American Armor Production Numbers

[Michael Emrys]

I pulled the following out of a post to the soc.history.war.world-war-ii newsgroup. I thought it might prove of some interest here.

<BLOCKQUOTE>quote:</font><HR>Richard Anderson of the Dupuy Institute sent me some figures from "Official Munitions Production of the United States, By Months, July 1, 1940- August 31, 1945," Initiated and Prepared Principally by the War Production Board, Program and Statistics Bureau, Issued May 1, 1947.

Total tanks produced in the period outlined (only M2A4s were being produced in July 1940). Total tanks 86,732, made up of T26E2 37, M26 2002 (all bar 40 in 1945), M6 40, M4 49,531 (production from February 1942 to July 1945) made up of 4,680 105 mm (2/44 to

6/45), 10,883 76 mm (1/44 to 7/45) and 33,671 75 mm (2/42 to 3/45) gun versions, M3 (Lee/grant) 6,248 (7/41 to 12/42), T25 40, T23 250, M7 7, M24 4,731 (4/44 to 8/45), M3 (Stuart/Honey) 13,859 (1/41 to 9/43), M5 8,884 (3/42 to 6/44), M2A4 330 (7/40 to 3/41 plus 10 made in April 1942), T16 240 and T9 (airborne) 830 (4/43 to 2/44)

According to the ETO loss reports to 20 January 1945 the US ETO armies had lost 2,855 76/76mm gun Shermans and had 4,561 on hand. As of 20 February 1945 the US ETO armies also had 804

105mm gun Shermans on hand after losing 124.

By 5 May 1945 the losses for the campaign for 1st army were put at 1353 75mm and 307 76 mm Shermans, 3rd Army 642 75 mm and 638 76mm, 9th Army 284 75 mm and 132 76 mm, total 2,279 75mm and 876 76 mm Shermans, grand total 3,055 losses for these three armies.

By the looks of the figures, to 20 January 1945, the 7th army had lost 542 75 or 76 mm gun Shermans.

So as a guess the allies in the ETO would have lost around 6 to 7,000

75/76 mm gun armed Shermans when you add the US 7th, British 2nd, Canadian 1st and French 1st Army losses, with around 6 to 7,000 on hand at the end of the war, so 12 to 14,000 out of a production of 44,554. Then there are the Allied army's losses in North Africa, Italy and the US forces in the Pacific. Plus around 4,100 Shermans sent to the USSR (and 1,400 M3s). The Commonwealth received around 17,000 US build medium (M3 or M4) tanks.

Given the above numbers and all the other uses the allies put Shermans

to, like DD, flame throwers etc, I doubt 19,000 Shermans never left the US, even given the training unit requirements and the 4,369 75/76 and 2,394 105mm gun armed tanks made in 1945 which would have had only a small chance of leaving the US.

As of 5 May 1945 the US 1st army had 18 operational (not on hand) M26 after losing 1, 3rd Army had 43, and 9th Army 47. Production to the end of February 1945 was 242, March 1945 production was 194.

Geoffrey Sinclair<HR></BLOCKQUOTE>

Michael

[ 10-20-2001: Message edited by: Michael emrys ]

[JasonC]

Yes they are interesting, thanks for posting them - but also a little confusing in spots. There were a couple places where it says "76/76mm", which may be typos for 75/76, meaning all Shermans besides the 105s. Or that may mean just the 76mms. I assume it is the former.

The loss figures of around 2850 through the end of the Bulge for the US alone, seem on the low side to me, but may be correct, especially if it is meant to be a battle loss figure. They don't fit the later army by army enumeration very well, though, unless anybody thinks the whole US army only lost 200 Shermans between the end of the Bulge and the end of the war. 2 per day? Seems pretty unlikely. I also wonder if these figures are battle losses only.

As for the overall accounting for US produced tanks, both from those figures and from others I have seen before, I think an outline of them goes like this.

22.5 K Stuarts, 1K sent to Russian and 7.5K to the UK. 14K for US forces, some in the far east, many used for training, enough leftover to account for the 1/3rd of TOE or so taken up by light companies and cavalry formations.

4.7K Chaffees built, but probably far less actually reached the front, perhaps only 1K.

6.2K Grants, 1.4K sent to Russia and perhaps 1.8K to the Brits, 3K retained for the US (used in North Africa, also the far east). Some converted to SPA, recovery vehicles, spotlight platforms, etc after their heyday.

49.5 K Shermans, 4.7K 105s and 10.8K 76s, leaving 34K originally made as 75mm. But 14K of those sent to the Brits, who upgraded up to 4K of them to Fireflies. Of the 76mm, 2.1K sent to the Russians, 1-2K to the Brits, leaving 7-7.5K for the US. Of which about 1K went to the east, leaving 6-6.5 for Europe. However, a portion made in 1945 probably never left the states; perhaps 5K did make it. Of the 75s, 14K to the Brits as mentioned and 2K to the Russians, leaves 18K. As many as 3K east, leaves perhaps 15K for Europe. Of the 4.7K 105s, only a few went to the Brits and a large number were late war and probably never reached the front. Perhaps 2K did reach US service in Europe.

So the US Shermans would be 5K 76, 15K 75, 2K 105. Some of them were used in the Med, not counted as "ETO" in US army parlance. Are some in a never left the states category? Training, maybe some. Some are probably in the depots, including those in various states of breakdown and repair, rather than operational on the one hand or dead on the other, and others reaching the front as upgrades and replacements.

If the loss figures given are for battle losses (likely, from their absolute number), then probably an equal number (about) were lost to breakdowns. Say the 2850 through January 20 figure is correct for battle losses, and losses after that date are 1/3rd again, and breakdowns are equal to combat losses. Then that would account for around 7500 vehicles out of service. That was based on an ETO figure, so perhaps the Med brings it to 10K. Perhaps 7.5K left in the fleets at the end, including the MTO. That leaves 4500 unaccounted for, for training, depots, in the states, etc.

Is that going to be completely accurate? No. But it may be in the ballpark, as an accounting for where all the produced US built tanks ended up. I hope it helps.

[Jeff Duquette]

<BLOCKQUOTE>quote:</font><HR>Jason Said: The loss figures of around 2850 through the end of the Bulge for the US alone, seem on the low side to me<HR></BLOCKQUOTE>

2855 not 2850

Your assertion that this number seems low is based on what? Real information, or speculation?

<BLOCKQUOTE>quote:</font><HR>Apparently from R. Anderson (or a Geoffrey Sinclair interprtation of Anderson)

According to the ETO loss reports to 20 January 1945 the US ETO armies had lost 2,855 76/76mm gun Shermans and had 4,561 on hand. As of 20 February 1945 the US ETO armies also had 804105mm gun Shermans on hand after losing 124.

By 5 May 1945 the losses for the campaign for 1st army were put at 1353 75mm and 307 76 mm Shermans, 3rd Army 642 75 mm and 638 76mm, 9th Army 284 75 mm and 132 76 mm, total 2,279 75mm and 876 76 mm Shermans, grand total 3,055 losses for these three armies.<HR></BLOCKQUOTE>

There is an addition error in the 76mm Sherman losses. Based upon the above figures through 5 May total 75/76mm losses should be:

2279 Sherman 75mm

1077 Sherman 76mm

Yielding total losses of 3,356 75mm/76mm Shermans as of 5 May.

This would imply a loss of an additional 501 75/76mm Shermans from the US Armies in ETO, 20 January to 5 May.

<BLOCKQUOTE>quote:</font><HR>Jason Said 22.5 K Stuarts, 1K sent to Russian and 7.5K to the UK. 14K for US forces, some in the far east, many used for training, enough leftover to account for the 1/3rd of TOE or so taken up by light companies and cavalry formations.

4.7K Chaffees built, but probably far less actually reached the front, perhaps only 1K.

6.2K Grants, 1.4K sent to Russia and perhaps 1.8K to the Brits, 3K retained for the US (used in North Africa, also the far east). Some converted to SPA, recovery vehicles, spotlight platforms, etc after their heyday.<HR></BLOCKQUOTE>

hmmm...

From Army Procurement in the “Green Book” series by the Office of the Chief of Military History

Total medium tanks procured (other than M-4) (note that these procurements include "stocks at depots and with troops on 1 July 1939 and procurement from 1 July 1939 through 31 December 1940"):

M2A1-----------------94

M3 series---------6,258

RAM/M4A5----------1,144

M7--------------------7

T23-----------------250

T25E1----------------40

Of these Britain received 2,887 M3 series, the USSR 1,386 M3 series, and American Republics 104 M3 series. The RAM tanks were US financed, but Canadian built.

Total light tanks procured:

M2A4----------------375

M3 series gas----12,365

M3 series diesel--1,496

M5 series---------8,884

M22 (T9E1)----------830

M24---------------4,731

In addition their were 240 Marmon-Harrington 'light tanks' built.

Of these Britain received 36 M2A4, 5,532 M3 series, 1,391 M5 series, 420 M22, and 289 M24. China received 100 M3 series. The French received 238 M3 and 413 M5 series (2eme DB was equipped with M3 Lights). The USSR received 1,676 M3 series, 5 M5 and 2 M24. American Republics received 653 M3 series and other countries received 50 M3 series.

Total heavy tanks procured:

M6 series------------40

M26, T26E1, T26E5----2,239

T26E2----------------185

Of these 12 M26 went to Britain and 1 to the USSR.

[JasonC]

My sense that it is low is based on unit histories, German claims, the flow of the battle and the various armor engagements involved.

Some examples - the Germans claimed 3750 Allied tanks in Normandy, while we know they lost 2200 themselves. The Allied losses may be weighted somewhat toward the Brits, but 1250 looks like a lower bound for the US losses there. In the unit histories, typical independent armor battalions working with the infantry divisions in the hedgerows turned over their whole tank strength once in the Normandy fighting. The armor divisions got off lightly by comparison, because most weren't engaged until they broke into the clear. But some of the breakout divisions (e.g. 2nd AD) report battalions losing half their initial strength by the end of Falaise. Both items point toward around 1500 tanks lost, plus or minus 15%.

Now, were the tank losses in Normandy alone more than half of all losses through the end of the Bulge? It seems unlikely, for several reasons. The front more than doubled in length, and the amount of armor in the line increased in proportion. Patton's ADs hadn't fought in Normandy until the breakout. 4 different US ADs saw their first action in the course of Cobra, and 3 more went into action in time for the Bulge. The independent armor battalions in action grew along with the infantry divisions. Then there are particular fights, like those around Metz and Nancy, Aachen and the Hurtgen, which were not costless. For their part, the Germans lost hundreds of tanks in identifiable battles in this period. All before the Bulge.

Then in the Bulge itself saw some of the most intense tank fighting of the war, and most of the show was US. Unit histories from Patton's ADs show half their tank losses for the whole war in their "peak" week in the Bulge. The opposing German armor force was about the same size as the Normandy force (2500 AFVs), and they lost about half of it in the course of the fighting. More with Nordwind/Alsace operations included.

It is hard to see that adding up to less than 1500 dead US tanks. If so, it would mean the US outscored the Germans heavily from Normandy on, in absolute terms, despite the armor quality differential and the fact that the attacking US tanks were facing more in the way of PAK, fausts, etc. Post Bulge, outscoring is believable due to odds and a collapsing front, but my impression from the histories is that the fall 44 through Bulge fighting was not so one sided.

The figure might be more believeable if thought of as "total write offs due to enemy action", meaning net of recoveries and in repair cases on the one hand, and not counting breakdown losses on the other.

On the tank LL totals, that is useful stuff again. It shows about 1K more Grants sent to the Brits than my rough estimate, and a few more Stuarts scattered hither and yon, but the basic picture remains.

A good discussion BTW...

[Michael Emrys]

It seems to me that in pondering numbers like these a few things need to be kept in mind. First of all, the fighting in Normandy was at close quarters in terrain that was mostly disadvantageous for armor. Also, the Allied armies, were still a ways down on the learning curve in applying combined arms doctrine. All of this led to higher losses than the pursuit across France on the heels of a broken, fleeing enemy. By the time the war had reached the borders of Germany and the German resistance had stiffened, the Allies had a beter idea of how to employ their forces.

Jason has already mentioned recovery and repair. I think this is critical. Most knocked-out tanks can be patched up and put back into the fight, but only if the owning army retains possession of them. A great many of the German losses in the Bulge were not battle casualties, but simply ran out of gas and were abandoned. Added to that were the actual losses due to enemy action. They all became permanent losses. Since the Alllies retained possession of the battlefield in the end, this was not so much of a problem for them. Therefore, it is possible to reconcile the apparent contradiction between German claims for kills and the actual permanent losses suffered, in this case, by the Americans. Aside from German over-claiming, there were the tanks recovered and repaired.

Jason has also mentioned losses due to breakdowns. I admit to being puzzled by this. My assumption has been that as long as the owning army retains possession of the vehicle in question, it can be repaired and returned to action. It doesn't seem to me that the Americans should have permanently lost many tanks to breakdowns. But I am open to persuasion if someone has documented numbers.

Michael

[JasonC]

"Aside from German over-claiming, there were the tanks recovered and repaired."

That is plausible. It still leads to a rather startling conclusion, that overall US tank losses for the war were within 3:2, and perhaps even, with the losses they inflicted on the Germans. Despite the armor quality differential and the greater propensity of many of the US types to burn. It is possible the tank recovery and maintenance questions, between them, outweigh any edge the Germans had in battlefield KOs. Not the usual picture of the armor war presented, but it might still be true.

"Jason has also mentioned losses due to breakdowns. I admit to being puzzled by this. My assumption has been that as long as the owning army retains possession of the vehicle in question, it can be repaired"

I don't think that is accurate in practice, because of cannibalization. It was very common practice to use vehicles that had broken down as standing parts supplies for others better off, needing comparatively minor repairs, etc. The tanks stripped were effectively written off, because replacing all the parts taken from them just didn't happen.

Also, my impression from unit histories is that US units drew replacement tanks as needed to replace losses and serious breakdowns, rather than nursing every tank as the Germans sometimes did, especially with the heavier makes. The focus was on keeping the unit near TOE in runners. This really reflects a different organizational and "army legalistic" attitude about equipment in the US and the German armies, which tied into replacement practices and all sorts of other things.

A German armor unit that has been in action a while quickly gets to a state where half its tanks are in the repair categories, while the overall number of tanks is shrinking below TOE. Half the fleet are "runners" after the first heavy action. Then both halves slower melt, as repairs almost keep pace with KOs and breakdowns but not quite. The runner TOE is often 25-50% of the nominal establishment strength, dealt with by leaving some elements out of battle and reducing platoon sizes.

All of which reflect the fact that the particular unit "owned" its issued tanks, and wasn't likely to get more short of a major overhaul, usually in the rear, that would bring the whole unit to full strength again. (I am talking about armor and SS branch forces). As strength dropped, the unit might be asked to do less on the front, but often it just got harder to do everything, until eventually the burn out was serious enough that the unit would be disbanded, refitted in the rear, or rolled into an ad hoc KG still capable of doing something.

US practice was quite different. You only see strengths down at 50% of TOE immediately after an engagement where the unit took serious losses, and the condition is temporary - typically a week or two long at most. Most of the armor battalions in the armor divisions are near 90% of TOE on other occasions. The independent battalions, with less overhead and clout for replacement and supply, are typically around 80% of TOE.

When a tank can be repaired rapidly by the battalion or division shops, it is returned to service quickly - or a portion of those lost are, cannibalizing the rest with little regard for their eventual repair. The rest of the shortfall is requesitioned, and typically the unit tops off after its next breather in action. Often the problem is supplying trained crews for the replacement tanks, not replacement tanks for surviving crews.

The front line unit is thought of as a conduit for a continual flow of weapons, men, and ammo, rather than an organization issued a certain number of weapons and personnel and then left to fight. In this situation, there is little reason to bend over backwards to return every wreck to service. There are plenty of additional tanks flowing to the front. Short term repair cases will just be back in action faster. The hard limit over a month or two time scale is going to be crews, not tanks.

When you look at QM reports from particular US armor battalions, you find between half and a third of their permanent losses were maintenance related. E.g. the unit history will record 40-50 tanks lost in combat, and the QM will report they drew ~75 new tanks over the course of the war, and ended it 5 tanks under establishment strength. Meanwhile crew casualties will run between half the strength of the unit and full strength, non-battle as well as battle, with a fifth of them or so returned to service, replacements arriving for half to two thirds, and ending ration strength anything from establishment to down a third.

These are based on only a smattering of unit histories, not an exhaustive accounting, so obviously there is room for error in any conclusions drawn from them. What is the wags line? One story is an anecdote, two stories are data. But I think it makes the reasons clear enough. Tanks weren't scarce but crews were. New tanks were reaching the front. So returning every major malfunction to service just wasn't critical - there wasn't anybody to crew it who couldn't get a new tank instead.

I hope this is useful.

[Michael Emrys]

Thanks for the illuminating reply, Jason. Makes sense.

One point I forgot to mention in my earlier post is that there are more ways to kill tanks than with other tanks. I think there is a consensus that the Germans for the most part had somewhat better tanks. I think a case can be made that they also had better AT guns as well. But they didn't have such massive amounts of artillery or air superiority available to them. I am unable to say how far those may have tilted the statistics, but they are a factor that must be considered also in evaluating and comparing the performance of the two armies.

Michael

[Slapdragon]

One fact that has always been interesting to me is that the Britishbuilt 24,000 tanks, versus 24,000 built by the Germans. If the US had never entered the war then the both of these two sides would have had comparable numbers if not quality. Then throw in 40,000 Russian AFVs and 100,000 American AFVs and you can see why every German squad had a tank killer in it.

[Jeff Duquette]

<BLOCKQUOTE>quote:</font><HR>Jason Said:

That is plausible. It still leads to a rather startling conclusion, that overall US tank losses for the war were within 3:2, and perhaps even, with the losses they inflicted on the Germans. Despite the armor quality differential and the greater propensity of many of the US types to burn. It is possible the tank recovery and maintenance questions, between them, outweigh any edge the Germans had in battlefield KOs. Not the usual picture of the armor war presented, but it might still be true. <HR></BLOCKQUOTE>

Gee I wonder where you got that idea from. Actual German irrecoverable loss figures portray a large number of Panzer casualties resultant from either abandonment or the crew destroying the vehicle. These are vehicles without spare parts and vehicles without gasoline that are subsequently recovered by the Allies.

Conversely Anglo-American tanks losses in NW Europe as well as Italy represent predominantly vehicles destroyed in combat…tank or anti-tank gunfire, panzerfaust\panzerschreack, mines, and a small number of losses to German artillery fire.

This has nothing to do with a lack of efficiency on the part of Germany repair & recovery. If one examines any operation or campaign in which one side or the other is forced to retreat, the retreating side will always incur far more irrecoverable tank losses. What the real numbers should tell us is that the German Army operational Panzer Strength at any given period of an ongoing operation\campaign might be running at anywhere from 50% to 25% of total strength within only a few days of the commencement of the operation. The remaining portion of total Panzer strength was being continuously cycled through maintenance units or sent back to depot facilities for more serious repair. The problem with the German Army was it was under an acute shortage of spare parts for its front line Panzers from about the time of Barbarossa all the way through the surrender. Production capacity was seemingly always focused on the production of new AFV’s rather than production of spare parts to keep existing machines running.

So although this paints a grim picture with regards to keeping existing panzers running, it also reflects on the actual accomplishments of those tanks and crews with operational vehicles. So although some might say gee “overall wartime loss statistics show the German tanks weren’t all that hot” the actual tactical picture is one of “many panzers were not making it to the battlefield, but those that did managed to inflict a huge amount of punishment on Allied tanks”.

By the way the American Tank loss figures originally derived from Richard Andersen represent irrecoverable tank losses (R. Anderson is one of the co-Authors\researchers for “Hitler’s Last Gamble”).

<BLOCKQUOTE>quote:</font><HR>Slapdragon Said: One fact that has always been interesting to me is that the Britishbuilt 24,000 tanks, versus 24,000 built by the Germans. If the US had never entered the war then the both of these two sides would have had comparable numbers if not quality. Then throw in 40,000 Russian AFVs and 100,000 American AFVs and you can see why every German squad had a tank killer in it.<HR></BLOCKQUOTE>

Soviet tank\assualt-gun production figures approached 100,000 during the war. In addition at the start of Barbarossa the Red Army had 20,000+ tanks. Soviet Irrecoverable tank\assault-gun loses for the war was 83,000!!!...not including lend-lease irrecoverable tank stats. So yes I agree it is easy to see why every German Infantry Squad probably had a tank killer in it.

[ 10-22-2001: Message edited by: Jeff Duquette ]

[Andrew H.]

These production numbers are interesting. I wonder whether they are really inconsistent with battalion reports though. That is, if a battalion counts, say, 5 tanks are write-offs and draws five new tanks, does that really mean that these tanks are write-offs, or are they only write-offs as far as the battalion is concerned.

In other words, might these tanks be overhauled by a higher level rear echelon repair park, sent to a replacement depot, and eventually reissued to another unit.

The same would be true wrt to cannibalized tanks: the local unit might use the tank as a source of spare parts for a while, but is it possible that the cannibalized tank would eventually be taken to a rear area and refurbished, and then sent to a different unit.

It wouldn't necessarily make military sense for a tank battalion with several severely damaged tanks to pause its advance until the tanks can be sent to a rear area, repaired, and returned to the battalion.

[Michael Emrys]

<blockquote>quote:</font><hr>Originally posted by Jeff Duquette:

Actual German irrecoverable loss figures portray a large number of Panzer casualties resultant from either abandonment or the crew destroying the vehicle. These are vehicles without spare parts and vehicles without gasoline that are subsequently recovered by the Allies.

Conversely Anglo-American tanks losses in NW Europe as well as Italy represent predominantly vehicles destroyed in combat…tank or anti-tank gunfire, panzerfaust\panzerschreack, mines, and a small number of losses to German artillery fire.<hr></blockquote>

All this was noted earlier.

Michael

[Michael Emrys]

<blockquote>quote:</font><hr>Originally posted by Andrew Hedges:

In other words, might these tanks be overhauled by a higher level rear echelon repair park, sent to a replacement depot, and eventually reissued to another unit.

The same would be true wrt to cannibalized tanks: the local unit might use the tank as a source of spare parts for a while, but is it possible that the cannibalized tank would eventually be taken to a rear area and refurbished, and then sent to a different unit.<hr></blockquote>

This was my thought as well, but whether this was the actual practice or not I do not as yet know.

Michael

[JasonC]

There is one major problem with Slap's story. The Germans didn't make 24K tanks, they made more like 50K. I don't know where the other figure comes from, but it would be reasonably close for turreted AFV only, during the war as opposed to before, and in Germany as opposed to Czech production.

The total for turreted AFVs alone was 30K, including pre-war and Czech vehicles. Might lower it to 28.5K leaving out the obsolete Pz Is, I suppose, and just count gun-armed AFV. Then there were 10.5K StuG, 5.5K armored TDs, 2K StuH and Brummbar, 3K SPAT (Marder and Nashorn). All identifiably what we mean by "AFVs". Total 49.5K.

It is probably best to leave out the 2K SPA and 1K SPAA. Then there is the light armor, 5K armored cars and 22K halftracks, but those aren't "real" AFV. But they do amount to another 30K armored items besides the AFVs counted above.

The total "production odds" the Germans faced over the course of the war were no higher than 5:1. And before the second large front opened in France, the production odds were only about 3:1 (Russians plus LL plus a few in the Med, vs. all German production).

As for the comment that retreaters always have more total write offs, it does not withstand scrutiny. The Russians had higher tank losses whether retreating or not, and possession of the battlefield afterward did not result in only even losses for them, compared to what they inflicted on the Germans. They had odds, too, and a better mix of tanks in armor terms, and late war in high end guns too. If the US only lost as many as they KOed, they were outperforming the Russians by a long way, despite their lighter mix.

It is a pecularity of the Germans that they waited until a retreat forced abandonment of workshops before they wrote tanks off. They did this because tanks were scarce and for the "unit ownership" sorts of reasons I already mentioned. Most of the tanks the Germans listed as "long term repair" would have been called total losses in any other army (except the Finns). And a lot of them probably never saw action again after reaching that category. They were used for parts and lost when the parent retreated, since there was no way to get them out. The other nations were not that hard up for parts, or tanks.

The apples to oranges nature of the comparisons this sets up disappear in a long enough time scale, because a tank that can't be got running again reasonably fast can't run when the front moves, which it always will in a moderate length of time (months).

But short term, it does mean total losses aren't talking about the same thing. A German tank would be listed as "abandoned" if it had 16 holes in it and a third of the parts stripped off. One knocked out in combat would be listed as "KOed by crew" if the hit hadn't made it burn, but the crew did. If you read accounts of particular fights, you find lots of tanks "blow up by crew" after they ran over an AT mine, etc. Everybody else would just list what caused the bail out.

As for the story that the few tanks were doing incredible things, I am aware of definite cases in Russia where that is accurate, and it might with some plausibility be applied to a case like Goodwood, I suppose. But the incredible things are less in evidence in the relative performance implied by the US loss numbers, if those are accurate. I'm not sure they are, but they are saying something that is not conventional wisdom.

The Germans lost around 1000 AFVs to the Americans in Normandy (slightly more vs. the Brits), at least 300 around Nancy, and 2000 more in the Bulge and Alsace. Not counting the Hurtgen-Aachen campaign, etc. If the US only lost 2850 Shermans in the same period, even allowing for TD losses too, the overall loss ratio would be around 1:1. And if the US only lost 500 more Shermans in the rest of the war, then the loss ratio for that part was probably above 1:1.

This is not the usual picture of vastly superior armor worn down by overwhelming numbers despite much higher kills. That might be true enough against the Russians, but it won't fit the loss totals given above. Either there is something wrong with those numbers (they are leaving something out e.g.), or somebody has been projecting things that were true vs. T-34s, and may have been true through Goodwood for the Brits for all I know, onto the American case.

See, if the Americans were doing no better than the Russians, you'd expect that loss figure to read 10,000 tanks. The Americans made enough to win even if the loss total was 10,000 tanks; the logic of attrition moving the current odds in favor of the guy who can replace losses faster would still have been enough. But if the real figure is more like 3000 tanks, then the US didn't just make more. It also lost only the same number.

Which would be news. It would mean having odds, having air superiority, good combined arms, better maintenance and recovery, more spare parts, not making boneheaded operational moves out of strategic desperation, the line moving your way, etc, were fully worth the quality edge the German tanks definitely possessed. And that the superior replacement rate was gravy or overkill, rather than cause.

I don't know if that picture is right, because I don't know if the loss figure is right. I already gave my reasons for doubting it, but those are based on maintaining my previous estimate of relative losses (which I had assumed were on the order of 2:1 for the westerners, vs. 3:1 for the Russians). If the loss figure is right, then US armor did rather remarkably well, not with "a few tanks" (like the Germans in Russia), but with "cheap" ones.

For what it is worth.

[Michael Emrys]

It will be interesting to see how this discussion progresses. While all the numbers that have been bandied about could surely stand some refinement and wise interpretation, the picture that is emerging reinforces my gut feeling that the US Army has sometimes been undersold as far as its performance in the ETO is concerned. As I posted earlier, there are lots of ways to kill tanks, just as there are to avoid having your own tanks killed. While the Americans often had a lot of learning to do, it seems clear to me that by the end of 1944 (rather sooner in the case of some formations) they had worked out ways to function efficiently and get the most out of their weaponry.

By contrast, the Germans had suffered heavy attrition among their most experienced formations, and though they continued to fight doggedly and well, their ranks must have inevitably have been diluted with the less able. I don't mean by this that the Panzerwaffe ever became a second rate force, but that it had a distinctly leveling effect, especially when comparing it to the Western armies.

Michael.

[Tero]

Originally posted by JasonC:

Most of the tanks the Germans listed as "long term repair" would have been called total losses in any other army (except the Finns).

Beggars can't be choosers. Waste not, want not.

The entire Finnish WWII armoured force was built around repaired captured Red Army tanks and a few Vickers 6tons (which were renamed T-26E and armed with guns from those Soviet BT's which could not be salvaged before our troops had to pull back or which could not be repaired to running condition). There were all in all over 144 ex-Soviet tanks in the Finnish inventory.

Only 8 Finnish Stugs were lost to the enemy, all the damaged vehciles that had to be salvaged were taken back. After the war they served until 1966 as runners and after that into the late -80's as static gun emplacements. A sizable portion of the worlds Stug-III runners are ex Finnish Army vehicles.

There were 10 T-26's repaired to more or less running conditions (one was powered with IIRC a Ford motor and a few were towed behind by those who could run under their own power) for the movie Winter War.

The other nations were not that hard up for parts, or tanks.

Our airforce had to keep in the air Soviet, British, American, French, German and Italian aircraft.

If you read accounts of particular fights, you find lots of tanks "blow up by crew" after they ran over an AT mine, etc. Everybody else would just list what caused the bail out.

There is a real diffrence. The self-destruction was a part of their SOP. In case the crew had to abandon the vehicles and there was no way to salvage it they were under orders to demolish it.

But the incredible things are less in evidence in the relative performance implied by the US loss numbers, if those are accurate. I'm not sure they are, but they are saying something that is not conventional wisdom.

The "problem" here is the fact that the numbers represent write offs reported up the chain of command.

An example (source Rick Andersons XLS sheet on weekly losses):

1st army, M4(75's)

9-15 Jul 1944, 7 days, 913 available, 75 reported lost

16-22 Jul 1944, 7 days, 1 102 available, 33 reported lost

From previous period the number of available rose 189 vehicles

23-29 Jul 1944 7 days, 748 available, 79 reported lost

From previous period the number of available fell 354 vehicles.

The total number of vehicles reported lost during those 3 weeks was only 187.

Disregarding any large formation transfers out of the 1st army during that period there were clearly vehicles "in the system" that were neither available for duty nor were they write offs.

The daily in's and outs of combat do not show but at the weekly level it can clearly be seen that things were not all hunky-dory for the US tankers.

This kind of fluctuation would explain the "inflated" German kill claims that are not supported by official US Army loss reports.

If the US only lost 2850 Shermans in the same period, even allowing for TD losses too, the overall loss ratio would be around 1:1.

Hmmmmmmm..... in addition to those 2 855 (the number as of May 12th 1945 was 3 139) there were:

M4(105) total write offs 124, TD (M10, M18, M36) 631, light tank (M5, Chaffee) 1 069, M8 AC 672, M20 AC 364, M8 AG 175, M7 SP 182, M32 recovery 114 (interestingly most of the M32 losses were sustained when the US Army was on the move ie. after the break out and again during the Bulge).

Your premise also disregards the facts that nearly all the German vehicles were write offs while the US losses were not all write offs.

And if the US only lost 500 more Shermans in the rest of the war, then the loss ratio for that part was probably above 1:1.

The write-off ratio can well be close to 1:1 but I think it is safe to say the kill ratio was in favour of the Germans.

This is not the usual picture of vastly superior armor worn down by overwhelming numbers despite much higher kills. That might be true enough against the Russians, but it won't fit the loss totals given above. Either there is something wrong with those numbers (they are leaving something out e.g.), or somebody has been projecting things that were true vs. T-34s, and may have been true through Goodwood for the Brits for all I know, onto the American case.

I think it is more a case of drawing conclusions from data compiled after the war and which was never toutched again when Cold War set in and all things American just had to be better than all things Soviet. Including the war record.

That is why German sources on the war were accepted at face value as they invariably spoke highly of the Americans while they painted a belitteling picture of hordes of formidable Soviet masses of tanks and infantry. It could be concluded that since we (the Americans) beat the Germans singehandedly all our stuff was better than the Soviet stuff since we had the irrevocable statistical proof.

See, if the Americans were doing no better than the Russians, you'd expect that loss figure to read 10,000 tanks.

Indeed. But how many were written off and how many were KO'd and repaired ? The total US AFV write off's were 6 470 (for the M4 3 263 alone) so how many were KO'd in combat and returned to active duty ?

It also lost only the same number.

Of that one type. The other AFV types do not count ?

If the loss figure is right, then US armor did rather remarkably well, not with "a few tanks" (like the Germans in Russia), but with "cheap" ones.

Hmmmmmm..... did the US armour face the German armour alone and kill ALL the German AFV's in the west ?

[ 10-23-2001: Message edited by: tero ]</p>

[JasonC]

The addition of the other types does not change the picture seriously. The 631 TD losses and 124 105s are actually less than I expected for the other mediums (I thought the TDs would be 1000, based on their portion of the force), and bring the total through the Bulge only to 3600. The US KOed permanently, by German loss records or unit histories, 3300 German mediums and heavies in just Normandy, Nancy, and the Bulge. That leaves out the rest of the Lorraine campaign, the whole southern sector (split with the French of course), and the Aachen-Hurtgen fight, throughout the fall. There are going to be more than 300 in all those.

Throwing in SPA and armored cars, let alone recovery vehicles, doesn't mean very much. The German figures above start at the StuG and Pz IV and go upward - they do not include their AC, SPA, Flakwagens, gun armed halftracks, yada yada. And no, I am not talking about all German losses in the west, just the ones lost in identifiable battles against US forces. The Brits got more than half the bag in Normandy, but not much more, once Cobra and Mortain are included. The breakdown there is around 1200 by Brits, and 1000 by Americans, 300 survive - based on where units were engaged when and their tank strength reductions when.

Then 1 Panzer and 2 Panzergrenadier divisions, plus 2 new full strength Panther brigades, were reduced to less than 50 survivors by the Nancy area fighting. Some of the German formations (but not the new Panther brigades) were below TOE, but the net reduction was at least 300 AFVs permanently lost. The losses in the Bulge and the sequential Alsace counterattacks, both of which fought mostly Americans (incidentally help from the Brits in the first, and the French in the second), ran to 2000 AFV.

Nor is it obvious, though it is possible certainly, that American tanks got knocked out more each, before being written off. The Germans were the ones that kept anything still resembling a tank ("waste not"), and got tanks back into operation that the other major powers would have ditched. The US types also burned more easily, and were more often penetrated catastrophically rather than marginally.

The average KOs per tank before complete loss might be something like 1.25 vs. 1.5, or 1.5 vs. 2 I suppose, but that is only going to make a small difference. And actually, the real figure is usually more like 1, because for every tank KOed twice there is one that breaks down mechanically without being KOed.

Also on the other side, the losses to PAK, AT mines, and fausts were high on the attacking US side, probably far higher than losses to US arty or air (the last is vastly overstated in Allied claims, and clearly a small factor for tanks themselves in German records. Trucks and such were another story there, of course).

If the US losses of mediums are really only ~3500 through the Bulge, then claims of high kill ratios per tank, by the Germans, just can't hold up. If to support such claims one has to trot out how many Stuarts (were killed by Panthers?) just to get the results up near 4 to 3 - ignoring the late war when it is probably worse - the stretch taffy is getting pretty thin.

Somehow the "5 Shermans to beat one Panther" stories don't sound so plausible when the Panthers are getting 1 Sherman and 1 Stuart, and everything else is getting 1 Sherman or TD, before being got itself. The Germans really did kill ~3 T-34s for each tank of their own they lost (although it wasn't just their tanks doing it - another story). If they didn't kill anything like 7500-10000 US mediums, then they didn't score 2-3 to 1 against US Shermans, even though the Shermans were a weaker tank technically. If true, that is news.

Oh, I also meant to address the point about week by week tank strengths at the army level. No, you cannot assume all net reductions in tanks in a whole army's returns are KOed tanks, because armies are meant to pass around subordinate units like divisions. In the week in question in the figures you gave - which is the week of Cobra, the end of July - US 3rd Army was activated.

[ 10-23-2001: Message edited by: JasonC ]</p>

[Jeff Duquette]

<blockquote>quote:</font><hr>Emrys Said:

All this was noted earlier.<hr></blockquote>

Yes I have mentioned these things before…in several past threads. Thus the “I wonder where you got that idea from”. Did you note the part where the original poster of your tank casualty figures did not add correctly?

<blockquote>quote:</font><hr>The total "production odds" the Germans faced over the course of the war were no higher than 5:1. And before the second large front opened in France, the production odds were only about 3:1 (Russians plus LL plus a few in the Med, vs. all German production).<hr></blockquote>

Only 5:1…Only 3:1?? Those sound like really light odds. Jason when the Germans went into Russia in 41 the odds were more like 5.5:1 in tanks\assault guns…in the Red Armys favor….and the Germans were attacking!!! Didn’t you imply in an earlier thread that the British held an "overwhelming odds" advantage at El Alamein with only a 2:1 numeric advantage over the Germans (or did you say 3:1...I cant recall)? If only the search engine worked here.

Just to get this straight you are implying that before the large front opened in France and Italy (say 1939 - 1942?) the Germans typically held the initiative and were in an attacking mode even though they had 1:3 numeric disadvantage in tanks\assault guns. And to further clarify you are saying that only when the odds started pushing to 5:1 do the Allies gain the initiative?

<blockquote>quote:</font><hr>As for the comment that retreaters always have more total write offs, it does not withstand scrutiny.<hr></blockquote>

You must be kidding. I would love to see you produce some real numbers on this one.

[PzKpfw 1]

<blockquote>quote:</font><hr>Originally posted by JasonC:

22.5 K Stuarts, 1K sent to Russian

6.2K Grants, 1.4K sent to Russia

49.5 K Shermans, 2.1K sent to the Russians, .<hr></blockquote>

Total US tanks shipped Lend Lease to the USSR:

M3A1 Stuart - 1,674

M5 - 5

M24 Chaffee - 1

M3A3 Grant - 1,386

M4A2 Sherman 75mm - 2,007

M4A2 Sherman 76mm - 2,095

M26 Pershing - 1

Of these numbers, many were lost in transit, Ie, 443 M3A1 Stuarts as well as, 417 M3A3, & M4A2 mediums etc, were lost.

Regards, John Waters

[ 10-23-2001: Message edited by: PzKpfw 1 ]</p>

[Michael Emrys]

<blockquote>quote:</font><hr>Originally posted by Jeff Duquette:

quote:

------------------------------------------------------------------------

Emrys Said:

All this was noted earlier.

------------------------------------------------------------------------

Yes I have mentioned these things before…in several past threads. Thus the “I wonder where you got that idea from”.<hr></blockquote>

Surely, Jeff, you are not going to claim that that idea is original with you? Or are you trying to imply that I have never read anything on the subject but what you have posted (and I frankly do not recall)?

Michael

[Andreas]

Jason, your figure of 300 German AFVs lost at Nancy seems quite excessive, what is it based on, if I may ask? von Mellenthin does not seem to remember 300 AFVs in the German force (although he is hazy on the numbers - he does not remember a second PzGr Division for that matter too. Which one was that?), and I am pretty sure that von Luck (copy waylaid at the moment) does not remember a lot of AFVs (if any) in 21st Panzer there.

Like Jeff, I would like to see the reasoning behind your statement that being on the retreat does not affect AFV write-offs, since it does not square with what I have read.

[Tero]

EDIT: Important I just noticed the weekly figures are for M4(75) only. I'll post the total medium figures posthaste.

Originally posted by JasonC:

Nor is it obvious, though it is possible certainly, that American tanks got knocked out more each, before being written off.

A kill racked up by the Germans could well have been a penetrating hit that did not do any serious damage but it did make the crew bail out.

The Germans were the ones that kept anything still resembling a tank ("waste not"), and got tanks back into operation that the other major powers would have ditched.

I am not convinced this is true. For one there are no records available to show how many Allied tanks were shipped over as replacements after the break out from Normandy was effected.

The high mark of American M4's reported as available for combat was 4561 in 20 Dec 44-20 Jan 45. 20 Oct-20 Nov 44 there were only 2832 reported available for combat and in 20 Nov-20 Dec 44 the number rose to 4076. (The low mark was 20 Jun-20 Jul 44 with only 2093 available for combat.)

The average available number of M4's a month was 2901. With total losses of 2855 the turn over rate was 98,41 %.

The highest loss percentage was 20 Jul-20 Aug 44 when the losses amounted to 21,78 % (2557 available, 557 lost). During the Bulge (going by the monthly periods 20 Nov-20 Dec 44 with 4076 available and 495 lost and 20 Dec 44-20 Jan 45 with 4561 available and 585 lost) the total losses during the Bulge were 1080 but the loss percentage was "only" 12,49 (average with montly percentages of 12,14 and 12,83 respectively).

I do not know how many new amroured formations were deployed to ETO between the periods of 20 Oct-20 Nov 44 and 20 Nov-20 Dec 44 but the number of M4's available for combat rose by 1244 vehicles (by ~44%). If we count 53 M4's in a medium battalion that would make around 23 new full strenght medium battalions. Or most of the formations were running at 50% strenght all the time until 20th Nov. When was the US advance stopped again ?

I seem to remember that at some point the ETO was in jeopardy of running dangerously low on serviceable medium tanks. Go figure.

The US types also burned more easily, and were more often penetrated catastrophically rather than marginally.

I saw a stat on that not long ago in a board somewhere but for the life of me I can not remember where. Anyway the stat showed that not all M4's KO'd by Pzfaust or Pzschreck had burned. The figure was suprisingly low.

And actually, the real figure is usually more like 1, because for every tank KOed twice there is one that breaks down mechanically without being KOed.

How will that broken down be listed in the stats ?

Also on the other side, the losses to PAK, AT mines, and fausts were high on the attacking US side, probably far higher than losses to US arty or air (the last is vastly overstated in Allied claims, and clearly a small factor for tanks themselves in German records. Trucks and such were another story there, of course).

That does not make them any less KO'd or written off. This is the distiction that has to be made when looking at the statistical figures.

If the US losses of mediums are really only ~3500 through the Bulge, then claims of high kill ratios per tank, by the Germans, just can't hold up.

Are the German claims made on battle by battle basis and not by the big sweeps you prefer to use ?

If to support such claims one has to trot out how many Stuarts (were killed by Panthers?) just to get the results up near 4 to 3 - ignoring the late war when it is probably worse - the stretch taffy is getting pretty thin.

Can you give the breakdown of the German AFV losses you use ? That would resolve if there were any "secondary" AFV's counted in the German list.

If they didn't kill anything like 7500-10000 US mediums, then they didn't score 2-3 to 1 against US Shermans, even though the Shermans were a weaker tank technically. If true, that is news.

You are in a pit here. You accept the German 3 T-34 for 1 German tank without any regard to the fact that any number of the T-34's could have been recovered and used again (which they IRL did) and killed again (which did happen). Then you disregard any other German kill against the Sherman and stick to the number of actual write offs. Nearly 50% of all written off US AFV's were M4's. The recovery units in US armoured formations did exists. What did they do if they were not recovering battle damaged tanks ? These recovery formations were sustaining losses themselves when the Germans were retreating. Did they lead the units into mined roads ?

In the week in question in the figures you gave - which is the week of Cobra, the end of July - US 3rd Army was activated.

You choose not to disclose which units exactly were transferred from 1st Army to the 3rd Army. Why ?

Were all the armoured units in the 3rd Army ex-1st army ?

1st army:

6 Jun-1 Jul 1944, operational 764, lost 187 (24%)

2-8 Jul 1944, operational 867, lost 21 (2,5%)

9-15 Jul 1944, operational 913, lost 75 (8%)

16-22 Jul 1944, operational 1102, lost 33 (3%)

23-29 Jul 1944, operational 748, lost 79 (10,5%)

30 Jul-5 Aug 1944 operational, 656, lost 68 (10%)

6-12 Aug 1944, operational 580, lost 76 (13%)

13-19 Aug 1944, operational 808, lost 39 (5%)

20-26 Aug 1944, operational 865, lost 19 (2%)

27 Aug-2 Sep 1944, operational 865, lost 35 (4%)

3rd Army:

1-11 Aug 1944, operational 301, lost 35 (11%).

12-18 Aug 1944, operational 794, lost 101 (13%).

19-26 Aug 1944 operational 782, lost 48 (6%)

27 Aug-2 Sep 1944, operational 780, lost 37 (5%)

3-9 Sep 1944, operational 614, lost 2 (0,3%)

9th Army

9-16 Sep 1944, operational 94, lost 2

17-20 Sep 1944, operational 94, lost 0

21-28 Sep 1944, operational 40, lost 0

29 Sep-5 Oct 1944, operational 79, lost 0

6-12 Oct 1944, operational 84, lost 0

13-20 Oct 1944, operational 81, lost 0

21-28 Oct 1944, operational 319, lost 19 (6%)

If all these big changes in the numbers of operational M4's can be explained with troop transfers then somebody better start explaining where did all those 1244 extra Shermans appear from in Nov 20th to the ETO figure.

[ 10-24-2001: Message edited by: tero ]</p>

[Michael Emrys]

One thing that has not been factored in is the quantity of armor in the 7th. Army. I'm sure they had tanks and that they had some shot out from under them. But the post I began this thread with only accounts for the 1st., 3rd., and 9th. Armies.

Then there is the armor in the French army.

Michael

[Tero]

Originally posted by Michael emrys:

One thing that has not been factored in is the quantity of armor in the 7th. Army. I'm sure they had tanks and that they had some shot out from under them. But the post I began this thread with only accounts for the 1st., 3rd., and 9th. Armies.

Yes. I'm using the same source.

Then there is the armor in the French army.

And the British, the Canadian and the Polish armies.

[JasonC]

"Only 5:1...Only 3:1??"

Yes. Slapdragon commented, in this thread (see the 8th post) that the Germans were outproduced 10 to 1 in tanks, which is false. He thought the Germans only had 24,000 AFVs over the course of the war. So it is necessary to point out the real figure. The issue under discussion, of course, is what the relative tank loss ratios were for the different fronts and forces and how it fits with tactical stories we are commonly told, of 5 vanilla Allied AFVs needed to tackle each German one, or of kill per vehicle claims of 5-10 for plain vanilla German AFVs, and higher ones for their heavies, like 17 per afternoon.

I've argued extensively in the past, on this board, that those tales cannot be literally true of average performance - though they may well be true of some portion of outliers. And the reasoning is simple enough. If you establish and maintain a 5 or 10 to 1 kill ratio and are only outproduced 3 to 1 or 5 to 1, you aren't going to run out of tanks and lose. On this basis and from other data and third party estimates, I have previously argued that the overall loss ratios probably ran around 3 to 1 on the eastern front (perhaps as low as 2.5), and probably around 2 to 1 on the western (perhaps 1.75).

That then fits a whole picture about how the Germans lost the war. They didn't get KOs in the ratio of production, but less than it. Which meant the higher the absolute loss rate went, the higher the odds would go in forces fielded against them. This subtraction effect of attrition fits the overall flow of the war, and can explain simultaneously the Germans running out of tanks, facing increasing odds, and losing (1) while still KOing far more tanks than they lost (2).

The fact that this basic explanation - call it "the attrition theory" if you like - explains both things is its main attraction. Explanations that play up only one of them often fail to explain the other half, and can thus be falsified. E.g. "Every vanilla German AFV KOed 5 Allied ones"; call it the uberpanzer theory. It fails because it cannot explain why the Germans lost. It waves the hands and says "numberless hordes". As one Korean vet said, how many hordes in a Chinese infantry platoon? The hordes are not numberless, and the degree of outproduction actually faced is not simultaneously compatible with uberpanzer type claims about kill ratios and the fact that the Germans lost.

And what is happening in this thread is Michael is presenting evidence that for the US, the loss ratio may have been more like 1:1, or perhaps 3:2 if you count all the light armor. Which is different from the 2:1 figure I believed previously - which I had already encountered significant doubts about from uberpanzer boosters, who think it would be higher still for "lousy" Shermans than for "revolutionary" T-34s. Which is related to another discussion, about the relative importance of tank specs, and how limited it was on the operational level, compared to doctrine, operational handling, replacement rates, logistics, etc.

Naturally if one achieves only a 1:1 kill ratio, being putproduced at all will eventually get you stomped, let alone being outproduced 5 to 1. But the point "in play" here, as it were, is not the eventual victory of the side with superior numbers. That is something all acknowledge as obvious, even the uberpanzer theorists, who tend to exaggerate the odds the Germans faced, as though they weren't high enough without exaggeration. What is in play here is the issue of relative outperformance per AFV - which is established beyond a doubt in the case of Russia, though the amount is disputed, but which this evidence makes murkier in the west.

Nor is the conclusion inherently unbelieveable, since the western powers had great advantages in odds, air, supply, initiative, etc. These may simply have counterbalanced the technical or tactical advantages of the German armor, leaving the loss rates close. Then attrition applies much more easily - a side with 3 to 1 exchanges 1 for 1, and is left with 2 to nothing and victory. Without a kill ratio to swim against, as it were.

I am not yet convinced the US loss figures are accurate, because I still regard the attrition theory (roughly, 3 to 1 east, 2 to 1 west, not enough against the production odds) as the most plausible overall explanation of the course of the war, including conventional wisdom about typical tactical match ups and performances. I acknowledge, however, that this evidence - near even US tank losses vs. inflicted in ETO - conflicts with my previous view.

I would have expected that number (dead US mediums) to be 7000 to 10000, not 3500. When a theory makes a wrong prediction about a number, there is either something wrong with the number - under discussion here - or with the theory - also under discussion here. The alternate theory is certainly not the uberpanzer one.

It would be something like - equalized doctrine and operational advantages (odds, supplies, air etc) are more important than tank specs, not only for the outcome of month long campaigns (operational success) but also for the losses incurred along the way (cost). Or, those factors are as important as tank spec match-ups, in predicting tank losses. Call it the "operational advantage" theory.

I am stating that theory rather than subscribing to it, because I still find it hard to believe the US only lost 3500 mediums. But I acknowledge that if that number were established as accurate, it would point away from the attrition theory and toward the operational advantage theory. Or, put in popular fashion, the operational advantage theory would say, "the Russians might have been wearing the Germans down at terrific cost (like in the attrition theory), but the westerners just plain beat them".

As for my earlier discussion of El Alamein, I said the British won through superior odds, which were 2 to 1, by achieving even losses. Actually they lost 500 tanks while the Germans and Italians lost 450, thus a loss ratio of 1.1 to 1. That left the Brits with a little under half their tanks and the Germans with 50 runners, and defeated - the odds afterward have risen to more than 8 to 1. Which is close to the picture presented by the operational advantage theory, for the western powers as a whole. Even losses, more forces to lose from, high absolute loss rate, yields a rapid swing in forces remaining in favor of the more numerous side.

It can be contrasted with what I'd call the attrition theory, which would see 2 vs. 1 initially, then losses of 1 and .5. Then the larger side fully replaces its losses and the defenders do not get any, leaving 2 vs. .5. The odds ratio has swung to 4:1. The process repeats with .5 and .25 lost, leaving 1.5 vs. .25. The former then has 6:1 odds and breaks through. The total forces used by the attackers in that example are 3 times those of the defenders, but some of them arrive as replacements in the course of the fighting. The loss ratio is 2:1 against the attackers - total losses 1.5 vs. .75. But since the loss ratio is under the total odds ratio including reinforcements, after attrition the attacker achieves high breakthrough odds (6:1) despite losing more than the defenders. The ratio at the front rises from 2 to 1, to 4 to 1, to 6 to 1 despite the defenders KOing twice as many as the attackers do.

The issue between myself and Michael, if I have understood him correctly anyway, is whether the first of those two is a more accurate picture of what was happening in the west (ETO), or with the Americans at least, than the second is. The uberpanzer boosters are extraneous to that debate, although they can't abide the fact that either explanation puts the relatively loss figure under 5. At around 2 for the west in the attrition theory, and around 1 in the operational advantage theory, at least for the Americans.

As for when the initiative shifts, it does so at the time of Stalingrad, when the overall odds are around 2:1. My explanation for the early war period is superior doctrine, as I have discussed at length, especially against the notion that technical tank specs (in tank dueling terms) made much of an operational difference. But the Allies certainly did not need 5:1 odds to take the initiative, and never would have if the loss rate against them had been the uperpanzer theorist's 5 to 1 indefinitely (though that was certainly not only achieved but exceeded in 1941, when the tech spec match up was the worst it ever got). The Russians maintained the initiative with 3:1 odds in fielded forces, even with losses running about that fast, too. They were helped by lousy German operational decisions (Stalingrad, Kursk, little things like that), and sound ones themselves.

Then you wonder what I can mean in saying retreaters do not always have more total write offs. I mean the whole Russian front from Stalingrad to Bagration. The Russians had far more total losses than the Germans, but the Germans were retreating. The Russians built their tanks to last weeks in combat, not years. They lost 1500-2000 total losses a month, while advancing. By 1944, the average life expectancy of a Russian tank caught up with the life expectancy of a German one, though their absolute losses remained higher because their fleet was larger. I am puzzled as to why you even consider it a controversial statement.

To John - thanks for the figures. I realize my numbers for Grants and Stuarts sent to Russia were arrivals rather than sent; the difference being the ones sunk, as you point out. On the Shermans, my 2.1K figure was part of an accounting for the 76mms only, which fits the number you have.

To Germanboy - About Nancy, part of the issue may be others considering only a portion of the campaign, the famous clash around Arracourt over the course of a few days, when the Germans counterattacked with 2 new Panzer brigades. The whole Nancy operation was a longer thing than that counterattack, which itself was larger than the Arracourt fight alone (where the US killed around 45 tanks). The whole fight involved approaching the river, establishing bridgeheads across it, encircling the city, meeting the German armor counterattack, consolidating the position afterward, then dealing with additional German counterattacks sans armor (in Gramercy woods in particular). And then the whole advance toward Metz afterward, which may be considered a seperate but follow-on fight.

The 2 defending Panzergrenadier divisions were the 3rd and the 15th, both veteran formations brought up from Italy to hold the line after the Normandy collapse. They held the river line on either side of Nancy proper, while a VG infantry division held the city itself. The Panzer brigades had not yet been brought to the sector. The Pz Gdrs had a mix of Pz IVs and StuG, below full TOE but not nothing, at least 100 AFV between them. One battalion size armor force from one was committed to an indepedent counterattack right at the point of penetration, before the city fell. It was run over by a full combat command coming down a single road, and destroyed. The rest of their AFVs were in hard defensive fighting at the river line and behind it, both before and after the Panther counterattack, and only a handful of runners remained after the fighting (but see below about the 113th Pz Brigade).

21st Panzer provided HQ staff, artillery, and such to coordinate the new brigades at the time of the actual counterattack (mid September), but it too had been in the area longer than just that attack. It had 20 Pz IVs running when it left Normandy; one of its Panzer battalions had been detached and rejoined only in September, and appears to have had as many again, as Normandy leftovers. The division refitted in Germany in August, before being sent back to the front in the Saar, but was not brought up to anything like full strength. It had 11,000 men by 1 September, which is 2/3rds of TOE, so it was not a shell but not in great shape either. At the end of September, around the end of the Nancy fight, it incorporated the 112th Pz Brigade, which was not one of the two in the Arracourt counterattack.

What was left of the 112th Panzer brigade was incorporated into 21st Pz near the end of September. What was left of the 113th Panzer brigade was incorporated into the 15th Pz Gdr at the begining of October. These formations had a Panzer battalion as per the independent brigade formation structure, plus a Panther battalion each, originally meant for a Panzer division, plus a company of StuG or Jagdpanzer (called a "battalion" but only 10 AFV). The 111th Panzer brigade was fought out by the time of Arracourt proper, with only 25 running tanks remaining, but in the same fighting. It was later incorporated into the 11th Panzer, fighting on the south part of the front.

The organization of the 111-113 Pz Brigades was considerably larger than prior Pz Brigades. Their Pz Gdrs were a full 2 battalion regiment, and they had 2 battalion sized armor units. Each had around 100 AFV. You can see a TOE table of the intended design for these, here - http://www.geocities.com/Pentagon/6931/brigade2.html

The total commitment of German armor to the area was between 350 and 450 AFV. The Pz Gdr divisions had only handfuls left afterward, and 21 Pz had only the survivors of 112th Pz Brigade. The Panzer brigades were gone by the end of the month in which they had formed. Surviving AFV probably totaled between 50 and 100. I am not talking about one afternoon of melee in the fog, I am talking about the collision between Patton's 3rd Army and the equivalent of 5 German divisions, grappling for about a month.

I hope there is somebody somewhere slightly interested in any of this, LOL.

[Slapdragon]

Athough in essence I agree with JasonC's, and certianly not ever Allied tank was a piker, I still think the productions numbers and the numbers on the ground need to be looked at to to figure the type of combat power in place, and to recognize what Allied tanks did the killing.

Taking the US, the huge production numbers for the M4 allowed them to equip all of their allies with this tank, equip a number of armored divisions with the tank, and then form enough tank battalions so that almost every Infantry Division in the ETO and Med theaters had one. The M4 though was not a great tank killer except at the beginning and the ending of its reign. The M18 and M36 battalions when they finally came on line did a huge chunk of the confirmed killing of the AGF.

So what you had was an armor poor German Army fighting the armor rich allies. In terrain that could handle it, the Allies could almost always get armor involved in the fighting, while Germans had to concentrate on ways that the average infantryman could kill of armor. This was true both in the east, where armor came in huge concentrations during breakthroughs, and in the west, where every infantry company had a chance to have some tanks supporting it.

Saying that tanks like the Cromwell and the M4 sucked of course fails to recognize what they were good for and how much they evolved to meet the needs of the war effort. The Cromwell eventually took a 77mm gun and increased armor in the form of the Comet, forming one of the best all round tanks of the war, while the M4A3E8 would someday fight the T34/85 on even terms, and could give the Panther a run for its money with its gyros and turret speed.

But, while the allies war equipment was good, it should never be ignored how good what the Germans had was. A properly sighted Jagdpanther could halt an allied advance for half a day while it was dealt with. On the Russian front a couple of Tiger tanks could blunt a Soviet breakthrough and force an exploitation to reroute around them. If Germany had any fault in its tanks, it was that they made them too good. Talk of German tank reliability problems in part comes from the use of patys that in war time where difficult to get. Eratz parts always gave troubles, and German units were progressively stripped of mechanical staff for front line duty to make up for the huge losses of manpower in Russia and the drain of a war that would eventually have 3.5 fronts (if you include Norway as a front). In fact, Germany would have done well to have built M4s, and the United States could have built Panthers with its overbuilt repair structure.

the 24,000 AFV figure comes from Chamberlien. The Russian figure is from Zaloga, but does not include assault guns which almost doubles the total, nor does it include 25,000 prewar tanks. These numbers do matter to some extent, not in proving they were all used, but in showing how wealthy the sides where. And they are proved just by looking at TO amd Es and seeing how many tanks the average US infantry division could muster to the front