Contact Us   |    Join   |    Donate


Published by Alfred A. Knopf. February 2013 Reviewed by Dr. Brian McCue.

Brian McCue is a civilian naval analyst with experience on both coasts, most of it pertaining to the analysis of Anti-Submarine W01fare exercises and operations. He is the author of U-Boats in the Bay of Biscay: An Essay in Operations Analysis (National Defense University Press I 990; Alidade Press 2008) as well as numerous mathematical papers and articles regarding the Second World War campaign against German submarines. He holds a Ph.D. and a Master’s degree from the Massachusetts Institute of Technology, and a Bachelor’s degree from Hamilton College.

BLACKETT’S WAR is aptly titled, since it not exactly a biography of Patrick Maynard Stuart Blackett. Rather, it is a history of the part of the Second World War over which he held sway: the creation in Britain of operations analysis, the scientific study and improvement of military operations themselves, as distinct from the weapons with which they were waged. Blackett, an outstanding experimental physicist who-unlike many other “men of the professor type” recruited for the more esoteric aspects of the British war effort-was a combat veteran of the First World War and had seen action at Jutland, conceived of operational analysis early in the war, took part in its application to defending London against the Luftwaffe, and was then drawn into the war against the U-boats.

Budiansky gives an excellent feel for the style and substance of Blackett’s work in operations research against the U-boats by presenting three important problems in some detail.

The first of these was the matter of the correct depth settings for air-dropped depth charges. The U-boats as yet lacked snorkels, and could be sighted by airplanes while running on the surface, but were likely to counter-detect the airplane and be submerged by the time the airplane came overhead. For this reason, the airplanes became equipped with depth charges in place of bombs. Initially, the depth setting was I00-150′, based on an estimate of how much a U-boat could submerge during the average time-50 seconds-that elapsed between its disappearance beneath the waves and the arrival of the attacking bomber. In that time, of course, the U-boat could (and would) also move horizontally in unknown ways, and a considerable dispersion of the depth charges was therefore advised.

Few U-boats were damaged in these attacks.

E.J. Williams, a physicist working with Blackett, realized that to attack the average U-boat was so difficult that doing so should not be attempted. Accurate drops would only be had on boats whose submergence had been tardy, so the depth setting should be reduced, and the pattern more tightly concentrated. Williams’s recommendations were adopted, and brought about a major increase in the proportion of successful attacks.

The second, and longest-running, was the question of how the available heavy bombers could best be employed-in the bombing of Europe for which they had been built, or in the emergent task of fighting the U-boats. When American production began to provide bombers in profusion, there arose a follow-on question: assuming that bombers were to fight U-boats, how would they best do so? The candidate assignments were bombing the U-boat pens in occupied France, searching for U-boats transiting the Bay of Biscay, flying in direct protection of Atlantic convoys, or bombing the German shipyards in which U-boat were built. While Blackett’s statistical analysis, and some experimental raids, showed that to bomb the U-boat pens would be almost exactly fruitless, the questions of if and how bombers were to be used against U-boats were inextricably tied up with organizational, intra-service, inter-service, international, doctrinal, and even moral questions, amid which Blackett’s analyses could gain no traction. Something approximating the right answer-protection of convoys and offensive search in the Bay-came only through US insistence as to how American-built bombers (including Navy’s purpose-built B-24 Liberator variant, the PB4Y Privateer) were to be used.

The third was the question of the correct number of merchant vessels in a convoy. Blackett examined data from convoys of various sizes and concluded that the number of vessels Jost per convoy was nearly independent of the convoy’s size, a finding that meant large convoys were better because a smaller proportion of ships would be lost. This finding contradicted Royal Navy doctrine, and perhaps also some residue of the eggs-in-one-basket thinking that had incorrectly opposed convoys in the first place. Blackett himself, recognizing the drastic nature and great importance of the change, did not finalize his recommendation until he had subjected it to the mental test of asking himself what size convoy would he prefer his children to be in, were they to have to voyage to America. He made his recommendation, his advice was taken, and a dramatic reduction in the losses of convoyed ships ensued.

Budiansky’s wide-ranging narrative includes some of the story of American antisubmarine operations research as well, particularly the work of Philip Morse and William Shockley, who worked in a decidedly different manner from that of Blackett.

Blackett’s work on the question of convoy size had begun with his assignment to perform a statistical analysis of the effectiveness of the ships and airplanes that protected convoys, and even when he had realized the importance of the question of convoys’ size, he addressed it in a statistical fashion, using available data on convoys of various sizes, and tested it by thinking of how he would his children to cross the Atlantic. Only afterward did another analyst provide the reasoning: the convoy’s ships filled its interior whereas its escorts populated its perimeter, so the number of convoyed ships was proportional to the square of the number of escorts.

Morse’s task was presented to him similarly: ” … we were shown a room full of reports of all actions against submarines, real or imagined,” wrote Morse later. “We looked at a few reports and talked to some of the officers who had participated in U-boat sightings and attacks. And we said we wanted to think about the problem before we started to read.”1 “Morse’s team,” continues Budiansky,” went into a one-week huddle’ while they worked out, from first principles, a mathematical theory of submarine search.” Without explicit comment, Budiansky has hit upon a key trans-Atlantic difference in how operations research tended to be performed: Blackett and the other British workers acted as strict empiricists, working with numerical data to find important regularities, such as the fact that the number of ships sunk was independent of convoy size. A theory came later, if at all. The Americans, in contrast, began with a theory, and then turned to data to fill in the numerical details.

On the other hand, the work on each side of the Atlantic was similar in a surprising way: the best questions were not those handed down by the Services, but those that the civilian analysts found for themselves.

In the case of convoy-sizing, Budiansky cites Blackett’s observation that “As in most of the important cases … the really vital problems were found by the operations research groups themselves rather than given to them to solve by the Service operational staffs.”” Indeed, this observation would seem to apply to the depth-charge problem and the bomber-allocation problem as well, and it stands in stark contrast to the Services’ view that the scientists should speak only when spoken to: ” … they must stick to their lasts,” wrote Air Marshal John Slessor later, when describing his opposition to the recommendation of Blackett and Williams that the bombers be assigned to patrolling for U-boats in the Bay of Biscay. “Statistics are invaluable in war if they are properly used-in fact, you can’t fight a modern war without them. But the Bay offensive was a battle, and a bitterly contested one, and nothing could be more dangerously misleading than to imagine that you can forecast the result of a battle or decide the weapons necessary to use in it, by doing sums.

And yet a correct forecast was exactly what Williams and Blackett had provided in the cases of the depth charges and convoy-sizing, and events were to prove that their predictions regarding the Bay were more correct than not.


THE SUBMARINE REVIEW is a quarterly publication of the Naval Submarine League. It is a forum for discussion of submarine matters. Not only are the ideas of Its members to be reflected in the REVIEW. but those of others as well, who are interested in submarines and submarining.

Articles for this publication will be accepted on any subject closely related to submarine matters. Their length should be a maximum of about 2500 words. The League prepares REVIEW copy for publication using Word. If possible to do so, accompanying a submission with a CD is of significant assistance in that process. Editing of articles for clarity may be necessary, since important ideas should be readily understood by the readers of the REVIEW.

A stipend of up to $200.00 will be paid for each major article published. Articles accepted for publication In the REVIEW become the property of the Naval Submarine League. The views expressed by the authors are their own and are not to be construed to be those of the Naval Submarine League.

Comments on articles and brief discussion items are welcomed to make THE SUBMARINE REVIEW a dynamic reflection of the League’s interest in submarines.

Articles should be submitted to the Editor, SUBMARINE REVIEW, 50250 Backlick Road. Annandale, VA 22003-6044.

Naval Submarine League

© 2022 Naval Submarine League