obviously the Abwehrschlacht involved no German operational encirclements nor opportunity intercept reservists. AGC was taking ~1,500k PoW/day during this period, which is short of the data picture for the post-Vyazma period but well ahead of the average for the rest of Barbarossa based on AGC having only <40% of Ostheer in the Abwehrschlacht. It's consistent with an upwards trend, in other words.
Question #3: Analytical discussion, including data issues
I see three main sources of error for quantifying Ostheer's tactical PoW haul:
- 1. Apportionment of intercepted reservists between operational and non-operational PoW totals. [Variable "X"]
- 2. Apportionment of OKH's December 20th correction to total PoW's between operational and non-operational hauls. [Variable "Y"]
- 3. Reporting delays from operational encirclements, causing a false impression of higher tactical PoW hauls. [Variable "Z"]
We can test the validity of my proposed trendline (significantly higher tactical PoW's in late fall) by positing whether the trend remains after positing different values for X, Y, and Z. First a bit more explanation of the variables:
This is the percentage of intercepted reservists who are properly attributable to operational encirclement. As the variable increases, Ostheer's total tactical PoW's declines. If, for instance, all reservist interceptions occurred during the Minks/Smolensk battles, then Ostheer's total tactical PoW count need not be decreased at all. Increasing the variable makes my hypothesis weaker and vice versa.
This is the percentage of downwards revision to Ostheer PoW totals (553,559) that should be apportioned to the operational encirclement totals. I.e. how many fewer men did the Germans really capture at Minsk/Smolensk/Kiev/etc.? Because I estimate tactical PoW's by subtracting operational PoW's from Ostheer's total, apportioning downwards revision to the operational total is the same as increasing Ostheer's total tactical PoW haul. That, in turn, weakens the trend apparent in the higher late-Fall tactical PoW data.
I have not found any document apportioning OKH's December 20, 1941 reduction of Ostheer PoW totals by period or by unit/army/group. Can anyone help? It would remove some uncertainty from this discussion...
This is the percentage of apparent tactical PoW's (i.e. those reported temporally/geographically outside of operational encirclement battles) that, in reality, should be attributable to operational encirclement. There would be two main sources for this error: reporting delay and stragglers from encirclement captured behind the lines.
Because the late-Fall period I highlight immediately followed big encirclements (Vyazma-Bryanks and Chernigov/Nogai Steppe), it's at least theoretically possible that Variable Z is sending me a false signal through the data. But for several reasons, I don't find it likely that this factor can account for a great part of the late-Fall increase in tactical PoW present in the data:
- AGC's rate of capture during the second half of November - over a month after the Viazma/Bryansk pockets collapsed - was higher (5,600/day) than its rate of capture between Oct 19 and November 14 (4,300/day). https://wwii.germandocsinrussia.org/ru/ ... ect/zoom/6 If delayed Taifun reports were really at fault, they'd likely have diminished in importance over time and therefore AGC would have reported fewer PoW in November than during latter October. The opposite is true.
- AGC's PoW hauls decline precipitously exactly when we would expect them to, given broadly timely reporting: when the German offensive ended. AGC reported only ~500 PoW/day during the last 20 days of December. https://wwii.germandocsinrussia.org/ru/ ... ect/zoom/6 If PoW reporting was delayed after Viazma-Bryansk, it suddenly became accurate during the panic and communications problems caused by the first Soviet counteroffensive.
- The foregoing is true for Ostheer as a whole: Its PoW numbers fall significantly in latter November when AGS and AGN stall out; its December figure is ~1/4 of November's. Once again it's either PoW reporting suddenly became accurate or reporting was reflecting what was happening at the front.
- The Germans corrected their PoW reports throughout the war; the AGC reports I cite are from December 1941 and later. If reports of PoW captures were delayed, staff would likely have properly assigned the PoW counts to the proper times of capture.
- No doubt some of AGC's post-Taifun PoW's were stragglers/escapees from the Kessels. As with delayed reports, however, we would expect this factor to diminish with time after the encirclements. Again, the opposite is true: AGC's capture rate was higher in latter November than in the immediate post-Taifun month.
Ok so let's play with the variables a bit to see under what parameters Ostheer's late-Fall apparent tactical PoW hauls don't reflect a trend.
- Z= ?? (dependent variable)
Here I'm assuming that half of intercepted reservists - 250k - are already included in operational encirclements. This seems excessive, as at no time did Ostheer encircle territory in which half of the Red Army was deployed. This parameter increases total tactical PoW by 85k from the OP calculation. (now 589k)
Y=65% is based on the proportion of PoW claims in operational encirclements (2.549mil) out of total Ostheer PoW claims prior to revision (3.92mil). It implies that the Germans double-counted their operational PoW's to the same extent as non-operational. Note, however, that AGC's post- and pre-revision statistics for operational encirclements reflect no significant downward revision of operational encirclements: https://wwii.germandocsinrussia.org/ru/ ... ect/zoom/6 https://wwii.germandocsinrussia.org/ru/ ... ect/zoom/6
Thus the error that OKH corrected probably stemmed from OKH-level tabulation errors rather than from local-level battle appraisals. As AGC's encirclements account for >60% of the operational totals, Y=65% seems implausible. Nonetheless, let's stick with it for discussion purposes. 65% of the 539,559 revision is ~350k.
Now our inferred tactical PoW total for all of 1941 is 939k.
Now we have 292k tactical PoW in November's 30 days (9.7k/day) and 647k in Barbarossa's other 163 days (4k/day).
So even under the extremely unfavorable parameter values for X and Y (IMO implausible), Z=60% is needed to equalize the two periods. I.e. a majority of Ostheer PoW reports were improperly time-stamped and/or tens of thousands of stragglers from the October pockets remained behind German lines and were captured in November. Doesn't seem plausible for the reasons I outline above, especially the upwards tick in AGC's PoW haul in latter November.
At some point I hope to expand this model to account for smaller-scale encirclements during the June battles in Lithuania and Ukraine. For example 87 and 124 ID's had highly unfavorable deployments and were encircled quickly by AGS. As these resulted from unique circumstances, it seems improper to consider them tactical surrenders along the same lines as happened in later weeks.
It would also be helpful to disaggregate offensive from defensive periods for comparison.
Absent some of the considerations I've presented, we must believe that over half a million Soviet soldiers surrendered to the Germans during Summer 1941 without being encircled. Nearly every German accounts from early Barbarossa mentions the intensity of resistance and high German casualties; these accounts seem irreconcilable with the usual reading of the data.
My proposed trendline revises this picture significantly: many fewer Soviet soldiers surrendered tactically in June-July 1941, most early non-operational PoW's were caught defenseless while mustering for service or trapped in untenable tactical circumstances owing to the initial surprise. Only later did tactical surrenders start to reach worrying proportions, when the Germans seemed poised to take Moscow and the average Soviet soldier was a barely-trained, poorly-armed/supplied replacement. The RKKA winter offensive restored morale and reduced tactical surrenders.
This revised picture is more line with both Soviet heroism and the rational individualism of people everywhere, who will make plans for surviving any eventuality including conquest of one's homeland.