|
|
Mythos revisited: American Historians and German
Fighting Power in WWII
by Thomas E. Nutter
|
Chapter Seven
Trevor N. Dupuy, NUMBERS, PREDICTIONS & WAR
Sharing with Martin van Creveld the unenviable distinction of being the
historian most despised by Mansoor, Bonn and their cohorts is a retired United
States Army Colonel, Trevor N. Dupuy. Dupuy was a graduate of the United States
Military Academy and well-known student of military affairs, and the author, by
himself and with others, of literally dozens of books on various aspects of the
topic, from ancient to modern times. In the post-World War II era, Dupuy became
a consultant to the defense establishment in the United States. He worked with
a network of former military officers, academics and other defense consultants
to supply the needs of the American military as it strove to prepare to fight
the next global war. Dupuy's affiliation with the Historical Evaluation and
Research Organization (HERO) resulted in, among other things, the study that
led to the publication of Numbers, Predictions & War in 1979.
Numbers represented Dupuy's effort to express to a wider audience the
analytical systems he and his colleagues at HERO had developed for use in
attempting to understand combat.
The language used by Dupuy's critics betrays the virulence of their reactions
to some of the conclusions set forth in Numbers. Peter Mansoor refers
to Dupuy's "assertion of the inferiority of American combat units on the
European battlefields of World War II" and avers that he "concluded that German
units were on the average 20 percent more effective than their British and
American counterparts." Mansoor says that the factors found by Dupuy to be
responsible for this outcome included "better utilization of manpower, more
experience, greater mobility, better doctrine, more effective battle drill,
superior leadership, and inherent national characteristics." He calls Dupuy
"the vanguard of a group of historians who trumpeted the tactical superiority
of the Wehrmacht at the expense of the American army" and comments
favorably on the criticism of Dupuy set forth in John Sloan Brown's Draftee
Division .(67)
In When the Odds Were Even , Keith Bonn excoriates Dupuy on two bases.
He argues first that the "incredibly complicated series of parameters" (the
availability of ammunition and fuel, the effects of weapons and morale, the
quantities of troops available, "et cetera ad infinitum") utilized by Dupuy to
analyze ground combat did not exist "as such during the period when the battles
being analyzed were fought." Consequently, in Bonn's view, those parameters
"are artificial and ex post facto at best, irrelevant at worst." The second
basis upon which Bonn criticizes Dupuy is that the latter included in his
analysis intangible variables not easily amenable to assessment. For these
reasons, Bonn pronounces "[T]he usefulness of [Dupuy's] work as anything more
than an interesting collection of conceptual ideas for commercial war games
…extremely limited." Some preliminary comments on Bonn's conclusions are worth
mentioning at this point. First, the notion that factors such as ammunition and
fuel, weapons and morale, and the number of troops available to a commander did
not exist during World War II, but are instead the artificial construct of
Dupuy, is so absurd that it would not merit comment but for the fact it is
asserted by a professional officer in the U.S. Army, and is therefore leant a
certain degree of credence. How anyone, let alone a graduate of the United
States Military Academy, could maintain that a person analyzing a particular
engagement should not take into account that one of the forces involved was
understrength, had been repeatedly pummeled by its adversary, was low on
ammunition and fuel and operated with weapons inferior to those of its
opponents is so fundamentally ridiculous as to be beyond comprehension. Second,
had Dupuy ignored the "intangible variables" of combat (the so-called "fudge
factors") in his analysis, he would have been guilty of the grossest
distortion, and would therefore have merited even more stringent criticism than
that offered up by the likes of Bonn.(68)
Michael Doubler evidently has so little regard for Dupuy that he mentions
neither him nor his work. A critique of Dupuy, however, is essential to John
Sloan Brown's Draftee Division, so much, indeed, that he devotes an
entire appendix (captivatingly entitled The Mythos of Wehrmacht Superiority:
Colonel Dupuy Reconsidered) to a supposed refutation of Dupuy's work. In
fact, however, Brown's appendix represents little more than an opportunity for
him to vent his spleen about a number of real or imagined slights that Brown
perceives to have been heaped upon the reputation of the American fighting man.
Most of these slights have little or no relation to Colonel Dupuy. For example,
in the very first paragraph of the appendix in question, Brown begins by
remarking upon the "pervasive adulation of the Wehrmacht " which he
claims to have infected the history of World War II, and ends by asking
rhetorically whether American soldiers, having captured thousands upon
thousands of German soldiers, believed that the latter were "better than
themselves." Neither of these points of view form any part of Dupuy's analysis.
In turn, Brown blames the exaggerations of German combat prowess by the British
and French, made for the purpose of explaining away their own defeats, German
bombast, and cartoonists such as Bill Mauldin for having laid the groundwork
for the "inflated" perception of German arms. Added to these elements were the
availability of German sources, the postwar disinterest of Americans in
military history, and the availability of Hitler as a scapegoat for German
failures, all of which tended to unfairly support the notion of German martial
superiority. There are, finally, certain prominent military historians, such as
S.L.A. Marshall ("suggesting that most American infantrymen spent World War II
cowering in the bottom of their foxholes") and B.H. Liddell Hart ("vent[ing]
his pique on Allied leaders who did not share his elevated impression of
himself"), as well as a number of unidentified behavioral scientists, whose
work had the effect of creating "images skewed to favor the German soldier over
the American". Brown's truly childish vitriol on the subject of these alleged
slights is best illustrated, however, by his conclusion that the unfair skewing
of history in favor of the German soldier at the expense of his American
counterpart "is not at all ameliorated by the fact that a significant fraction
of the public buying World War II books consists of enthusiasts who collect
Nazi memorabilia, construct plastic panzers, and energetically seek to be the
German player in hex-grid war games." This work has already shown that Brown's
characterization of Marshall is totally without foundation. One also wonders
what scientific investigation formed the basis of Brown's conclusions about the
character of the "significant fraction" of purchasers of books about the Second
World War he describes. And even if his statements about these so-called
"enthusiasts" were true, one would want to know precisely what such "facts"
have to do with the issue at hand in any case. In fact, there is little to
support the "thesis" urged by Brown's "analysis" of Dupuy's work other than the
author's own venomous bias.(69)
What is it about Numbers, Predictions & War that so inflames the
critics of its author? In order to answer this question, it is necessary for
one to consider the work in some detail. Dupuy says that the purpose of the
book "is to describe the Quantified Judgment Method of Analysis of Historical
Combat Data" or QJMA, a major component of which "is a long but simple
mathematical equation" known as the Quantified Judgment Model ("QJM"). Of this
simple mathematical equation Brown concludes that it "simply demonstrates the
intellectual intimidation wrought when complex calculations are unleashed upon
a liberal arts community." On this point Dupuy opines in a rather different
tone than Brown. " Let me hasten to reassure those who feel at all
uncomfortable when faced with pages of complex mathematical equations; I feel
exactly the same way….I am not a mathematician, and all of the formulae which
appear in this book are merely means for organizing numbers or quantities, and
can be understood by anyone who is able to add, subtract, divide, and
multiply."(70)
Dupuy begins by pointing out that there are two principal problems in the
quantitative analysis of "war data", namely the reliability of the numbers, and
the interpretation of them. The first involves that which none of Dupuy's
critics has done---"time-consuming, frustrating, tedious, and expensive
research, with frequent comparison of differing data sources, different
engagements, and different campaigns." The second calls for a rejection of the
view that military history is irrelevant, embracing instead its opposite,
taking into account the significant impact upon modern warfare of new weapons
and technology, particularly those relating to transportation, observation and
communications. Dupuy then discusses at length the pratfalls attendant to the
use of "war data", describing in some detail a list of ten commonly presented
propositions often depicted as being based upon military experience. Among
these are the notion that as weapons have become more lethal, changes in
warfare have become correspondingly more radical. Dupuy rebuts this proposition
by showing that the technological change that has most influenced modern ground
warfare was the introduction of the muzzle-loading rifle musket, firing an
elongated bullet, a change that occurred in the decade between 1850 and 1860.
This change reversed the relation of lethal capability between artillery and
infantry weapons that had previously obtained, so that during the American
Civil War, small arms accounted for over 85% of the casualties, as opposed to
the 10% caused by cannon fire.(71)
Dupuy also disputes other commonly accepted maxims about military history.
Notable among these are the proposition that an attacker should have a
three-to-one superiority over the defender, that the numerically inferior force
is usually successful, and that modern technology permits faster advance rates
in combat. In each case, Dupuy marshals data to show the contrary of the axiom.
Dupuy engages in this exercise in order to "illustrate the problems of trying
to analyze trends in ground combat by making sense out of the anarchical masses
of data that lie in the dark and musty records of warfare", and to show why, in
view of the general inaccessibility of data, military historians and analysts
regularly engage in guesswork, assumptions and generalizations. In contrast,
Dupuy says, the purpose of Numbers is to translate the numbers of
military history into a coherent, consistent, quantitative theory of combat and
combat relationships.(72)
The first step taken by Dupuy in this regard is to analyze the effects of
weapons lethality upon combat. This comparison of weapons to casualties is not,
Dupuy points out, a simple matter of matching numbers to numbers; instead, the
analysis requires knowledge not only of the numbers of weapons, but also of
their various types and respective lethality. This latter Dupuy and his
colleagues define as the ability of a weapon to kill personnel and render
equipment ineffective in a given time period, where the capability of the
weapon depends upon weapon range, rate of fire, accuracy, radius of effects and
battlefield mobility. Since it was not possible to give precise values to the
effects of these variables, it was necessary to "postulate a standard,
theoretical, laboratory-like environment which could be common for all
weapons." This standard Dupuy calls the Theoretical Lethality Index (TLI), a
composite of thirteen variables common to all weapons, from the Roman short
sword to a nuclear warhead. The TLI of an individual weapon, however, had to be
converted into an Operational Lethality Index (OLI), which is done by applying
to the TLI the "Dispersion Factor", the area in square kilometers occupied by a
tactically deployed military force of 100,000 soldiers. Both the TLI and OLI
are values that assume "proving ground" circumstances; in order to find the
actual battlefield value of an individual weapon, it was necessary for Dupuy to
account for such variables as weather, terrain, season, mobility
characteristics and vulnerability, and to determine how these factors affected
weapon effectiveness.(73)
The "concept of variables…is the essence of the Quantified Judgment Model, and
of the Quantified Judgment Method of Analysis of Historical Combat Data." What
are the variables that form this essence, that define the battlefield values of
weapons? They fall into three categories, namely (a) environmental (affecting
weapon effectiveness) and operational (affecting employment of weapons and
forces) variables, (b) tabular (represented in simple tables) and formular
(represented in formulae) variables, and © tangible (representing specific,
quantifiable factors of practical effect on weapons) and intangible
(qualitative factors that resist quantification) variables. Dupuy and his
colleagues isolated seventy-three variables; among these are a number of fairly
obvious tangible, quantifiable factors such as the effects of terrain, weather,
season and air superiority, as well as offensive/defensive posture, all of
which are amenable to tabularization. Two other tangible variables, mobility
and vulnerability, involve the interaction of several variables, and must
therefore be expressed in formulae. More intangible, but no less significant
variables, include the concepts of leadership, training and morale, logistics,
time and space, momentum, intelligence, relative technological development and
initiative. A final intangible, combat effectiveness, represents an amalgam of
all or most of the others, among which the most important seem to be
leadership, training/experience, morale and logistics. As Dupuy points out
clearly, however, comparisons of combat effectiveness as between opposing
military forces represents "an oversimplified statement of a complex
relationship."(74)
From this point, Dupuy moves to the difficult task of constructing a model
within which the variables could be seen to operate. The notion that a model
might be useful was suggested to Dupuy and his colleagues during the course of
performing various investigations into historical data for, among other
clients, the United States Air Force. In large measure, however, the decision
to seek a model was taken because it was felt that members of the operations
research community, which included many who either doubted or rejected the idea
that history could teach anything meaningful on the subject of modern warfare,
could be appealed to best by resort to quantitative facts marshaled in support
of qualitative observation. This they would do by using the OLI to quantify the
total weapons firepower of opposing forces in a given engagement, and applying
to this figure reasonable factors for identifiable and quantifiable variables.
The first effort to do this was in connection with a study done for the British
Ministry of Defence, which inquired about the relationship between tactical air
support and land combat. In this particular study, Dupuy's group studied 60
division-sized engagements between September 1943 and June 1944 in the area of
operations of the U.S. Fifth Army in Italy. It is worth noting that the
analysts selected Italy because the ground and air operations of Fifth Army
were confined by the terrain to a narrow space, thus allowing more confident
assumptions about weapons interactions. The result of this study was the QJMA,
"a method of comparing the relative combat effectiveness of two opposing forces
in historical combat". In Dupuy's view, the QJMA resulted from a comparison of
(1) the Quantified Judgment Model formula for ascertaining the theoretical
winner of an engagement, in which historical data for the weapons inventories
of each side of an engagement are used to arrive at a figure for the Power or
Power Potential of a each force; the force enjoying a resulting ratio of 1 or
more would be regarded as the likely "winner"; and (2) the Result Model formula
for quantifying the actual outcome of the engagement; this model would in turn
take into account (a) the degree to which the respective sides achieved their
mission goals; (b) the ability of the respective sides to gain or hold ground,
taking into account their relative weapons inventories; and © the relative
"casualty effectiveness" of the two sides, arrived at by comparing casualties
to the starting strengths of the respective sides.(75)
The QJMA is ultimately based on three comparisons. In the first, the combat
power potentials of two opposing forces in a given engagement are compared.
These combat power potentials are derived by applying the effects of all of the
environmental and operational variables that can be identified to the
Operational Lethality Index (OLI) values for the total weapons inventories of
each combatant. As previously indicated, a ratio of greater than 1.0 is
regarded as predictive of success. In the second comparison, the actual
battlefield performance of the combatants is compared in the Result Model as
described above. The three performance criteria in the Result Model are
analyzed, and each side is assigned a resultant R value; the R values are
subtracted, and the side with the larger value is determined to be the actual
"winner" in the engagement. The final comparison is between the calculated
results of the two models. If the QJM ratio is greater than 1, then the Result
Model ratio should be positive; for example, when the ratio of power potential
between forces A and B is greater than 1 in favor of A, the ratio of AR minus
BR should be positive. Dupuy's analysis shows that this comparison is
consistent throughout; when it is not, it is "certain that the inconsistency is
due to some exceptional combat phenomenon, which is usually explicable after
further study and analysis. In fact, it has been through the exploration of the
causes of such inconsistencies that the value of the QJMA as an analytical tool
has been the greatest."(76)
Of particular importance to Dupuy's critics is his analysis of combat in World
War II. Criticism, in fact, formed a background for the development of the QJM
almost from the start, and found its focus on the subject of Dupuy's study of
the conflict in Italy in 1943-44. Foremost among the early criticism was the
notion that HERO's consideration of the 60 engagements involved amounted to
what Dupuy calls a "curve fitting" exercise. In this case, the term in question
supposes that if one employs as many variables as equations, the values of the
variables can be mathematically adjusted "to fit", thus potentially rendering
the variables meaningless for "equations" (or battles) fought under different
circumstances. Dupuy and his colleagues began to address this issue by
expanding their database. They first added data on eighteen engagements between
US and German formations in France between August and November 1944; they added
a further three engagements (at Kursk in July, 1943, in Italy in late 1944, and
in the Battle of the Bulge) by studying the use of obstacles and barriers in
World War II. Dupuy refers to these twenty-one engagements as the "Validating
Data Base" (in contrast to the "Development Data Base", a term used to describe
the original 60 engagements); nearly identical results were achieved by the
application of the same formulae, procedures and variables to both
databases.(77)
The problem with Dupuy's analysis, from the point of view of his critics, is
that it demonstrates an average German combat effectiveness superiority factor
of about 23 percent. More importantly, the study results indicate that this
superiority was manifest not only in the conditions prevalent in Italy, but
also in the very different situations presented by the Allied pursuit of the
German army in France during the summer and autumn of 1944, and in the Ardennes
in December of that year. Yet while the focus Dupuy's critics has been upon the
alleged unfairness of the study results, the author and his associates directed
their attention instead to the "problems" that the analysis uncovered. One of
these "problems" concerned the role of Allied airpower. On the one hand, it was
apparent that most deviations from the norm (of German combat superiority)
established by the overall study occurred where Allied airpower was either
unemployed or inconsequential; on the other, the study results showed that
Allied success was only assured (with some exceptions) where their overall
power superiority was very great. In cases where the power ratio favored the
Allies only marginally, the Germans were usually successful. In situations
where the power ratio suggested an indeterminate outcome, the Germans were
invariably the "winner", just as they were when the ratio was in their
favor.(78)
Dupuy and his associates had begun their investigations of the campaign in
Italy with an assumption that the Germans possessed a 10 percent advantage in
combat effectiveness over their American and British counterparts at the time
of the Salerno landings, based upon relative experience. They also believed,
however, that the Allies would erase this gap by the middle of 1944, an
estimate not proven by the study. Analyzing this problem, and the issue
described in the preceding paragraph, the Dupuy group found that the
consistency of most of their results involved a balance struck between their
underestimation, on the one hand, of German combat effectiveness, and on the
other of the effect of greater Allied air strength. As a result, the group
began applying a 1.2 relative combat effectiveness value to show German
superiority, as well a doubling of the values they had been applying for the
effects of air weapons.(79)
Dupuy's findings about German combat effectiveness in the Developmental Data
Base were confirmed in the analysis of the Validating Data Base. In an effort
to investigate the reason for consistent findings of German combat superiority,
the group made a comparison of fighting strength versus overhead for German and
American infantry divisions, based on tables of organization for 1943-44. This
comparison showed that in a German infantry division, 59.83 of the personnel
strength was involved in serving or manning weapons in a normal combat
situation. The relative number in an American infantry division was 50.26
percent. This comparison "suggests" that "part" of the finding of overall
German superiority "probably" was the result of better utilization of
manpower.(80)
A third "problem" encountered by the Dupuy group in their initial evaluation of
World War II data was the "perturbation" in results created by combat surprise.
On this subject, Dupuy discusses in particular the very poor showing of the
British 56th Division against an understrength German 65.Infanterie-Division
in an engagement along the Moletta River between 16-19 February 1944. The Dupuy
group concluded that the factor contributing substantially to the unanticipated
outcome was the surprise achieved by the Germans. This led the group to
theorize that three major effects proceed from tactical surprise. These
effects, which the group then incorporated into the QJM model, are (i) that the
mobility of the surprising force is enhanced because of its ability to position
its troops for optimum effect before the attack; (ii) that the vulnerability of
the surprised force is made worse by its opponent's ability to place fire
unexpectedly and accurately; and (iii) that the vulnerability of the surprising
force is reduced through its ability to more effectively plan and position its
troops. Although Dupuy observes that the general validity of this thesis was
confirmed for several other engagements in World War II, as well as in the
Arab-Israeli Wars, he cautions that "[T]here are, however, undoubtedly other
effects of surprise, and further research should be undertaken to attempt to
ascertain these."(81)
Of special significance with regard to the criticism leveled at Dupuy is his
conclusion about the probable outcome of the struggle in northwest Europe in
the summer of 1944. On the basis of their analyses of the engagements examined
in the Validating Data Base, the group concluded that "had the Allies made a
concentrated thrust on a relatively narrow single army front, the Germans would
have been unable to withstand it. I am now convinced that had this been done
the Allies could have at least reached the Rhine by September 1944. Beyond
that, of course, one can only speculate, but certainly under such circumstances
a German Ardennes offensive in December would have been impossible." Such
conclusions are manifestly not those of a person intent on denigrating the
combat prowess of the United States Army.(82)
Dupuy also addresses the issue of accounting for the effects upon a land battle
of air power. He identifies two conceptual problems in this regard. First,
there is the fact that aircraft "loiter" over a battlefield for only a fraction
of the engagement. Second, even when aircraft are in the area of the ground
battle, they may be assigned tasks other than ground support. The real
conundrum, however, is presented by the fact that even when not directly
engaged in ground support roles, aircraft may profoundly affect the ground
battle indirectly by, for example, interdicting the enemy's supplies away from
the battlefield. The issue presented by this latter phenomenon is how to take
into account such indirect ground support missions in the QJM, while at the
same time not including the weapons possessed by aircraft on such missions in
the battlefield inventory.(83)
The Dupuy group's assumptions concerning air interdiction were threefold: (1)
that the supply capability of the interdicted force is reduced through both the
destruction of supplies and the difficulties created for moving such supplies;
(2) that the ability of the interdicted force to move reinforcements as well as
troops actually engaged is impaired; and (3) that the damage done to the
interdicted force's communications impairs its command and control function.
Dupuy and his colleagues were only able to investigate the effect of
interdiction upon the supply function. When this element was considered in
connection with the Developmental Data Base (the original 60 engagements
studied), it was found that it was decisive in 25% of the engagements in which
interdiction effects were discernible. But whereas it was somewhat difficult
for the group to evaluate the effects of interdiction, the effects of direct
interaction between air and ground weapons were much more susceptible to
evaluation. The group identified eight different ways in which airpower affects
ground action; most of these effects increase the combat power of the side with
air superiority. The group applied the eight air superiority elements to the
Developmental Data Base, in which there were 38 instances of Allied success, as
well as seven in which the outcome was inconclusive. Dupuy and his colleagues
found that if the air component had been removed from these 45 engagements, a
German success would have been either predictable or very likely in 20-24 of
them. Put another way, in this subset of 45 engagements, "airpower provided the
margin which provided victory or prevented defeat" in at least 44% (perhaps as
high as 53%) of Allied successes and inconclusive engagements.(84)
Dupuy is frank to observe that his group's use of data related to the influence
of airpower on the ground battle has relevance to the conflict in Italy in
1944, because of the peculiar nature of that particular situation, viz., the
relative and absolute superiority of Allied airpower, and the methodical manner
in which it was brought to bear on the enemy. Accounting for the effects of
airpower is one of the areas upon which Dupuy's critics dwell. Others are what
Dupuy refers to as "fudge factors". These are principally the behavioral
variables of surprise and combat effectiveness, the significance of which
"cannot be perceived or appreciated unless the observer sees what the combat
power ratio would be without the factors." The attitude of Dupuy's critics is
that the HERO staff resorts to the use of these factors only when they are
needed to make the formulae work correctly. In fact, however, Dupuy points out
that the relevance of these factors was discovered only when they were ignored
in particular instances, with the result that peculiar outcomes developed.
Because of this, Dupuy and his colleagues automatically include a factor for
surprise when they are aware that it was achieved by one or the other side in
an engagement. Surprise as a verifiable phenomenon of significance in combat
has been illustrated by the Dupuy group in both Italy and France.(85)
The finding of superior German combat effectiveness, and its application in the
QJM, is the "fudge factor" that most infuriates critics of Dupuy and his work.
The group therefore had to address the question whether this was truly a
measureable factor, rather than an arbitrary "fudge factor" used to make the
use of the QJM and the results of that usage believable. They confirmed German
combat effectiveness by assessing casualty-inflicting capability, a measure
that first became apparent to them through a study of the American Civil War.
Dupuy and his group applied the same assessment method to both the First and
Second World Wars. In both of these struggles, German combat effectiveness was
superior to that of its western European and American opponents by nearly
identical figures. In the same manner, German combat effectiveness superiority
with respect to that of the Russians was nearly consistent over both World
Wars.(86)
Dupuy was quite aware of the criticisms leveled against his analysis. As he
observes, the most common critique of his work is that such an historical
approach is scientifically invalid. Dupuy points out, however, that while the
scientific techniques and experience of his technically-oriented critics (or
those who purport to rely upon such techniques and experience) are reliable in
dealing with scientific questions, they are less so when applied to human
behavior in the historical context. History, Dupuy argues, is the most reliable
guide to evaluating or indeed predicting human behavior. This, in his view,
gets at the crux of the matter. At the time Numbers, Predictions & War was
published, the most vehement critics came from the so-called OR (Operations
Research) community, a group composed of mathematicians and scientists whose
careers as analysts of current military affairs and supposed predictors of the
nature and outcome of future wars not only made them suspicious of
professionals with credentials other than their own, but also biased them
against the value of historical analysis and its applicability to the present
and future. Dupuy counters the OR critics by observing that (1) the comparative
analyses of the 1973 Arab-Israeli War by the OR community and the HERO group
showed that the real world validity of the mathematical models of the former
were no more realistic that the OLI representation of the QJM; (2) the
imprecision and unreliability of modern OR models has been pointed out by a
member of the OR community, Dr. J.A. Stockfish of the Rand Corporation, who
contended that modern mathematical models lack reliability at least in part due
to a failure to enlist empirical data; and (3) whereas the OR community was
unable to reproduce or represent modern war with modern weapons (as in the case
of the 1973 October War), the QJM model proved itself capable of doing so.
Dupuy also addresses criticisms of the QJM with particularity, with special
reference to the oft-repeated claim that the good results of the QJM can be
explained by the use of "manipulated data". There are, Dupuy says, three
different ways in which numbers could be manipulated: (a) weapons and other
variable values could be specifically selected to fit the circumstances of a
given engagement; (b) the results themselves could be made to fit the
preconceived notions harbored by the analyst, should the formulae not yield the
desired results; and © , in a case where the QJM and Result formulae had been
applied, one of the two truly judgmental inputs to the Result formula (assessed
mission accomplishment or the distance-advanced figure) could be changed.
Dupuy's response to these arguments is a simple one: All of the numbers,
formulae, values, tables and the like utilized in the QJM are fully
substantiated and justified in the historical record, "[A]nd that record is
there…for the review and scrutiny of anyone else who wishes to check on either
the data or the process."(87)
Numbers, Predictions & War is an affront to its critics not
because its methodology is faulty, as those critics would have us believe, but
because its conclusions about the relative combat effectiveness of the German
army and its Allied counterparts are unpalatable to them. The solution to the
critics' conundrum---the supposed invalidity of Dupuy's findings, resulting
from the application of flawed statistical analysis to suspect data---is for
the critics to correct the record by conducting their own statistical analysis
of the same (or "better") data, and publish the results for the profession to
scrutinize. This they have regrettably and unaccountably failed to do.
Footnotes
(67) Mansoor, pp. 6-8. Brown's Draftee Division is discussed more fully below
(68) Bonn, pp. 7-8.
(69) John Sloan Brown, Draftee Division (Presidio Press, 1998), pp. 168-169.
(70) Ibid., p. 169; Dupuy, pp. ix-xvii.
(71) Dupuy, pp. 3-9.
(72) Ibid., pp. 9-18.
(73) Ibid., pp. 19-30.
(74) Ibid., pp. 32-39.
(75) Ibid., pp. 40-42. It is worth observing that the "team" which worked on
HERO's Italian project on behalf of the British Ministry of Defence, and whose
efforts laid the groundwork for the QJMA, was not comprised of persons whose
knowledge of the subject matter was abstract. In addition to Col. Dupuy, the
"team" included Brig. Gen. Edwin S. Chickering, USAF, Ret.; Col. Ashton Crosby,
USA, Ret.; Col. Angus M. Fraser, USMC, Ret.; Col. Harold Quackenbush, USA,
Ret.; and Col. John A.C. Andrews, USAF, Ret.
(76) Ibid., pp. 55-56.
(77) Ibid., pp. 57-58.
(78) Ibid., p. 61.
(79) Ibid., pp. 61-62. Dupuy observes that "[W]e didn't like one of the two
conclusions which this adjustment forced upon us---that 100 Germans were
roughly the combat equivalent of 120 Americans or British---but we could not
ignore the fact that our numbers demonstrated that this was so."
(80) Ibid., pp. 62-63. It is in this context that Dupuy makes the comment that
"[T]he remainder (of the finding of German superiority) could possibly be the
result of such factors as more experience, greater mobility, better doctrine,
more effective battle drill, superior leadership, or inherent national
characteristics", a remark that has been singled out by at least one of his
critics as clear evidence of his pro-German bias. This critique ignores Dupuy's
very next sentence, in which he states that "[A] serious analysis of the
reasons for this greater German effectiveness is an important research
requirement", thereby confirming the author's view that the question at issue
is an unsettled one. Dupuy's comparison on the subject of combat personnel is
generally confirmed by both Martin van Creveld and Robert Sterling Rush.
(81) Ibid., pp. 63-64. Other engagements in which the effect of surprise proved
critical include the Anzio breakout of 23-25 May 1944 (American success),
Velletri, 26 May 1944 (German success), Ardennes-Sauer River engagement, 16-17
December 1944 (German success) and the Suez Canal Crossing of 6 October 1973
(Egyptian success).
(82) Ibid., p. 69. Ibid., pp. 71-72.
(83) Ibid., pp. 71-72.
(84) Ibid., pp. 76-78. The eight effects of airpower are: (1) the force
strengths of both sides are increased directly by the OLI value of direct air
support aircraft; (2) relative mobility is enhanced for the side with air
superiority; (3) vulnerability is reduced for the side with air superiority;
(4) the effectiveness of artillery is enhanced for the side with air
superiority; (5) vulnerability is increased for the side without air
superiority; (6) artillery effectiveness is reduced for the side without air
superiority; (7) direct air support is degraded for the side without air
superiority; and (8) supply capability is degraded by air interdiction.
(85) Ibid., pp. 95-97.
(86) Ibid., pp. 98-110. Noteworthy is the fact that HERO studied the Battle of
Kursk in 1979. In doing so, they checked the data given by Soviet secondary
sources with those shown in the captured German records at the National
Archives in Washington, D.C., a notion apparently never considered by
contemporaries of the HERO analysts, and most assuredly not by their critics.
Dupuy observes that "[I]n only one respect was there any significant difference
between the Soviet and German figures; the Soviets assessed the German
casualties as almost three times greater than the data recorded in the German
files, and showed German tank losses about five times greater than the German
figures. Since there was no reason for the Germans to falsify their records,
and since we have found their casualty and loss figures always quite reliable
in other theaters, we have taken their figures for these losses rather than the
Soviet assessments." No one in the historical community, least of all any of
Dupuy's critics, bothered to examine these selfsame records for over twenty
years, until the publication by Anders Frankson and Niklas Zetterling of Kursk
1943 (Frank Cass, 2000). Frankson and Zetterling confirmed both the
evidence used and the conclusions reached by Dupuy and his group concerning the
Battle of Kursk in every detail. It should also be observed that while Dupuy's
critics energetically dispute his findings with regard to German combat
effectiveness superiority, they do not do so with respect to his findings
concerning the combat effectiveness superiority of the Israelis in their wars
with the Arab states.
(87) Ibid., pp. 140-146.
Copyright © 2004 by Thomas E. Nutter.
Please send comments to Thomas E. Nutter at: tenutter@gmail.com.
|
|
|
|
