Book Review | Astral Codex Ten | How the War Was Won

Created

Nov 27, 2024 2:27 AM

The Wikipedia-Level Story of World War II (and O’Brien’s Counterargument)

To understand why O’Brien’s argument is so novel, you need to know the modern-day conventional understanding of the story of World War II. Here is my summary of the conventional narrative of World War II:

Germany conquered Poland and France. It tried to bomb the UK into submission/maybe enable an invasion. That effort failed when Germany was defeated in the Battle of Britain, thanks largely to the plucky efforts of British airmen (memorably summarized by Winston Churchill: “Never in the field of human conflict was so much owed by so many to so few.”)
Stymied in the West, Germany invaded the Soviet Union, won a bunch of crushing victories, but then got turned back at the gates of Moscow. The Soviets moved all of their factories east of the Ural Mountains and produced a vast tide of T-34 tanks that overpowered the Germans.
The Germans suffered a catastrophic defeat at Stalingrad and a bloody strategic defeat at Kursk, after which the Soviets relentlessly pounded Germany to defeat.
The US and the UK sent a lot of material help and eventually fought the Germans too, most notably in the D-Day invasion and the Battle of the Bulge. However, most of the fighting was done by the Soviets.
It is very difficult to say how important the aerial bombing campaigns of the Western Allies were in defeating Germany. The Germans moved much of their production underground, insulating them from truly disastrous effects.
The U.S. mostly fought alone against Japan, which won a series of impressive early victories (e.g., Pearl Harbor, the conquest of Singapore) until the decisive Battle of Midway, after which the vastly larger US industrial base outproduced Japan into oblivion.
The US bombed the Japanese into submission by destroying Japanese cities, ultimately by dropping atom bombs on Hiroshima and Nagasaki.

By examining where the Axis focused their productive capacities and how the Allies disrupted those capacities, O’Brien challenges virtually every part of that narrative:

The Battle of Britain was not a close-run thing. The fact that British fighter planes were flying over their own territory meant their attrition rate of pilots and aircraft were far lower than the Germans’.
American and British bombing mattered far more to the war’s outcome than the battles of the Eastern Front, which consumed a much smaller portion of German expenditures.
American and British airpower made German battlefield victories on the Western Front virtually impossible and dramatically limited the force Germany could bring to bear in the East.
Japan (really, Japan plus the giant empire it conquered at the beginning of the war) was an industrial behemoth to rival the Soviet Union. However, the destruction of the Japanese merchant fleet by American air and sea forces wrecked Japan’s economy.
The firebombing of Japanese cities and the atomic bombings of Hiroshima and Nagasaki had an ambiguous strategic effect. American air power played a much more important role in severing Japan from the natural resources it had conquered in the early part of the war.

Battles are Overrated

Take another look at the conventional narrative. Almost every key event involves a battle, a period of time in a relatively localized area where combatants slugged it out to see who would occupy some bit of land or sea. To O’Brien, this focus is silly, a relic of long-ago wars in ages with far less industrial capacity.

Start with theory. States fight to impose their will on another state in pursuit of some political goal. To do that requires that they achieve sufficient local military superiority that the other state can’t stop them from achieving their political goal.

Nazi Germany wanted to be the new administrators of the agricultural area of the western Soviet Union. To do that, they had to evict the Soviet military, whether through direct destruction or forcing the Soviet government to withdraw their armed forces. Individual battles for control of a localized area only matter if they are a means to that end.

Does the occupation or non-occupation of that point on the map affect the ability of a combatant to keep fighting?

In some limited cases, yes. Battlefield victory enabled Germany to overrun France before France could really focus its productive effort on the war. After their surrender, the French could not produce weapons, and they functionally could not organize their manpower to fight the Germans. But if the German army conquered, say, a random city in the Soviet Union, like Stalingrad, Soviet production and manpower was barely affected. The war goes on.

In theory, the German army could destroy so much of the Soviet military in one battle (or even a few discrete battles) that the Soviets run out of men or weapons. If there was ever a time this could have happened, it would have been the German invasion of the Soviet Union in 1941, when the Germans basically won a series of crushing victories.

The problem for the Germans was that by World War II, people in the combatant countries were good at building stuff in vast quantities, and the major combatants of World War II generally had access to sufficient natural resources. Even massive armies could not destroy produced weapons systems (e.g., tanks, airplanes) on the battlefield fast enough to remove the other side’s ability to continue fighting.What could (and did) happen was the destruction of the other side’s ability to produce and distribute weapons.

Sure enough, if you look at the actual data from even the largest battles, neither side really destroys a hugely significant amount of stuff. Take the Battle of Kursk—the largest tank and air battle of World War II. Wikipedia will dazzle you with the numbers of soldiers involved (millions), tanks deployed (in the ballpark of 10,000), and aircraft in the sky (in the ballpark of 5,000).

In this entire vast battle that supposedly dictated the outcome of the Eastern Front, the Germans lost approximately 350 armored fighting vehicles (AFVs) during the most intense 10 days of fighting. In the two months around when the battle took place, the Germans lost 1,331 AFVs on the entire Eastern Front. In the year of the battle, 1943, the Germans built more than 12,000 AFVs. Also worth noting: they disproportionately lost older, obsolete tanks at Kursk, and built new, capable tanks. The Germans lost a very manageable amount of equipment at Kursk—less than a month’s worth of AFV production.

If modern war means you cannot realistically destroy enough weapons in one battle to matter—if the largest battle of all time didn’t really matter—what did?

Allied Air and Sea Operations Won the War

In O’Brien’s methodology, we should look at what the Axis spent its productive effort making and consider what Allied actions slowed that productive effort. In both theaters, the answer is shocking. The Germans spent relatively little productive effort on tanks, focusing far more on aircraft, submarines, and vengeance weapons (i.e., proto-cruise missiles and rockets). The Japanese spent heavily on aircraft as well, but also a tremendous amount on freighters and oil tankers.

The Allies won the war by using air power to destroy the German and Japanese capacity both to produce military equipment and to transport it to the battlefield. By 1944-45, the Germans and Japanese could not use their economies to arm and supply their armies on the battlefield, leading to their inevitable defeat.

In the European war, American and British airpower: (a) directly destroyed a significant amount of productive capacity, (b) rendered remaining capacity far less efficient, (c) made it impossible for the Germans to defeat western ground forces, and (d) compelled the Germans to waste tremendous resources on air defense and exorbitant, ultimately ineffective vengeance weapons.

In the Pacific, the United States used carrier-based airpower, submarines, and bomber-deployed mines to isolate Japan from the resources of the empire it conquered in 1941-42. American bombers also directly destroyed factories and transportation systems, leading to similar levels of economic dysfunction as in Germany.

Amateurs Discuss Destruction; Professionals Discuss Non-Operational Losses

O’Brien is at his absolute best describing the subtle factors that whittled away Axis combat power. Air and sea power created a situation where the Axis war machine simply could not function anywhere near as efficiently as it needed to.

For example, after the Allied air bombings started, Germany built vast underground aircraft factories to protect production. But that move carried a host of negative side effects. To name a few:

The direct cost of building new factories in inconvenient places was very manpower intensive.
The old factories had been sited convenient to resource bases. The new factories were necessarily not near resource bases—they were in areas where one could dig out big new facilities.
Railroads, by far the most efficient means of transportation, were set up to efficiently move goods to and from the old factories, not the new ones.
Those factories had to be optimized for things like size and compactness, not efficiency and quality control. Aircraft frequently broke down on their way to the front lines. Once damaged, they could not be fixed on the front lines and were effectively useless.

These effects ultimately mean fewer airplanes produced as the war went on, and dramatic increases in non-operational losses. Citing the German field marshal in charge of aircraft production, O’Brien assessed that the Germans lost approximately half of their planned fighter production in this way. This comports with post-war American assessments, which assessed total German aircraft losses at the front as 15,327 in 1944, and non-operational losses at approximately 15,000. For comparison: total German aircraft losses at Kursk were approximately 159(!)

Data from HtWWW, recreated to improve image quality

The inefficiencies stemming from bombing ruined several would-be German technological panaceas. Germany developed the world’s first operational jet fighter, the Me-262. Lack of fuel meant there was not enough training for its pilots, and maintenance shortfalls meant that about half of the 1,400 Me-262s produced by Germany were lost outside of combat. The Germans developed a dangerous, relatively modern submarine, the Type XXI. They intended to deploy dozens in a way that the Allies would have been hard pressed to fight, but production delays meant that only one ever actually went on a mission.

Allied Bombings Provoked Vastly Expensive Reactions

O’Brien thoroughly documents how expensive Germany’s reaction to Allied bombings was. First, expenditures on anti-aircraft weaponry and fighter planes skyrocketed. The Germans practically denuded the Eastern Front of fighter planes to have more to throw at the bombers. By late 1944, a bare 15% of German aircraft were fighting on the Eastern Front. In the second half of 1943, significantly more concrete was devoted to the construction of protected aircraft factories in Germany than to the entire Eastern Front. The amount of concrete devoted just to protecting Hitler personally from air attack was almost a third of the entire total for fortifications on the Eastern Front

Second, and perhaps even more importantly, the bombings caused Hitler to authorize the most expensive German program of the war, the V-2 rocket, with essentially no goal in mind other than the psychological importance of striking back at Allied cities. The V-2 program cost the Germans proportionally as much as the U.S. spent on the Manhattan Project. According to O’Brien, the design and construction of V-2 rockets cost as much as all German AFV construction between 1939 and 1945(!)

It may be surprising to learn that the V-2s were basically irrelevant to the war. Launched primarily against UK cities, the V-2s killed several thousand civilians. However, more German slave laborers died building the V-2s than British civilians died from their use. The stupidity and expense of building the V-2 probably saved tens of thousands of lives elsewhere, which is ultimately yet another benefit of the Allied bombing campaign.

O’Brien’s production-focused approach yields some surprising insights about what the Germans should have done. The most cost-effective effort was certainly the use of submarines (U-Boats) to attack American shipments of military equipment across the Atlantic Ocean. For example, data suggest that the German navy destroyed at least twice as many American aircraft in the pre-production phase by destroying resource shipments as the German air force did in combat in 1942 and 1943.

Japan Was Far More Powerful Than We Usually Think

O’Brien goes to great lengths to illustrate that Japan was not just a small island power easily subsumed by American production. The Japanese economy, at its peak, produced about as much as the Soviet Union. Its industrial base was mostly untouched until mid-1944. In 1943, it produced as much steel as the Soviet Union. The Japanese navy’s planes doubled between 1943 and 1944.

Famously, the Soviets focused on producing tanks. The Japanese focused on freighters and oil tankers. They had to—they had gone to war to obtain natural resources by conquest away from their home islands, and to use those resources, they had to ship them back to the home islands. The problem was that once the American navy had conclusively defeated the Japanese navy (certainly no later than mid-1943), nothing could stop American submarines and carrier-based aircraft from savaging Japanese shipping.

But just as the bombing of Germany weakened German production in several complementary ways, the American war on Japanese shipping caused cascading logistical problems. For example, one very successful initiative was the aerial mining of Japanese ports. The mining didn’t start until March 1945, but it still sank more tonnage than U.S. submarines did in the entire war. Beyond that, the mining forced Japanese ships to use smaller, less efficient ports with bad communications and dock facilities, reducing the value of the small amount making it through to port.

The Morality of Strategic Bombing

One small but noteworthy argument in HtWWW relates to the “area” bombing of German cities, the firebombing of Tokyo, and the atomic bombings of Hiroshima and Nagasaki. Usually, air power enthusiasts are apologists for the indifferent (or even intentional) bombing of Axis civilians. They portray the fire/atomic bombings as difficult, but necessary and effective. O’Brien calls that logic into question.

As we’ve seen, strategic bombings that targeted specific factories or mined harbors were extremely useful. O’Brien writes, however, that civilian-centric bombing had ambiguous effects. Obviously, killing workers hurts productivity. But killing their spouses or children or destroying their houses does not immediately lead to unsolvable resource dilemmas.

It is perhaps too obvious to bear mentioning, but to the extent the civilian-centric bombings were not as effective as the rest of the strategic bombing campaign, they were immoral. O’Brien does not shy away from this conclusion, and shows a commendable willingness to gore sacred cows. He writes that Arthur Harris, leader of the British bombing campaign, resisted attempts to shift bombing away from cities generally and toward fuel or transportation targets, even when the evidence was clear that bombing was more effective. He takes the unusual step of effectively calling Churchill a moral coward:

From the autumn of 1944 onwards, it becomes difficult to justify any of the area attacks on German cities as important in winning the war. However, removing Harris, which might have allowed for such a change, was beyond the Churchill government’s courage.

O’Brien is similarly critical of Curtis LeMay, the American general who oversaw the firebombings. In his autobiography, LeMay justified the firebombings on the vague claim that they damaged Japanese morale. His evidence was a decline in Tokyo’s population, but population tended to decline after bombing raids anyway because production was relocated after raids. O’Brien concludes:

LeMay’s view of warfare was definitely a step backwards – and possibly self-defeating. His notion of causing justified destruction with little evidence beyond the physical action of destruction added an unnecessary air of irrationality to the American campaign.

Another important consideration in the debate over using the atomic bomb that I had not seen before: the firebombings were declining in effectiveness over time for the obvious reason that the best targets were already gone and the remaining cities were taking better precautions. The argument that firebombings alone would drive Japan to surrender without need of the atomic bomb must account for this awkward fact.

Death by Oil Austerity

Oil was a particular problem for Japan. The Japanese had gone to war with the United States in no small part because the U.S. cut off oil exports to Japan. The Japanese attempted to replace U.S. oil with oil from southeast Asia. Again, this was far less efficient than the pre-war arrangements, and once the U.S. Navy shut down shipping, the Japanese had to make drastic cuts to conserve oil.

Perhaps the single worst way to conserve oil was in flight training. The Japanese air forces entered a death spiral. To replace veteran flyers lost at Midway or off Guadalcanal, the Japanese parsimoniously supplied oil for limited training flight hours. This famously led to their being massacred by better-trained American pilots in the Great Marianas Turkey Shoot.

Another point brilliantly made by O’Brien: reductions in fuel expenditures meant Japanese pilots did not have sufficient training in navigation. Early in the war, aircraft were delivered to forward operating bases by aircraft carriers, limiting the ability of pilots to get lost. Once the American Navy had driven Japanese carriers from the scene, Japanese pilots had to make several over water hops to fly from the home islands to forward bases. Shocking numbers were lost along the way—up to 50%. (HALF) (!!!!) (I CAN’T EMPHASIZE ENOUGH HOW CRAZY THIS IS).

In addition to reduced training, Japan found another terrible way to conserve fuel: do not test engines for very long on the ground before sending single-engine fighter planes off to distant island deployments. Maintenance factors were one reason that on just one leg of the trip from Japan to forward bases, 5% of aircraft that took off from one island never landed at the next.

By 1945, the Japanese economy was so desperate for fuel that the government set up more than 34,000 small stills in the home islands to distill the oil from pine needles into aviation fuel.

In the European theater, the Allies specifically targeted German coal-to-oil conversion plants and Romanian oil facilities, which became far less productive. Over the course of one year, 1944, the western Allies destroyed the German energy market, and with it the German economy writ large.

Data from HtWWW, recreated to improve image quality.

German oil shortages caused exactly the same training problem Japan had faced, with a slightly different but similarly disastrous outcome. Japanese training and production problems led to planes not arriving where they were supposed to in fighting condition (perhaps as few as 10% were actually combat capable when they arrived!) For Germany, training shortfalls meant annihilation for their air force as inexperienced pilots were forced to fight numerically and qualitatively superior American and British pilots. German monthly aircraft lost/damaged rates increased from 52.5% in January 1944 to 96.3% in June.

One particularly illuminating episode illustrates how these problems manifested for Germany. The German air force had a reserve of 800 aircraft to counter the D-Day landings. The pilots of that force were used to only flying under expert control systems in Germany (countering bombing raids). When they went to France, they had trouble navigating and often landed on the wrong fields. Ultimately, they were poorly prepared to fight. The head of German fighter command was certain that the entire reserve did not destroy even two dozen Allied aircraft.

American/British Airpower Decided the Outcome of Land Battles

Beyond the strategic effects of bombing, tactical airpower (i.e., airplanes attacking land forces) gave an insurmountable advantage to the western Allies’ land forces. After D-Day, the Germans had a very strong defensive position in the hedgerows of northwest France. Allied aircraft literally carpet bombed one of the strongest divisions in the German army out of existence, with 70% casualties in one day. That division would normally have approximately 200 AFVs. At the end of that one day of bombing, it had 14.

The Battle of the Bulge, the last offensive by the Germans to drive back the western Allies’ advance, was almost pathetic in its hopelessness. We Americans tend to focus on the hard fighting at the outset of the battle, and the stout resistance of the 101st Airborne at Bastogne. Knowing that airpower would make their attack impossible, the Germans timed the battle for bad weather and prayed it lasted as long as possible. Prayer was really the only option. Once the skies inevitably cleared after a little over a week of bad weather, more than 2,000(!) Allied bombers destroyed the German offensive. With most logistical support wiped out, one famous German division had to abandon all its vehicles and walk back to Germany.

Criticism of HtWWW as a Book: Love the Data, (Mostly) Don’t Care About the People

My single biggest criticism of HtWWW is O’Brien spends a lot of time (I would estimate 20% of the book) discussing the relative importance and influence of various people in the United States and United Kingdom. The section on Doug MacArthur is worth a longer digression, which I have included below. The problem is that focusing on personnel is almost completely irrelevant to the main argument of the book.

For example, it is modestly interesting that Franklin Roosevelt, consistent with advice from Harry Hopkins and Admiral Ernest King, focused America’s productive effort on air and sea power. It is not at all central to the argument that air and sea power won the war. The fact that these particular people thought it was a good idea to build planes and ships matters less than the outcome that the U.S. did exactly that.

I am very much interested in World War II history, and on an interestingness scale of 1-10, I found this discussion to be at about a 4. The central argument of the book about German and Japanese production was a consistent 10.

Sidenote: MacArthur Was a Disastrous General

In the part of the book focused on personnel, the one discussion that hit around a 9 or 10 was of Douglas MacArthur and the invasion of the Philippines. MacArthur was the American general commanding the defense of the Philippines. The Japanese conquered the Philippines, and MacArthur slipped away to Australia, heroically vowing, “I shall return.” He did in December 1944, and some of the worst fighting of the war took place, with massive casualties for the Americans, Japanese, and Filipino civilians. Fighting was still ongoing in the Philippines when the war ended in August 1945. The Americans took more than 220,000 casualties, the Japanese 430,000. Estimates vary on Filipino civilian deaths, but 750,000 is a credible middle of the road estimate.

O’Brien’s contribution here was pointing out the strategic pointlessness of MacArthur’s invasion. The big American strategy in the western Pacific was to penetrate the Japanese defensive line of islands to link up with China. The northern Marianas Islands also were within heavy bomber range of Japan, and so would allow for efficient, effective bombing. (Bombing Japan from bases in China were logistically impractical, with virtually all materials being flown in over the Himalayas—another fascinating logistics discussion in this book.)

The Americans had already conquered the Marianas Islands and had total air and sea dominance in the western Pacific. The forces the Japanese had in the Philippines could have been simply left to wither, as they had been on other islands bypassed by the island-hopping campaign.

So, why did the Philippines invasion happen? The inescapable conclusion is that MacArthur was too politically formidable to risk angering, and he personally wanted to invade the Philippines to make good on his promise to return. Not coincidentally, the Philippines also offered some prospect of an extended land campaign where MacArthur could improve his reputation after his disastrous original defense of the Philippines.

Also relevant, in O’Brien’s words: “MacArthur [] dazzled Roosevelt with tales of easy victories and grateful Filipinos and American voters.”

Criticisms of HtWWW’s Central Argument

I think it is clear from the data that O’Brien’s argument, that air and sea power played a more important role than land battles in deciding the war, is fundamentally right. Still, one can raise a few objections.

Individual naval battles were capable of destroying a significant percentage of overall production. O’Brien discusses the Battle of Midway, where the Japanese lost four aircraft carriers (37 percent of their navy’s aircraft carriers at the time, 22 percent of all carriers they had during the war). This point doesn’t really disprove O’Brien’s core argument—it is basically a footnote saying that individual naval battles are more likely to matter than individual land battles.

Politics and psychology matter tremendously in war, sometimes more than productive effort. O’Brien tacitly acknowledges this in the V-2 weapons discussion when he notes that the Germans spent all this money and effort on a psychological salve to the trauma of Allied bombing. The Japanese did ultimately surrender after the atomic bombings. (Or, if you are more on the revisionist end of the spectrum, they surrendered after the Soviets declared war.) France surrendered after a few disastrous battles. The productive effort lens might be useful, but subject to important caveats.

Why Does the Conventional Narrative Focus on Battles?

A perfect companion book to HtWWW would examine why military historians and the broader public have focused inordinately on battles. Here are some plausible factors:

Battles are more dramatic. Propaganda during the war focused on battles so that there would be more inherent drama. Working twelve hour shifts in a factory to win the great battle is probably psychologically easier than thinking your work is going to disappear into an inchoate slog.
The battle-focused narrative empowers the blue-collar men who did the hard fighting. This is politically convenient. Those blue-collar men vote and do not want to be told that what they did was relatively less important. It is also commercially advisable to play up the everyman. People want to watch movies and buy books showing people like them (or their fathers or grandfathers) making an individual difference in the war. They don’t want to watch a movie showing how important it was that the bomber offensive led to Germany decentralizing German manufacturing, thereby increasing the attrition rate of German fighters on their way to the front.
People want to believe that individual effort matters. They want to believe that the endless suffering means something, that their loved ones’ deaths meant the world could be free. This is particularly true in the heartbreaking Soviet context where, for example, 80 percent of males born in 1923 did not survive the war.

Broadly Applicable Lessons from HtWWW

One can obviously draw specifically military lessons from HtWWW. O’Brien concludes his book by noting that Allied air and sea power fundamentally took away the ability of the Germans and Japanese to move resources or even their own armed forces. Even horrifically destructive battles cannot do this on their own. O’Brien notes that the U.S. won many battles in Afghanistan, but because the Taliban could still move forces and resources, they still won.

Another obvious lesson: target productivity, not forces. Destroying a tank on the battlefield is the least efficient way of defeating an enemy than sapping away his ability to replace losses. Target as many different parts of the supply chain as possible. There truly was no single shortage that doomed the Axis. Oil is arguably the closest tangible resource. The sense one gets from HtWWW is that the Allies’ “solution” of air and sea power was robust; there truly was no single thing that could have changed the outcome of the war. Conversely, a war based on battlefield destruction alone quickly turns into a bloody quagmire, like World War I.

Abstracting away from just military operations, the complexity of the modern world means that success or failure rarely hinges on a single factor. It is not impossible for a single technology or genius leader to accomplish great things, but it is far likelier that success is built on interlocking efforts aimed at the same general goal. As I read HtWWW, I thought about Amazon, Google, and Apple. None of those companies are built on a single “killer app”. They are built on doing a lot of little things well.

Finally, we should be wary of romanticized narratives and simplistic conclusions about why something happened. Finding the true story takes a nuanced, un-biased interpretation of both data and first-hand accounts. HtWWW is a fantastic example of that.

Bonus: Great Statistics/Logistics Stories I Couldn’t Work in Anywhere Else

HtWWW spoils the reader with an endless parade of fascinating statistics. As I said in the opening, when I first read the book, I excitedly texted friends these numbers every few minutes as I was reading. These stories didn’t fit in my review, but I could not in good conscience omit them. Below are just a few.

Germany, Japan, the United Kingdom (including the British Empire), and the United States all devoted between 65 and 80 percent of their economic output to the making and arming of aircraft, naval vessels, and anti-aircraft equipment.
During the three most intense months of the Battle of Britain, the German air force landed only 17 bombs that caused “severe” damage to aircraft and aircraft engine production, electricity services, gas supplies, water industry, oil infrastructure, and food service industries combined. This underscores just how ineffective Germany’s campaign was and how little chance there was a of an actual British defeat.
According to Albert Speer, the German minister for armaments during the war, the Germans had 2.33 million workers building aircraft, which was not only more than all the works employed building all weapons and ammunition for the army, it was equal to or more than the number of American workers building aircraft. Germany built about 65,000 planes in 1943-44. The United States built 182,000.
On December 27, 1941, when the climactic battle outside Moscow was taking place, the Germans had deployed almost a hundred more aircraft to the fight the Royal Air Force than they had on the Eastern Front.
Railway activity in Germany and occupied territories declined by almost 40% between August and December 1944.
The Japanese navy lost 50% more pilots at the Battle of Santa Cruz (off Guadalcanal) than at Midway—and that was only one of six major naval battles in the Solomon Islands.

Subscribe to Astral Codex Ten

P(A|B) = [P(A)*P(B|A)]/P(B), all the rest is commentary.