The relative effectiveness of the allied and axis armies in the ETO

[womble]

This is pretty much exactly what I was driving at - that the focus on the relative effectiveness of individual units, while interesting academically, misses the broader point about the importance of logistics, replacements and economic support.

However, it's not a very interesting question to ask at that scale, because the answer can be quickly ascertained by finding out who won the war. On a broad scale, with the exception of some sort of drastic moral about-face, the "better army" (at the scale of leadership, logistics, economics etc) will pretty much always belong to the victor.

I find it very difficult to reconcile some authors' view that the German army was the most professional army of the modern era...

That's because you're looking at it from a different angle. While not supermen, the Germans had hit upon a way of training their troops that was perhaps more modern in many regards than its enemies, giving lower ranks more opportunity and permission to exercise their initiative, as C21st 1st rank armies tend to, earlier than other nations did. From a straight "bashing heads together" perspective, too, a lot of their gear has better numbers than the Allied equivalents. MG42 has better RoF and a quick change barrel; the Panther can put a hole in any plate the allies ever put in its considerable range etc, etc, etc. At that scale, it's easy to make an argument that the Germans were a smidge better off than the allies.

But as you say, that's all largely irrelevant, if the enemy can just bury you in empty SPAM cans dropped from the sky, when you're talking about a war to determine the fate of the largest (? I think Eurasia is bigger than Africa, and I'm thinking of N and S America as separate cos there's on'y a real thin connecty bit) landmass. So, myself, I'd rather say I don't understand why anyone would worry too much about the converse viewpoint.

[JonS]

ya, but the point Overy makes (conclusion to Why the Allies Won) is that the Allies still had to use all their spam cans. The Russians had a crap-ton cans of spam in 1941, but it didn't help them much. If they'd continued using them that way in 1942 and 1943 they'd have lost. Same with the US and UK. Having the New York and Liverpool docks groaning under untold pallets of spam doesn't do much to win a war.

[warrenpeace]

Tarquelne wrote:

Hmm... and the Soviets were supposed to be hot-**** operationally late in the war, which wouldn't contradict the 2:1 German to Russian CEV at all, assuming CEV is focused just on engagements.

I'm not sure I buy that the Soviets were all that great. For Bagration, they outnumbered the Germans 12:1 in guns, 10:1 in tanks/assault guns, 9:1 in aircraft and 5:1 in manpower. Give me those ratios in a quick battle .

[Chek]

Market Garden possibly displays the German military's greatest strength and that must surely be it's command staff.

When the first Allied paratroopers began landing there was virtually no substantial defence in place. But with the ability to set up integrated command posts at what appears to be, "the drop of a hat", and scrounge together a defeated and fractured army and stop the Elite of the Allied forces is pretty impressive. Especially when you consider that the bulk of the German forces were inexperienced, handled themselves poorly in battle and suffered accordingly.

There was also another huge contributor to the German success and that was their logistical situation, especially around Arnhem. They had discovered the relevant information to set up the supply drop zones on a British officer. Consequently they were amply supplied with British food,weapons and ammunition.

Actually I'd probably put logistics first. Boring I know but hard to escape

[Sublime]

I'm not sure I buy that the Soviets were all that great. For Bagration, they outnumbered the Germans 12:1 in guns, 10:1 in tanks/assault guns, 9:1 in aircraft and 5:1 in manpower. Give me those ratios in a quick battle .

Thats for the whole battle, which compromised AG Center. Local engagements wouldnt have always had those odds (obviously). Plus Bagration isn't a great example. Look however at Saturn, or Typhoon, or stopping the Kursk offensive. I'd argue that the Allies and Soviets were in fact better than the Germans simply because in the end they won. You can argue (sucessfully) the Germans may have done individual things or aspects better, which all matters for squat if you can't put it all together and win.

I honestly think its difficult to quantify one side being 'better' than another. Every nation had elite units that would give the other nations regulars a run for their money, etc.

Id add that the legendary German aces, U-Boat, tank, aircraft, etc. had different policies towards other nations - (and ones that Id say were remarkably NOT 21st century and backwards) for example 'fly until you die' and other ideas that kept men in combat way too long, and hurt training new men. Over time it was less policy and more necessity with manpower shortages - but that harkens to my original point that no matter how good at tactical combat the Landsers were, it doesn't matter for squat when the General Staff is scared of Hitler and the government is making idiotic and insane decisions. The Germans also of course were in combat about the longest in WW2, besides of course Britain, Japan, Poland and the others in the original lineup. This means there was obviously more chances for combat for Germans, over longer amounts of time.

What would the U.S. Army of 1945 have looked like with another year of combat under its belt in Europe?

Finally - warrenpeace, I was neither giving Dupuy a chance, nor not giving him a chance. You simply stated he gave the US/Brits a value, then Germans, then Soviets. So I pointed out that hardly seemed an accurate model. And I fail to see how he could have seen the British/US forces as roughly equal. They were very different armies with different experiences, in different positions. I also think Mr Dupuy is doing some serious cherry picking of his own, if he's only assessing the German Army until 1944. You can make any Army look great if you only tell the story up until the defeat, and leave that part out. Of course the German Army was pretty soundly beaten by 44 anyways.

[Tarquelne]

I'm not sure I buy that the Soviets were all that great. For Bagration, they outnumbered the Germans 12:1 in guns, 10:1 in tanks/assault guns, 9:1 in aircraft and 5:1 in manpower. Give me those ratios in a quick battle .

I am sure that operational ability - just like tactical performance - can and should be judged independently of numerical advantage.

It's very easy to do a headcount and attribute the Allied victory to sheer quantity, but I'm not so comfortable with assuming that tactical ability - what Dupuy demonstrated - and numbers - rather obvious - were the only categories where the armies involved had significant differences. Especially if we want to seriously address "relative effectiveness."

[Tarquelne]

All factual claims about Dupuy's work are "IIRC":

And I fail to see how he could have seen the British/US forces as roughly equal. They were very different armies with different experiences, in different positions.

That's what statistics do, Sublime. They allow you to take an arbitrary number of different factors and scales and boil things down - if you want - to a single number. That doens't necessarily mean that number will be of any practical value, but - if you've set things up right - the number will accurately reflect whatever it's supposed to describe. (Heck - the Germans and Western Allies were also very different armies with different experiences and in different positions, and they came out only 20% different.)

Dupuy basically had a model that he'd use to crunch the numbers for different battles. The model would describe the expected result. Differences from that result would indicate performance above or below average from one side or both. Do enough of this math stuff and you can pop-out the CEV numbers.

The most likely place Dupuy messed-up would be the model. (I wonder how well he accounted for the defenders often-touted advantage.) By necessity - war not being as simple as baseball - it has fudge-factors built into it. Though that sort of mistake would be far more likely to throw all the ratings off, not simply assign the UK and the US similar numbers. (Err.. assuming he even tried. Warrenpeace can probably clear this up, but I don't remember if he looked for differences between the Western Allies, or just differences between the Western Allies as a whole, the Germans, and the Soviets. But I think sample size would be the only thing preventing a UK vs. US analysis.)

I also think Mr Dupuy is doing some serious cherry picking of his own, if he's only assessing the German Army until 1944. You can make any Army look great if you only tell the story up until the defeat, and leave that part out. Of course the German Army was pretty soundly beaten by 44 anyways.

Unless he was an idiot that wouldn't effect Dupuy's results one way or the other.

Well, OK, arguably the very nature of the war could have changed enough to throw off his model, but that's sheer speculation. I very much doubt the model was anywhere near that sensitive.

BTW: For at least the first round of studies Dupuy crunched the numbers on the battles the Army chose. And, again, that shouldn't matter when looking at expected vs. actual.

It's like a CM scenario where the victory conditions favor one side because - assuming equally skilled opponents - that side is going to get crushed.

[Andrew H.]

I'm not sure I buy that the Soviets were all that great. For Bagration, they outnumbered the Germans 12:1 in guns, 10:1 in tanks/assault guns, 9:1 in aircraft and 5:1 in manpower. Give me those ratios in a quick battle .

But this is where the focus on tactical wargames distorts what war is really about. The great accomplishment of the Soviets lay *precisely* in getting those odds over the Germans. That didn't happen by accident; the USSR didn't have 12 times the population of Germany, and its economic output wasn't greater than that Germany.

So while the Germans are fiddling around with unimportant crap - "How can we build a tank to take on the T-34?", the Allies are asking the real war winning questions, like "How can we build 50,000 tanks, 100,000 planes, and 200,000 trucks".

The US Army had something like 1000 airplanes in May 1940. After the fall of France, Roosevelt asked the right question ("How can we build 50,000 airplanes a year?" and not the German question ("How can we build a plane better than the Bf 109?")

[kensal]

But this is where the focus on tactical wargames distorts what war is really about. The great accomplishment of the Soviets lay *precisely* in getting those odds over the Germans. That didn't happen by accident; the USSR didn't have 12 times the population of Germany, and its economic output wasn't greater than that Germany.

So while the Germans are fiddling around with unimportant crap - "How can we build a tank to take on the T-34?", the Allies are asking the real war winning questions, like "How can we build 50,000 tanks, 100,000 planes, and 200,000 trucks".

The US Army had something like 1000 airplanes in May 1940. After the fall of France, Roosevelt asked the right question ("How can we build 50,000 airplanes a year?" and not the German question ("How can we build a plane better than the Bf 109?")

Well the German question was often "How can we build a tank weighing 1,000 tonnes powered by u boat engines or an 80cm railway gun with a crew of 250 men to assemble it, 2,500 men to lay the railway tracks and which needs two flak battalions to protect it"

[warrenpeace]

I don't dispute that Allied production was critical in getting the huge advantages in material that won the war. I guess in my mind this does not go to the question of Military Effectiveness, but political effectiveness and economic effectiveness. If one looks at say, percentage of woman in the workforce, the USA was more mobilized for war than was Germany. Not sure where things like arms production and logistical decisions belong. These are part military and part civilian. If you include these in Military Effectiveness, then I certainly agree with what has been said.

I read Dupuy a while ago, but if I remember correctly, he takes into account a variety of factors including terrain, defender vs. attacker, etc. He also applies his model not just to ww2 but to battles in a variety of wars both older and newer. However, as has been pointed out this is a tactical evaluation, i.e. the model starts at the time the battle begins and does not take into account the strategic and operational perspective.

That being said, I really don't think the Germans were operationally inferior. In my mind the Allies in ww2 were a little like Grant and Sherman in the Civil War. Grant recognized his strategic advantage and pressed them to victory. Not elegant or flashy, but effective.

[Chek]

Logistics is most definitely a military matter and is what differentiates the good commanders from the merely adequate.

[warrenpeace]

Chek

I agree that logistical understanding is important for Generals, although two of the best, Rommel and Patton, were not known as great logisticians. My point is that there are two types of logistics, strategic and operational. Strategic logistics are determined at the national level, e.g. things like: How many tanks do we need to make?, How many landing craft?, should we engage in the Manhatten project? etc. Operational logistics are how do we take the stuff that we have and distribute it too the troops such that they have what we need. I would argue that strategic logistical questions are not strictly military but are a combination of military and political. Furthermore, I would state that WW2 was won by the Allies primarily on the basis of these decisions and their superior execution.

I believe, overall, the German Military was quite competent with regards to operational logistics given the strategic decisions that were made. There are a few exceptions (winter gear for 1941 comes to mind). Much of their poor strategic logistic decision making stems directly from their political process in which Hitler made many decisions.

[womble]

I believe, overall, the German Military was quite competent with regards to operational logistics...

They did, however suffer from the constraints of those logistics decisions: large numbers of horse drawn transports and guns, for example. Such things often weigh lighter in the minds of those who espouse Axis superiority than the "tactically superior" tanks and small arms, which are also only present because of those "strategic logistic" decisions made at the political level. You can't, realistically, divorce the two.

[ASL Veteran]

Alrighty then, since this thread has become a thread about Dupuy let’s have a look at what his methodology is exactly

1.Seventeen forces that were engaged in combat during the period from the start of World War II to the present were selected.

The evaluations were based on forces in combat about which the evaluators had detailed knowledge. These forces also had to be ones that were involved in a number of combat engagements, so that the evaluation would reflect a true combat effectiveness rather than a nonrepresentative performance in one battle.

So far so good.

2. The Judgments of a group of military historians about the comparative effectiveness of these forces were obtained.

The second step involved asking three military historians to compare the combat effectiveness of all the forces, two at a time. The historians were asked to compare "troop quality," because it was the investigator's judgment that this would be the most familiar designation forcombat effectiveness, as distinguished from the effects of numbers and weapons. It was recognized, however, that it would be impossible for these historians, or anyone, to judge troop quality at the force level, and that what was actually teing judged was tWe emobat effectiveness of the forzes. Each historian was asked, "Which force had the higher troop quality, the Germans or the British in 1943-44?" and then, "Which force had the higher troop quality, the Germans in 1943-44 or the Chinese in 1950-52?" The procedure was carried out with all possible pairs of the forces listed above, In general, the number of comparisons is given by the formula n (n-1)/2, in this case 136 comparisons. Figure 1 shows the form on which these comparisons were made.

It should be noted that, although the size of the panel may seem small, in establishing a consensus of informed judgment the qualifications of the panel members were judged to be much more important in this preliminary effort than the size of the sample. This panel was the best informed one that could be questioned in the time available. The form has been circulated to six adaitional highly qualifited military historians, and their responses will be used to adjust the data obtained, if necessary. Since the agreement among the three queried was close, it is not expected that there will be significant changes when all the responses are received.

So they bring in three military historians to indicate who is better ‘qualitatively’ between each of two armies that are compared side by side. Who are these three historians? This is also obviously a set of subjective opinions that are being used as the basis for the creation of a quantifiable number.

3. The judgments were converted into a scale on which the effectiveness of each force was quantified.

Each historian's judgments were converted into a scale, ranging from a value of zero for the force with the lowest effectiveness or capability, to a value of 100 for the force with the greatest effectiveness or capability. This conversion was carried out by the Method of Paired Comparions. This method was devised in the early 1930s by Thurstoneone of the fathers of psychometrics. It is based on a mathematical fact that may not be readily apparent, unless one has explored the mathematical implications of choosing between two alternatives. The fact is that when a person makes a series of such choices among two alternatives, he has carried out the mathematical equivalent of assigning all of the alternatives to locations on a scale. This scale can be normalized so that its maximum and minimum values are any preselected values, often zero and one hundred. While the individual has difficulty in assigning items being evaluated directly to the scale, he has no difficulty in evaluating which of any two of them is preferable to the other. Since Thurstone's day, his original idea has grown into a field of mathematics with a large literature in its own right, nonmetric scaling. Of the more recent texts on the subject of scaling, listed in the bibliography, is that of van der Ven.

Using this mathematical methodology, the choices of each of the three historians were converted into points on a scale running from zero to 100. The results were listed statistically to determine whether the three scales might be significantly different from one another. This was done by applying the Friedman Two-Way Analysis of Variance test. A null hypothesis of no significant difference among the three could not be rejected at either a I%or 5% level of confidence. Accordingly, the judgments were pooled, and the scale value assigned to each force was taken as the average of the values resulting from the judgments of the three military historians. In a large number of judgments, a rater might make logical lapses. That Is, he might state that Force A was superior to Force B and that Force B was superior to Force C, and then state that Force C was superior to Force A. To test the possibility of this having occurred, Kendall's Coefficient of Consistency was calculated for the evaluations ofeach of the three raters. The results were greater thanfor eacn of ther't indicating a high degre of internal logic in each of the three sets of Judgments.

That’s an awful lot of mumbo jumbo explaining that they used the opinions of three military historians to rate an army’s quality. No matter how much mumbo jumbo you toss in there it is still a compilation of the subjective opinions of three military historians.

4, Historical data about the demographic, economic, political, military, educational, and climatic characteristics of the nation to which each force belonged was assembled.

Data was next accumulated about each of the nations whose forces was being evaluated. The sources of the data were the cross-national data publications of the United Nations, the World Bank, and related sources. (See bibliography.) These general collections of data were supplemented in a number of cases by single items of data from other publications (also listed in the bibliography).

Data was accumulated on the following characteristics:

Demography

Birth rate

Household size

Male life expectancy

Population density

Protein consumption per person

Education

Male literacy (%)

Primary schooling (%)

Secondary schooling ()

Economics

Energy consumption per person

GNP per person

Physical quality of life Index

Politics

Civil disorder index

Civil liberties index

Political power index

Political rights index

Military

Battle deaths per month per million inhabitants

Civilian/military ratio at start of war

Military power index

Military spending per person

Climatic

Rainfall

Temperature in coldest month

Temperature in hottest month

These characteristics were selected according to two guidelines. First, the Book of World Rankings was used as a rough screen of what might be worthy of further investigation. This volume ranks the nations of the world according to approximately 450 characteristics. In cases where a large number of the nations being evaluated were at either the high or the low end of a ranking, that characteristic was tentatively selected for evaluation. Second, an attempt was made to get at least two characteristics that were logically related to each broad area, and for which a plausible argument could be made that they might be related to troop capability. Thus, Average Household Size was selected because it was felt that the type of family structure in which a soldier grew up might be related to his capability in combat.

Several of the items of data are index numbers, well supported by discussion in the scholarly literature, that measure characteristics too complex to be reflected in a single item of data. Among these are the Physical Quality of Life index, devised by the Overseas Development Council, and the Political Power index devised by the Georgetown University Center for Strategic and International Studies, An attempt was made to pick indexeý that are supported by explanations published in the open literature of the particular discipline involved.

The data was gathered for a period that was approximately contemporary with the war in which the troops were being evaluated, Most items were for the nation within two or three years of the time of the war and none were more than 10 years before or after it. Where data this current could not be located or calculated from other data, the item wasleft blank. Seven of the 24 characteristics had 12 or more items of data, while 5 characteristics had 7 or fewer items of data.

Seriously .. rainfall? Birthrate? Quality of life?

5. The degree of association, or correlation, between each characteristic and the quantified estimates of force effectiveness from Step 3 was calculated.

The product-moment correlation coefficient ® was calculated between the scaled value representing force effectiveness (or troop capability) and each of the characteristics of the nations listed previously. These coefficients were then interpreted to answer the following question: Is the association between effectiveness and a particular characteristic strong enough to reflect an actual association between the set of characteristics for all countries and all wars during the period or could It have been due to the fact that only a small sample of data was employed?

After applying a standard statistical procedure, it is possible to make an inference about whether the association or correlation in the particular sample is "significant," that is, whether it arose from an underlying relation or from the particular sample of data employed. A high value of the correlation coefficient does not imply a cause and effect relation between the two variables. It only means that a positive change in one is accompanied by a change (positive or negative) in the other.

The correlation coefficients were evaluated at two levels of significance, the 5% level implying one chance in 20 of the association's resulting from chance, and the 1% level, implying one chance in 100. These are the levels of testing comnonly employed in most statistical analysis. The results are shown in Figure 3.

Of the national characteristics that were strongly associated with effectiveness (1%confidence level), three had a negative correlation, that is,increases in the characteristic were associated with decreases in tombat effectiveness. These were:

Birth rate

Household size

Temperature in the hottest month

One strongly associated characteristic (1% confidence level) had a positive correlation. This was:

Male literacy

None of the characteristics reflecting the political or military characteristics of the society were strongly associated with combat effectiveness.

The high level of correlation noted above indicates the probability that a formula could be derived to estimate combat effectiveness for the forces of any two forces that are being studied in a war game. No significant differences are observable between a scale of combat effectiveness or troop capability derived from informed military

So you take the subjective opinions of three historians and you then go and find data such as rainfall and GNP per person in order to confirm why the historians opinions are what they are. So three historians say “Germany in 1944 ROCKS!!” then you go and sift through rainfall data and make a conclusion that “Germany must ROCK because there is more rainfall in Berlin than in London.” Science is awesome. So that’s it. That’s what Dupuy did and what he measured. It should probably be noted what he didn’t measure. Let’s have a look at this CEV thing that he compares his results to

As part of this same task, a review of the findings of HERO's Quantified Judgment Model (QJM) on the combat effectiveness of national forces was carried out. Analysis of a large number of combat engagements – analysis carried out during the past 10 years -- has shown that there is a quantifiable factor in addition to numbers of men and firepower that helps determine the outcome of battles, and that this factor differs from division to division anc from national force to national force.

This factor is termed by HERO the relative combat effectiveness value (CEV). The CEV is assumed to include troop capability, along with leadership training. and tactics. Thus far it has not been possible definitively to isolate any of these subfactors, including troop quality, from the total CEV, but the unquestionable existence of CEVs does show objectively that numbers and weapons alone do not determine battles.

So the CEV is assumed to include a lot of militarily important stuff, but nobody can isolate any of these factors. So the factors he deems to be important such as troop capability, leadership, training, and tactics are all folded into this CEV value but they can’t be isolated so nobody actually knows what they are. Note that Dupuy definitely didn’t measure anything relating to those things he deems important to the CEV such as leadership, training, or tactics. Nope, Dupuy measured the amount of rainfall that a certain nation has and what the national birthrate was. What else does Dupuy say?

A search was made in the HERO Engagement Data Base for units that had unusually high CEVs, in the hope that demographic and other relevant data about the troops composing these units might help establish a link between individual capability and unit performance. The division identifed as having a particularly nigh CEV is the 88th Division, in its performance against German units in Italy in 1944. It was not possible within the time frame of this study to carry out any research on the relationship of the relative combat effectiveness of divisions and demographic or other statistics, but follow-on research would appear promising.

So in other words, after taking the opinions of three military historians and finding out through rainfall data why Germany ROCKS, Dupuy freely admits that when he looks at a specific unit that was listed as having a high CEV he can’t figure out why by examining birthrates and rainfall data. Presumably American infantry divisions such as the 45th that came from New Mexico (which is hot) would perform more poorly than an American division that originated from New York. What else does he say?

There is almost certainly some tradeoff between the factors that compose combat effectiveness; that is, leadership and training almost certainly can compensate to some extent for low troop quality. However, it is obvious that superior leadership and training would be required if quality were low, not just average leadership and training.

There also seems to be some synergistic effect among troop capability, training, and leaaership. Good commanders raise the capability of their troops, and with good troops a commander performs better. Leadership is especially important, because of its impact on training, and the impact, In turn, of training on troop capability. This does not mean that low troop capability does not have a degrading effect on combat performance, and it does not mean that a predictor for degraded combat performance cannot be found, but it does complicate the task.

Of course, Dupuy didn’t measure any of that did he? He did this

Preliminary exploration of the relationship between the combat effectiveness of armies during the past 40 years and a variety of national demographic, economic, and other factors indicates a strong association between a number of these factors --

male literacy, household size (negative), birth rate (negative), and temperature of capital city in the hottest month (negative) – with national combat effectiveness.

Yeah, he measured household size and the temperature of the capital city in the hottest month. This study is almost, dare I say, Kettlerian. He goes on to say

Two points may be made about this preliminary finding. One is that some of the statistics that characterize an industrialized society are also associated with high troop quality in the middle-to-late 20th Century. Another is that one of these factors, male literacy, is one that can be identified in the individual soldier and thus controlled for the army. Also, it is possible that it could become the basis for a predictor of army combat effectiveness. It is important to understand, however, that the findings so far do not give any reason to believe that increasing literacy in a given amy will significantly increase combat effectiveness.

So that’s Dupuy. He goes and finds three military historians and asks them to subjectively compare army effectiveness at various times in the 20th century two armies at a time. He then goes out and finds rainfall and temperature data to find out why the armies that the historians rated as better were better. He then goes and mentions factors such as leadership, training, and tactics in passing, but he doesn’t measure any of those factors and the CEV value that he obtains from other research is supposed to include those factors but none of those factors can be isolated. Good stuff.

[warrenpeace]

ASL,

WHere are you getting those Dupuy quotes. That is not the methodology used in "Numbers, Predictions and War".

[warrenpeace]

Actually, I see where you are getting this and how you are deliberately misinterpreting it. This appears to come from a report the Dupuy did for the Army. Note, the historian part has absolutely nothing to do with the CEV modeling (This is why he says "As part of this same task, a review of the findings of HERO's Quantified Judgment Model (QJM) on the combat effectiveness of national forces was carried out." ) Here is a good post that describes the basics of QJM methodology and how it has been validated. If your going to criticize it, at least spend the time to understand it.

http://h-net.msu.edu/cgi-bin/logbrowse.pl?trx=vx&list=h-war&month=0602&week=d&msg=k4qdqzoz%2Bwegn5N/r8zL3g&user=&pw=

[ASL Veteran]

Actually, I see where you are getting this and how you are deliberately misinterpreting it. This appears to come from a report the Dupuy did for the Army. Note, the historian part has absolutely nothing to do with the CEV modeling (This is why he says "As part of this same task, a review of the findings of HERO's Quantified Judgment Model (QJM) on the combat effectiveness of national forces was carried out." ) Here is a good post that describes the basics of QJM methodology and how it has been validated. If your going to criticize it, at least spend the time to understand it.

http://h-net.msu.edu/cgi-bin/logbrowse.pl?trx=vx&list=h-war&month=0602&week=d&msg=k4qdqzoz%2Bwegn5N/r8zL3g&user=&pw=

I am not deliberately misinterpreting anything. I was reading directly from the document as it was written. I will check your link though and see what it says.

[ASL Veteran]

In “When the Odds were Even” the author summarizes the situation for the German Army of 1944 as the following

German tactical and operational doctrine was theoretically sound and battle proven, but this doctrine was not always reflected by the organization of army units for combat by 1944. For a variety of reasons, many new types of units were introduced by late 1944, each with varying organizations and types of equipment. Many of them were unsuited for the execution of doctrinal tactical or operational maneuvers, and this only added to the Clausewitzian friction always present in war. Further, the growing disparity between Waffen SS and army units caused counterproductive rivalries that detracted from the common effort against the Anglo American Allies and the Soviets.

By late 1944, these units, already handicapped by doctrinal and organizational inconsistencies, were often manned by ill trained and sometimes unmotivated soldiers. In attempting to deal with the rapidly worsening battlefield situation, the German army adopted measures that often rebounded to its disadvantage. The failure to prepare a replacement system that could adequately cope with the demands of a protracted, high intensity conflict resulted in the adoption of increasingly counterproductive stop gap measures. The most billiant set of tactics and operational techniques are useless if they are not carried out by formations suited to the accomplishment of the required tasks; by late 1944, the German army in the west was suited neither by organization nor by personnel and training for the execution of its mission to hold back the Allies from the gates of Germany.

As to the Americans he states the following

American tactical and operational doctrine closely resembled that of the Germans in its essentials by late 1944. As such, it adhered closely to those Clausewitzian precepts that were still valid. It stressed the importance of combined arms operations in both the attack and in the defense, and made considerable demands for initiative and good judgement on the part of its soldiers.

Unlike the Germans, however, whose organization for combat sometimes failed to accurately reflect the needs of tactical and operational doctrine, the Americans organization was admirably suited to the task. By uniformly organizing and equipping their units and by the institution of carefully tailored regimental combat teams in their infantry divisions and combat commands in their armored divisions, the Americans ensured flexibility and cohesion in the execution of their doctrine.

Although American unit training was not ideally conducted due to the need for replacements for units already in combat, it nevertheless basically satisfied the requirements of cohesion and bonding by ensuring that most of the soldiers in combat formations had served and trained together for a considerable amount of time prior to deployment to a combat theater. In marked contrast to the Germans, the Americans integrated replacements from other branches and specialties into their units, rather than the wholesale creation, in short periods of time, of units from personnel originally trained for duty other than ground combat. During lulls in the action, US units conducted training designed to sharpen combat skills and to fully integrate replacements for men lost in combat.

Of course, strategic considerations dictated many of the differences in organization and training for the respective foes armies in the Vosges by the autumn of 1944. It should be remembered, however, that both nations were conducting two front coalition warfare, and that their strategic situations were mostly of their own making. Nevertheless, the fact is that, by 15 October 1944, American doctrine was more clearly reflected in the organization of its troops for combat, and US soldiers were far better trained and led that most of their Wehrmacht counterparts. The result of these differences was the success of American arms in the Vosges campaign.

Of course the author explains in detail all the factors he considered before writing that summary. I didn’t copy all of that information because it would be much more than I want to type. I can access the specific information if necessary though.

[JasonC]

I think Dupuy's results are quite questionable, but not for the sort of reasons ASL Veteran is talking about. ASL Veteran is being unfair to Dupuy and to the historical process that led him to postulate a national CEV. He is confusing Dupuy's later and rather naive attempts to *explain* the origins of an observed CEV, with how Dupuy found a CEV factor in the first place.

The CEV is an empirical fudge factor that was derived from actual combat outcomes in the Italian campaign in 1943, division level engagements. Dupuy found that in that era, controlling for stance (offensive defensive), terrain, weather etc, the Germans won the engagements that had a very close to even, objectively calculated combat power on the two sides, very consistently. It was not a large effect - if the Allies had a modest edge in objective combat power they predictably won the engagements - but it was present. He could improve the fit of his calculated objective combat values to outcomes mapping if he added a pro German fudge factor of 1.2 times. This was an empirical fact about the data from the Italian campaign, not something he derived upward from any theory of the causes of combat effectiveness.

Later, having seen national CEVs in multiple eras and match ups - e.g. the Israeli's outperforming the Arabs by 200% aka 3 to 1 based on objective combat power calculations alone - Dupuy then sought to explain CEVs. The naive analysis that ASL Veteran is going on about was just an attempt to take fields available from sources like the CIA World Factbook and see if they might be predictive of national CEVs.

I think it is fair to call that attempt a failure, but it is certainly not where Dupuy got CEVs from in the first place. They come from the data of who won the battles. Arabs that outnumbered Israelis 2 to 1 by combat system firepower measures and the like, uniformly lost to the Israelis.

As for what the weather and rainfall stuff is actually about, they are Montesquie-esque proxies for race and civilizations, pretty transparently. Similarly the measures like literacy and GDP per capita are proxies for industrial civilization aka the developed world vs the third world. Dupuy was trying to be data-objective about the difference between an NVA rifle regiment and a regiment of US Marines, beyond their firepower support elements. And make no mistake, there is such a difference and it is huge.

When two forces from very similar civilizations with very similar standards of civilian personnel skill levels and education, with similar standards of professionalism and training, fight with one another, empirically calculated CEVs based on who won the even engagements are going to come out quite close to each other. But that is not what we see when it is Germans fighting WW II Russians, or Israelis fighting 1967 or 1973 Arabs, or US Marines fighting North Vietnamese.

In all of those cases, one of the armies is coming from a much richer and more industrial society with vastly higher standards of education, professionalism, discipline, and all the rest of it. And we don't see empirical CEVs like 1.2 times, we see CEVs like 3 and 4 times. Even controlling for better fire support systems and such, the individual men are just fighting much tougher, smarter, more disciplined etc and getting more out of each of their weapons than the less sophisticated side manages to get out of theirs.

That is a fact, Dupuy is not making it up. He is trying to explain it quantitatively.

As for the reasons I have for finding some of Dupuy's results questionable, they don't have to do with any of the above. Instead they stem from how he measured combat outcomes. Basically he puts a large emphasis on mission success and ground control. Losses are also involved and important, but the weight given to which way the front line moves is to me far too great in his outcomes scoring. I simply do not regard ground control as important for measuring outcomes, because to me the proper large scale "frame" for assessing combat performance is attritionist. I would put a far greater emphasis on the losses and much less on the other aspects of combat outcomes. But in addition, the time and operational scale used has to be large enough to include the losses caused by occasional "line collapse". That is, you can get a very unrepresentative sense of the exchange ratio between say the Allies and Germans in Normandy if you focus on "divisional level engagements" in the push to St. Lo period, than you will if you include episodes like the surrender of Cherbourg or the Falaise battle. The offensive stance hurts in the smaller scale and helps in the larger; it "cashes" operational superiority for lopsided defender losses only after forcing local line collapse events.

That is the place I see room for serious revisions to Dupuy's methodology. Basically, the outcomes should be measured on a larger, operational campaign scale, less on a tactical division battle for 2-3 day scale, but should then be measured by losses not ground taken or other immediate mission objectives. If the ground control or mission objectives are truly important, they will show up as forced losses to the enemy on that larger, operational campaign scale. If they can't be translated into such objective losses on that larger scale, then they are not in fact important combat outcomes.

Anyway, I wanted to make clear where I do see an issue with Dupuy's methods, as well as where I find them to be basically sound. Above all, he tells you what he is doing, and that is great.

[ASL Veteran]

Actually, I see where you are getting this and how you are deliberately misinterpreting it. This appears to come from a report the Dupuy did for the Army. Note, the historian part has absolutely nothing to do with the CEV modeling (This is why he says "As part of this same task, a review of the findings of HERO's Quantified Judgment Model (QJM) on the combat effectiveness of national forces was carried out." ) Here is a good post that describes the basics of QJM methodology and how it has been validated. If your going to criticize it, at least spend the time to understand it.

http://h-net.msu.edu/cgi-bin/logbrowse.pl?trx=vx&list=h-war&month=0602&week=d&msg=k4qdqzoz%2Bwegn5N/r8zL3g&user=&pw=

I glanced at the link and there was nothing in there that changed my opinion. I suppose this part is the part you want me to see the most

Like all combat models, the QJM is a theory of combat. As Col. Dupuy clearly pointed out, it is not a theory of war. The theory behind the QJM is actually quite simple: in combat, all things being equal, the side with the superior combat power wins. Col. Dupuy defined combat power (P) as force strength (S); times the environmental and operational factors (V) affecting the circumstances of combat on the force; times the relative quality (Q), or combat effectiveness, of the troops. Put quantitatively: P = S x V x Q. Dividing the P of the friendly side vs. the P of the enemy (Pf/Pe) produces a ratio. If the Pf/Pe ratio is higher than 1.0, the friendly side will win; if it is less than 1.0, the friendly side loses. The higher the ratio, the greater the magnitude of the win, or vice-versa, the greater the magnitude of defeat. A Pf/Pe ratio of between 0.9 and 1.1 is considered to be an indeterminate outcome.

Col. Dupuy's theory was based upon analysis of a mass of real world

observations. While social science data can often be highly variable and

"fuzzy," in this instance, the patterns were quite clear. As TDI President

Chris Lawrence often puts it, "combat is largely deterministic with some

stochastic elements."

Col. Dupuy's theory has been put to the test several times in the form of

validation testing of the QJM and its successor, the Tactical Numerical

Deterministic Model, or TNDM. The QJM was validated to the data base it was initially created from, 60 division-level engagements between US, British, and German forces in Italy in 1943-44. It was then tested to a Validation Data Base of 21 division-level engagements drawn from US and German forces in France in 1944, the Battle of the Bulge, and German and Soviet forces at the Battle of Kursk. Col. Dupuy stated that the QJM correctly picked the winning side in 55 of the original 60 engagements, and 20 of 21 Validation Data Base engagements, for an accuracy rate of 93%. [This validation is detailed in _Numbers,

Predictions & War_, pp. 57-70.]

That's all fine, plugging numbers into an equation, but there doesn't appear to be anything in those equations that account for most of the real world factors that affected the German army in 1944. He is also, by default, cherry picking engagements to make his calculations since he isn't selecting every division sized engagement. It's a data set, but it isn't very persuasive to me since I'm well read enough to know about the substantial problems faced by the German Army in 1944. Which battles during the Battle of the Bulge did he select? The ones involving the 101st Airborne at Bastogne or the ones involving the green American forces in the St Vith area?

[Tarquelne]

I'm not terribly interested in defending Dupuy.

OTOH, I'm somewhat interested in explaining the difference between magic, guessing, and work in the social sciences.

Alrighty then, since this thread has become a thread about Dupuy let’s have a look at what his methodology is exactly

The overall point of may be: When you take a methodological description, then reject the relevant details and leave yourself with only a dismissive over-simplification, you're hardly looking at it "exactly."

This is also obviously a set of subjective opinions that are being used as the basis for the creation of a quantifiable number.

Yes. Most researchers have limited budgets and limited amounts of time. They have to replace asking everyone in the world, time machines, or god-like omniscience with DUM DUM DUM!.... statistics.

No matter how much mumbo jumbo you toss in there it is still a compilation of the subjective opinions of three military historians.

The mumbo jumbo tossed in there describes why you'd be justified in trusting those subjective opinions more than you could under other circumstances.

Seriously .. rainfall? Birthrate? Quality of life?

I guess the theory was that soldiers of different backgrounds might tend to be of different qualities, and they wanted to explore what may or may not make a difference. Of the 20+ things they tested only a handful much of an impact. The only one that surprised me was "temperature". And this study pointed out these were all just correlations.

"Temperature" probably says more about "first world" vs. second or third status, or the general history of the nation than temperature itself.

One might find out something interesting from following that up. Or you could just mock it. Tire pressure? What the heck does that have to do with saving energy?

So you take the subjective opinions of three historians and you then go and find data such as rainfall and GNP per person in order to confirm why the historians opinions are what they are. So three historians say “Germany in 1944 ROCKS!!” then you go and sift through rainfall data and make a conclusion that “Germany must ROCK because there is more rainfall in Berlin than in London.” Science is awesome.

I suspect you're being sarcastic.

If so, it looks like you misunderstood what was going on. They got their subjective measures, then they crunched some numbers to make them not quite so subjective, then they compared them to some even-less subjective numbers, then they compared all that to some objective numbers in an attempt to figure out which of those objective measures are important.

Yes, yes, yes, that's all based on the assumption that three military historians they cross-checked against each other aren't *all* idiots, that the work in the QJM they also used as a cross-check wasn't a waste of time, that the people compiling the hard data didn't lie, and that math actually works.

And, yes, it would have been better if they'd compared only objective measures to objective measures. Too bad nobody has any.

OTOH, if they keep pursuing this line of research, maybe some can be generated one day.

So the CEV is assumed to include a lot of militarily important stuff, but nobody can isolate any of these factors.

Yes. They'd be much happier if they could simply pretend they could isolate that stuff. I assume these guys are all too honest to be politicians and too ugly to be actors.

So the factors he deems to be important such as troop capability, leadership, training, and tactics are all folded into this CEV value but they can’t be isolated so nobody actually knows what they are. Note that Dupuy definitely didn’t measure anything relating to those things he deems important to the CEV such as leadership, training, or tactics. Nope, Dupuy measured the amount of rainfall that a certain nation has and what the national birthrate was.

Yes. When you can't measure something you find things you *can* measure, and then try to suss out any relationships. "Hard" scientists do this, too. A lot.

So in other words, after taking the opinions of three military historians and finding out through rainfall data why Germany ROCKS, Dupuy freely admits that when he looks at a specific unit that was listed as having a high CEV he can’t figure out why by examining birthrates and rainfall data.

No, he said he didn't have time to even try, but it seemed like a good idea. *Now* I can understand why warrenpeace said "deliberately misinterpreted." That's a pretty egregious mistake you made there IMO.

Based on your next post - and the one quoted from above - I think you may have completely missed that the CEV numbers mentioned earlier in the thread came from a completely different method from the one described in the SCACE document.

The "QJM" method uses a model - which does have subjective stuff in it - to crunch numbers on battles and, eventually, spit out the CEV scores.

The stuff with the three historians and the demographic data was one part of a quick (less than 2 months, I think - one part mentions a "45 day timeframe") study. It was designed to identify what influences troop quality, not present a cast-iron conclusion about who had the best troops.

But - and I think this isn't unreasonable - if you want to identify influences on troop quality it's handy to have some idea of the quality of the troops. And since you'll be using statistics, you need numbers.

OMG, ninjad by Jason. Events beyond my control interrupted the composition of my message. And I had to move big rocks this morning - my fingers are tired.

JasonC

I simply do not regard ground control as important for measuring outcomes, because to me the proper large scale "frame" for assessing combat performance is attritionist. I would put a far greater emphasis on the losses and much less on the other aspects of combat outcomes.

Good point. It'd go a long way toward correcting the method's blindness with regard to the larger scales. At the very least it'd aid in the examination of how willing armies are - or should be - in accepting losses when achieving objectives.

[Michael Emrys]

I agree that logistical understanding is important for Generals, although two of the best, Rommel and Patton, were not known as great logisticians.

Which suggests to me that their reputations are at least somewhat inflated. Rommel, for instance, never grasped how unimportant the North African theater was in the grand scheme of things. This led him to constantly attempt more than his army with its limited logistical support was capable of.

Michael

[MikeyD]

Currently I'm rereading Hans Delbruck's absurdly thick multi-volume "history of the art of war" (nothing on WWII, he died in 1929). Delbruck seems to take particular glee in taking up his colleagues' academic arguments on particular armies and battles and utterly eviscerating them point-by-point. What we're arguing over in this post, they were arguing back in 1898 over the composition of the Hussite wagon armies of 1421. Right down to the finer points of methodology and numerical analysis. In other words, I'm really enjoying this discussion.

[Childress]

This led him to constantly attempt more than his army with its limited logistical support was capable of.

Michael

This is the sort of English grammar up with which I will not put.

[Michael Emrys]

This is the sort of English grammar up with which I will not put.

You fail to boldly go where no grammar has gone before.

Michael