Category Research & Analysis

Reviews of NPW

I was just surfing Amazon.com yesterday and noticed that Trevor Dupuy’s book Numbers, Predictions and Wars had 8 reviews. This is odd as I think the book went out of print before Amazon.com existed. The version they were reviewing was the hardcover from 1979. All the reviews were from 2012-2018, all four and five stars. Interesting. Even I don’t have a copy of the hardback version.

Numbers, Predictions & War

Now, if only people would pay more attention to his greatest work, Understanding War. It only has two reviews.

Understanding War

The Battle of Britain Data Base

The Battle of Britain data base came into existence at the request of OSD PA&E (Office of the Secretary of Defense, Program Analysis and Evaluation). They contacted us. They were working with LMI (Logistics Management Institute, on of a dozen FFRDCs) to develop an air combat model. They felt that the Battle of Britain would be perfect for helping to develop, test and validate their model. The effort was led by a retired Air Force colonel who had the misfortune of spending part of his career in North Vietnam.

The problem with developing any air campaign database is that, unlike the German army, the Luftwaffe actually followed their orders late in the war to destroy their records. I understand from conversations with Trevor Dupuy that Luftwaffe records were stored in a train and had been moved to the German countryside (to get them away from the bombing and/or advancing armies). They then burned all the records there at the rail siding.

So, when HERO (Trevor Dupuy’s Historical Evaluation Research Organization) did their work on the Italian Campaign (which was funded by the Air Force), they had to find records on the German air activity with the Luftwaffe liaison officers of the German armies involved. The same with Kursk, where one of the few air records we had was with the air liaison officer to the German Second Army. This was the army on the tip of the bulge that was simply holding in place during the battle. It was the only source that gave us a daily count of sorties, German losses, etc. Of the eight or so full wings that were involved in the battle from the VIII Air Corps, we had records for one group of He-111s (there were usually three groups to a wing). We did have good records from the Soviet archives. But it hard to assemble a good picture of the German side of the battle with records from only 1/24th of the units involved. So the very limited surviving files of the Luftwaffe air liaison officers was all we had to work with for Italy and Kursk. We did not even have that for the Ardennes. Luckily the German air force simplified things by flying almost no missions until the disastrous Operation Bodenplatte on 1 January 1945. Of course, we had great records from the U.S. and the UK, but….hard to develop a good database without records from both sides. Therefore, one is left with few well-documented air battles anywhere for use in developing, evaluating and validating an air campaign model.

The exception is the Battle of Britain, which has been so well researched, and extensively written about, that it is possible to assemble an accurate and detailed daily account for both sides for every day of the battle. There are also a few surviving records that can be tapped, including the personal kill records of the pilots, the aircraft loss reports of the quartermaster, and the ULTRA reports of intercepted German radio messages. Therefore, we (mostly Richard Anderson) assembled the Battle of Britain data base from British unit records and the surviving records and the extensive secondary sources for the German side. We have already done considerable preliminary research covering 15 August to 19 September 1940 as a result of our work on DACM (Dupuy Air Combat Model)

The Dupuy Air Campaign Model (DACM)

The database covered the period from 8 August to 30 September 1940. It was programmed in Access by Jay Karamales.  From April to July 2004 we did a feasibility study for LMI. We were awarded a contract from OSD PA&E on 1 September to start work on the database. We sent a two-person research team to the British National Archives in Kew Gardens, London. There we examined 249 document files and copied 4,443 pages. The completed database and supporting documentation was delivered to OSD PA&E in August 2005. It was certainly the easiest of our campaign databases to do.

We do not know if OSD PA&E or LMI ever used the data base, but we think not. The database was ordered while they were still working on the model. After we delivered the database to them, we do not know what happened. We suspect the model was never completed and the effort was halted. The database has never been publically available. PA&E became defunct in 2009 and was replaced by CAPE (Cost Assessment and Program Evaluation). We may be the only people who still have (or can find) a copy of this database.

I will provide a more detailed description of this database in a later post.

Has The Army Given Up On Counterinsurgency Research, Again?

Mind-the-Gap

[In light of the U.S. Army’s recent publication of a history of it’s involvement in Iraq from 2003 to 2011, it may be relevant to re-post this piece from from 29 June 2016.]

As Chris Lawrence mentioned yesterday, retired Brigadier General John Hanley’s review of America’s Modern Wars in the current edition of Military Review concluded by pointing out the importance of a solid empirical basis for staff planning support for reliable military decision-making. This notion seems so obvious as to be a truism, but in reality, the U.S. Army has demonstrated no serious interest in remedying the weaknesses or gaps in the base of knowledge underpinning its basic concepts and doctrine.

In 2012, Major James A. Zanella published a monograph for the School of Advanced Military Studies of the U.S. Army Command and General Staff College (graduates of which are known informally as “Jedi Knights”), which examined problems the Army has had with estimating force requirements, particularly in recent stability and counterinsurgency efforts.

Historically, the United States military has had difficulty articulating and justifying force requirements to civilian decision makers. Since at least 1975, governmental officials and civilian analysts have consistently criticized the military for inadequate planning and execution. Most recently, the wars in Afghanistan and Iraq reinvigorated the debate over the proper identification of force requirements…Because Army planners have failed numerous times to provide force estimates acceptable to the President, the question arises, why are the planning methods inadequate and why have they not been improved?[1]

Zanella surveyed the various available Army planning tools and methodologies for determining force requirements, but found them all either inappropriate or only marginally applicable, or unsupported by any real-world data. He concluded

Considering the limitations of Army force planning methods, it is fair to conclude that Army force estimates have failed to persuade civilian decision-makers because the advice is not supported by a consistent valid method for estimating the force requirements… What is clear is that the current methods have utility when dealing with military situations that mirror the conditions represented by each model. In the contemporary military operating environment, the doctrinal models no longer fit.[2]

Zanella did identify the existence of recent, relevant empirical studies on manpower and counterinsurgency. He noted that “the existing doctrine on force requirements does not benefit from recent research” but suggested optimistically that it could provide “the Army with new tools to reinvigorate the discussion of troops-to-task calculations.”[3] Even before Zanella published his monograph, however, the Defense Department began removing any detailed reference or discussion about force requirements in counterinsurgency from Army and Joint doctrinal publications.

As Zanella discussed, there is a body of recent empirical research on manpower and counterinsurgency that contains a variety of valid and useful insights, but as I recently discussed, it does not yet offer definitive conclusions. Much more research and analysis is needed before the conclusions can be counted on as a valid and justifiably reliable basis for life and death decision-making. Yet, the last of these government sponsored studies was completed in 2010. Neither the Army nor any other organization in the U.S. government has funded any follow-on work on this subject and none appears forthcoming. This boom-or-bust pattern is nothing new, but the failure to do anything about it is becoming less and less understandable.

NOTES

[1] Major James A. Zanella, “Combat Power Analysis is Combat Power Density” (Ft. Leavenworth, KS: School of Advanced Military Studies, U.S. Army Command and General Staff College, 2012), pp. 1-2.

[2] Ibid, 50.

[3] Ibid, 47.

Historians and the Early Era of U.S. Army Operations Research

While perusing Charles Shrader’s fascinating history of the U.S. Army’s experience with operations research (OR), I came across several references to the part played by historians and historical analysis in early era of that effort.

The ground forces were the last branch of the Army to incorporate OR into their efforts during World War II, lagging behind the Army Air Forces, the technical services, and the Navy. Where the Army was a step ahead, however, was in creating a robust wartime historical field history documentation program. (After the war, this enabled the publication of the U.S. Army in World War II series, known as the “Green Books,” which set a new standard for government sponsored military histories.)

As Shrader related, the first OR personnel the Army deployed forward in 1944-45 often crossed paths with War Department General Staff Historical Branch field historian detachments. They both engaged in similar activities: collecting data on real-world combat operations, which was then analyzed and used for studies and reports written for the use of the commands to which they were assigned. The only significant difference was in their respective methodologies, with the historians using historical methods and the OR analysts using mathematical and scientific tools.

History and OR after World War II

The usefulness of historical approaches to collecting operational data did not go unnoticed by the OR practitioners, according to Shrader. When the Army established the Operations Research Office (ORO) in 1948, it hired a contingent of historians specifically for the purpose of facilitating research and analysis using WWII Army records, “the most likely source for data on operational matters.”

When the Korean War broke out in 1950, ORO sent eight multi-disciplinary teams, including the historians, to collect operational data and provide analytical support for U.S. By 1953, half of ORO’s personnel had spent time in combat zones. Throughout the 1950s, about 40-43% of ORO’s staff was comprised of specialists in the social sciences, history, business, literature, and law. Shrader quoted one leading ORO analyst as noting that, “there is reason to believe that the lawyer, social scientist or historian is better equipped professionally to evaluate evidence which is derived from the mind and experience of the human species.”

Among the notable historians who worked at or with ORO was Dr. Hugh M. Cole, an Army officer who had served as a staff historian for General George Patton during World War II. Cole rose to become a senior manager at ORO and later served as vice-president and president of ORO’s successor, the Research Analysis Corporation (RAC). Cole brought in WWII colleague Forrest C. Pogue (best known as the biographer of General George C. Marshall) and Charles B. MacDonald. ORO also employed another WWII field historian, the controversial S. L. A. Marshall, as a consultant during the Korean War. Dorothy Kneeland Clark did pioneering historical analysis on combat phenomena while at ORO.

The Demise of ORO…and Historical Combat Analysis?

By the late 1950s, considerable institutional friction had developed between ORO, the Johns Hopkins University (JHU)—ORO’s institutional owner—and the Army. According to Shrader,

Continued distrust of operations analysts by Army personnel, questions about the timeliness and focus of ORO studies, the ever-expanding scope of ORO interests, and, above all, [ORO director] Ellis Johnson’s irascible personality caused tensions that led in August 1961 to the cancellation of the Army’s contract with JHU and the replacement of ORO with a new, independent research organization, the Research Analysis Corporation [RAC].

RAC inherited ORO’s research agenda and most of its personnel, but changing events and circumstances led Army OR to shift its priorities away from field collection and empirical research on operational combat data in favor of the use of modeling and wargaming in its analyses. As Chris Lawrence described in his history of federally-funded Defense Department “think tanks,” the rise and fall of scientific management in DOD, the Vietnam War, social and congressional criticism, and an unhappiness by the military services with the analysis led to retrenchment in military OR by the end of the 60s. The Army sold RAC and created its own in-house Concepts Analysis Agency (CAA; now known as the Center for Army Analysis).

By the early 1970s, analysts, such as RAND’s Martin Shubik and Gary Brewer, and John Stockfisch, began to note that the relationships and processes being modeled in the Army’s combat simulations were not based on real-world data and that empirical research on combat phenomena by the Army OR community had languished. In 1991, Paul Davis and Donald Blumenthal gave this problem a name: the “Base of Sand.”

How Many Confederates Fought At Antietam?

Dead soldiers lying near the Dunker Church on the Antietam battlefield. [History.com]

Numbers matter in war and warfare. Armies cannot function effectively without reliable counts of manpower, weapons, supplies, and losses. Wars, campaigns, and battles are waged or avoided based on assessments of relative numerical strength. Possessing superior numbers, either overall or at the decisive point, is a commonly held axiom (if not a guarantor) for success in warfare.

These numbers of war likewise inform the judgements of historians. They play a large role in shaping historical understanding of who won or lost, and why. Armies and leaders possessing a numerical advantage are expected to succeed, and thus come under exacting scrutiny when they do not. Commanders and combatants who win in spite of inferiorities in numbers are lauded as geniuses or elite fighters.

Given the importance of numbers in war and history, however, it is surprising to see how often historians treat quantitative data carelessly. All too often, for example, historical estimates of troop strength are presented uncritically and often rounded off, apparently for simplicity’s sake. Otherwise careful scholars are not immune from the casual or sloppy use of numbers.

However, just as careless treatment of qualitative historical evidence results in bad history, the same goes for mishandling quantitative data. To be sure, like any historical evidence, quantitative data can be imprecise or simply inaccurate. Thus, as with any historical evidence, it is incumbent upon historians to analyze the numbers they use with methodological rigor.

OK, with that bit of throat-clearing out of the way, let me now proceed to jump into one of the greatest quantitative morasses in military historiography: strengths and losses in the American Civil War. Participants, pundits, and scholars have been arguing endlessly over numbers since before the war ended. And since nothing seems to get folks riled up more than debating Civil War numbers than arguing about the merits (or lack thereof) of Union General George B. McClellan, I am eventually going to add him to the mix as well.

The reason I am grabbing these dual lightning rods is to illustrate the challenges of quantitative data and historical analysis by looking at one of Trevor Dupuy’s favorite historical case studies, the Battle of Antietam (or Sharpsburg, for the unreconstructed rebels lurking out there). Dupuy cited his analysis of the battle in several of his books, mainly as a way of walking readers through his Quantified Judgement Method of Analysis (QJMA), and to demonstrate his concept of combat multipliers.

I have questions about his Antietam analysis that I will address later. To begin, however, I want to look at the force strength numbers he used. On p. 156 of Numbers, Predictions and War, he provided the following figures for the opposing armies at Antietam:The sources he cited for these figures were R. Ernest Dupuy and Trevor N. Dupuy, The Compact History of the Civil War (New York: Hawthorn, 1960) and Thomas L. Livermore, Numbers and Losses of the Civil War (reprint, Bloomington: University of Indiana, 1957).

It is with Livermore that I will begin tracing the historical and historiographical mystery of how many Confederates fought at the Battle of Antietam.

The 3-to-1 Rule in Recent History Books

This seems to be the rule that never goes away. I have a recent a case of it being used in a history book. The book was published in English in 2017 (and in German in 2007). In discussing the preparation for the Battle of Kursk in 1943 the author states that:

A military rule of thumb says an attacker should have a superiority of 3 to 1 in order to have a chance of success. While this vague principal applies only at tactical level, the superiority could be even greater if the defender is entrenched behind fortifications. Given the Kursk salient’s fortress-like defences, that was precisely the case.

This was drawn from Germany and the Second World War, Volume VIII: The Eastern Front 1943-1944: The War in the East and on the Neighboring Fronts, page 86. This section was written by Karl-Heinz Frieser.

This version of the rule now says that you have to have a superiority of 3-to-1 in order to have a chance of success? We have done a little analysis of force ratios compared to outcome. See Chapter 2: Force Ratios (pages 8-13) in War by Numbers. I never heard the caveat in the second sentence that the “principal applies only at tactical level.”

This rule has been discussed by me in previous blog posts. Dr. Frieser made a similar claim in his book The Blitzkrieg Legend:

The 3-to-1 Rule in Histories

These books were written by a German author who was an officer in the Bundeswehr, so apparently this rule of thumb has spread to some of our NATO allies, or maybe it started in Germany. We really don’t know where this rule of thumb first came from. It ain’t from Clausewitz.

Questioning The Validity Of The 3-1 Rule Of Combat

Canadian soldiers going “over the top” during an assault in the First World War. [History.com]
[This post was originally published on 1 December 2017.]

How many troops are needed to successfully attack or defend on the battlefield? There is a long-standing rule of thumb that holds that an attacker requires a 3-1 preponderance over a defender in combat in order to win. The aphorism is so widely accepted that few have questioned whether it is actually true or not.

Trevor Dupuy challenged the validity of the 3-1 rule on empirical grounds. He could find no historical substantiation to support it. In fact, his research on the question of force ratios suggested that there was a limit to the value of numerical preponderance on the battlefield.

TDI President Chris Lawrence has also challenged the 3-1 rule in his own work on the subject.

The validity of the 3-1 rule is no mere academic question. It underpins a great deal of U.S. military policy and warfighting doctrine. Yet, the only time the matter was seriously debated was in the 1980s with reference to the problem of defending Western Europe against the threat of Soviet military invasion.

It is probably long past due to seriously challenge the validity and usefulness of the 3-1 rule again.

Are There Only Three Ways of Assessing Military Power?

military-power[This article was originally posted on 11 October 2016]

In 2004, military analyst and academic Stephen Biddle published Military Power: Explaining Victory and Defeat in Modern Battle, a book that addressed the fundamental question of what causes victory and defeat in battle. Biddle took to task the study of the conduct of war, which he asserted was based on “a weak foundation” of empirical knowledge. He surveyed the existing literature on the topic and determined that the plethora of theories of military success or failure fell into one of three analytical categories: numerical preponderance, technological superiority, or force employment.

Numerical preponderance theories explain victory or defeat in terms of material advantage, with the winners possessing greater numbers of troops, populations, economic production, or financial expenditures. Many of these involve gross comparisons of numbers, but some of the more sophisticated analyses involve calculations of force density, force-to-space ratios, or measurements of quality-adjusted “combat power.” Notions of threshold “rules of thumb,” such as the 3-1 rule, arise from this. These sorts of measurements form the basis for many theories of power in the study of international relations.

The next most influential means of assessment, according to Biddle, involve views on the primacy of technology. One school, systemic technology theory, looks at how technological advances shift balances within the international system. The best example of this is how the introduction of machine guns in the late 19th century shifted the advantage in combat to the defender, and the development of the tank in the early 20th century shifted it back to the attacker. Such measures are influential in international relations and political science scholarship.

The other school of technological determinacy is dyadic technology theory, which looks at relative advantages between states regardless of posture. This usually involves detailed comparisons of specific weapons systems, tanks, aircraft, infantry weapons, ships, missiles, etc., with the edge going to the more sophisticated and capable technology. The use of Lanchester theory in operations research and combat modeling is rooted in this thinking.

Biddle identified the third category of assessment as subjective assessments of force employment based on non-material factors including tactics, doctrine, skill, experience, morale or leadership. Analyses on these lines are the stock-in-trade of military staff work, military historians, and strategic studies scholars. However, international relations theorists largely ignore force employment and operations research combat modelers tend to treat it as a constant or omit it because they believe its effects cannot be measured.

The common weakness of all of these approaches, Biddle argued, is that “there are differing views, each intuitively plausible but none of which can be considered empirically proven.” For example, no one has yet been able to find empirical support substantiating the validity of the 3-1 rule or Lanchester theory. Biddle notes that the track record for predictions based on force employment analyses has also been “poor.” (To be fair, the problem of testing theory to see if applies to the real world is not limited to assessments of military power, it afflicts security and strategic studies generally.)

So, is Biddle correct? Are there only three ways to assess military outcomes? Are they valid? Can we do better?

The (Missing) Urban Warfare Study

[This post was originally published on 13 December 2017]

And then…..we discovered the existence of a significant missing study that we wanted to see.

Around 2000, the Center for Army Analysis (CAA) contracted The Dupuy Institute to conduct an analysis of how to represent urban warfare in combat models. This was the first work we had ever done on urban warfare, so…….we first started our literature search. While there was a lot of impressionistic stuff gathered from reading about Stalingrad and watching field exercises, there was little hard data or analysis. Simply no one had ever done any analysis of the nature of urban warfare.

But, on the board of directors of The Dupuy Institute was a grand old gentleman called John Kettelle. He had previously been the president of Ketron, an operations research company that he had founded. Kettelle had been around the business for a while, having been an office mate of Kimball, of Morse and Kimbell fame (the people who wrote the original U.S. Operations Research “textbook” in 1951: Methods of Operations Research). He is here: https://www.adventfuneral.com/services/john-dunster-kettelle-jr.htm?wpmp_switcher=mobile

John had mentioned several times a massive study on urban warfare that he had done  for the U.S. Army in the 1970s. He had mentioned details of it, including that it was worked on by his staff over the course of several years, consisted of several volumes, looked into operations in Stalingrad, was pretty extensive and exhaustive, and had a civil disturbance component to it that he claimed was there at the request of the Nixon White House. John Kettelle sold off his company Ketron in the 1990s and was now semi-retired.

So, I asked John Kettelle where his study was. He said he did not know. He called over to the surviving elements of Ketron and they did not have a copy. Apparently significant parts of the study were classified. In our review of the urban warfare literature around 2000 we found no mention of the study or indications that anyone had seen or drawn any references from it.

This was probably the first extensive study ever done on urban warfare. It employed at least a half-dozen people for multiple years. Clearly the U.S. Army spent several million of our hard earned tax dollars on it…..yet is was not being used and could not be found. It was not listed in DTIC, NTIS, on the web, nor was it in Ketron’s files, and John Kettelle did not have a copy of it. It was lost !!!

So, we proceeded with our urban warfare studies independent of past research and ended up doing three reports on the subject. Theses studies are discussed in two chapters of my book War by Numbers.

All three studies are listed in our report list: http://www.dupuyinstitute.org/tdipub3.htm

The first one is available on line at:  http://www.dupuyinstitute.org/pdf/urbanwar.pdf

As the Ketron urban warfare study was classified, there were probably copies of it in classified U.S. Army command files in the 1970s. If these files have been properly retired then these classified files may exist in the archives. At some point, they may be declassified. At some point the study may be re-discovered. But……the U.S. Army after spending millions for this study, preceded to obtain no benefit from the study in the late 1990s, when a lot of people re-opened the issue of urban warfare. This would have certainly been a useful study, especially as much of what the Army, RAND and others were discussing at the time was not based upon hard data and was often dead wrong.

This may be a case of the U.S. Army having to re-invent the wheel because it has not done a good job of protecting and disseminating its studies and analysis. This seems to particularly be a problem with studies that were done by contractors that have gone out of business. Keep in mind, we were doing our urban warfare work for the Center for Army Analysis. As a minimum, they should have had a copy of it.

My Response To My 1997 Article

Shawn likes to post up on the blog old articles from The International TNDM Newsletter. The previous blog post was one such article I wrote in 1997 (he posted it under my name…although he put together the post). This is the first time I have read it since say….1997. A few comments:

  1. In fact, we did go back in systematically review and correct all the Italian engagements. This was primarily done by Richard Anderson from German records and UK records. All the UK engagements were revised as were many of the other Italian Campaign records. In fact, we ended up revising at least half of the WWII engagements in the Land Warfare Data Base (LWDB).
  2. We did greatly expand our collection of data, to over 1,200 engagements, including 752 in a division-level engagement database. Basically we doubled the size of the database (and placed it in Access).
  3. Using this more powerful data collection, I then re-shot the analysis of combat effectiveness. I did not use any modeling structure, but simply just used basic statistics. This effort again showed a performance difference in combat in Italy between the Germans, the Americans and the British. This is discussed in War by Numbers, pages 19-31.
  4. We did actually re-validate the TNDM. The results of this validation are published in War by Numbers, pages 299-324. They were separately validated at corps-level (WWII), division-level (WWII) and at Battalion-level (WWI, WWII and post-WWII).
  5. War by Numbers also includes a detailed discussion of differences in casualty reporting between nations (pages 202-205) and between services (pages 193-202).
  6. We have never done an analysis of the value of terrain using our larger more robust databases, although this is on my short-list of things to do. This is expected to be part of War by Numbers II, if I get around to writing it.
  7. We have done no significant re-design of the TNDM.

Anyhow, that is some of what we have been doing in the intervening 20 years since I wrote that article.