In October 1944, the Imperial Japanese Navy sent 761 sorties against American carrier groups off Formosa over five days. Nearly 180 of those aircraft never returned, and another 150 were lost over the island or burned on the ground where they sat. Those numbers only surface weeks later, after someone has sifted through squadron returns and tallied the losses. On the day itself, what a navy has is aircrews coming back in the dark, a room with a table, and an officer asking each crew what they saw.
Vice Admiral Shigeru Fukudome watched the opening of that battle from a hillside in southern Formosa, his command post a cave dug into the rock about four thousand yards from the building his staff had been using. On the morning of October 12, a radar station on the east coast picked up a large formation approaching. Fukudome had 213 fighters up over Takao and Taihoku waiting for it. From the cave mouth, he saw aircraft catching fire and falling, and he was delighted.
He said so out loud. Then he looked harder and understood that every aircraft burning in the sky above him was his own, and that the tidy formations still circling overhead were American. He had expected the odds over his own island to run six to four in his favor. By his own later reckoning, they ran nine to one against him, and the naval air arsenal the Japanese had built at Takao came apart under a single heavy attack.
Fukudome published that account in December 1952, in the journal of the United States Naval Institute. The reliability of that article is what the whole story turns on, because what happened in the paperwork was stranger than what happened in the air. Somewhere in the reporting chain, a beating turned into the largest naval victory Japan had announced in two years. On October 16, Imperial General Headquarters told the country that twelve American warships of cruiser size or larger had been sunk and twenty-three more destroyed by other means.
The actual damage to the United States Third Fleet was two cruisers hurt hard enough to need towing. No American ship was sunk at all. Any country at war lies to its own people, and Japan had a great deal to lie about by the autumn of 1944. But the harder question is narrower.
How does a navy come to tell that story to itself, in its own signals, in the traffic its own admirals then act on? The answer lies in the shape of the fighting. The daylight attacks, flown in large formations, achieved almost nothing. Most crews never found the American ships at all.
Every hit that was claimed came from one specialist unit that attacked in darkness, a group built around torpedo aircraft and trained for filthy weather, sending out some thirty aircraft a night on three consecutive nights. Night attacks are where claims go wrong. A crew sees a flash on the water and cannot tell whether it was their own torpedo or somebody else’s gunfire. Two crews hit the same ship a few moments apart and come home having sunk two ships.
A destroyer becomes a cruiser. A cruiser becomes a battleship. And a battleship at night, with tracer coming up, becomes a carrier. Fukudome had seen the problem coming.
He had gone around his units asking individual aircrew whether any of them had ever laid eyes on a submarine, and plenty had never done so. Half the night unit was army aircrew, some with two months of training over water, some with six, against a naval assumption that a man needed three years before he was fit to attack ships in the dark. Their ability to identify what they were looking at was the part Fukudome later described as the most deplorable of all. The claims came in inflated, sent up by crews who were not competent to judge what they had done.
That much is ordinary. Aircrew overclaim in every air force that has ever existed, and staffs everywhere apply a discount. What matters here is what the claims were allowed to do next. The officer commanding the night unit was a captain named Kuno, and his force was based in Kyushu, several hundred miles from Fukudome’s cave.
His figures went by radio straight to the commander-in-chief of the Combined Fleet and to Imperial General Headquarters in Tokyo. At the same moment they went to Fukudome, who was his superior for the operation on paper. The claim passed through nobody. It went up three pipes at once, and the one man with the standing to slow it down received it as a copy after the fact.
By October 15, Admiral Toyoda, commanding the Combined Fleet, had decided the moment was too good to waste and ordered a general pursuit of the beaten American fleet. Surface units sitting in the Inland Sea, hundreds of miles from anything, were told to join in. The chase found empty ocean. Eight days later, the same fleet sailed into the Philippines for the Battle of Leyte Gulf carrying a picture of American carrier strength that bore no relation to the strength waiting there.
Ten months after that week of fighting, the Imperial Japanese Navy’s archive became a problem to be solved. The Japanese services had been thinking about how to destroy their own paperwork since well before the war. A February 1938 study sent to every army unit found that paper resists fire. A stack of bound volumes will smolder, char at the edges, and survive in the middle even when soaked in fuel, because flame needs air between the layers.
The recommended fix was to interleave the pages with celluloid sheets. As the war turned, the rules for using all this loosened in stages. In June 1944, units were told to send unneeded top-secret papers back for pulping. That November, commanders won the right to destroy classified files without waiting on approval from Tokyo.
Then came March 1945, with American forces about to land on Okinawa, and the geographic limits came off. The emergency rules now applied everywhere, at home as much as overseas. On August 14, 1945, the day before the surrender broadcast, the order itself went out. Every classified holding was to burn at once, and the telegram carried an instruction that it too was to burn on receipt.
Units worked at it for three weeks. A circular on August 20 complained that some files were half burned and others tucked in among personal papers. The fires went out on September 3, on an occupation order. By the autumn of 1945, Japan’s Navy had lost the war, lost most of its ships, and burned an unknown but enormous quantity of its own files.
The Americans, who badly wanted to understand what had happened on the other side of three and a half years of fighting, found themselves holding almost no enemy paper at all. What they had instead was the men themselves, alive in Tokyo and available. The interrogations began that October and ran to the end of December, the work of the Naval Analysis Division of the United States Strategic Bombing Survey. Men who had been trying to kill each other eight weeks earlier faced one another across a table with an interpreter in the middle, working over carrier movements and fuel states and why a given decision was taken on a given afternoon.
The transcripts went out in two volumes in 1946. Around the same time, a second effort got underway, much larger. General MacArthur’s headquarters ran a historical section, a hundred people at its peak. The section commissioned former enemy staff to write up their own operations in a long series of monographs.
Everyone involved understood the weakness of that method, and nobody had an alternative. Between roughly 1946 and 1955, the entire English-language account of the Japanese war was assembled out of the memories of the men who had fought it. The memoirs followed straight on. Fukudome’s account of Formosa appeared in December 1952.
The files that could have tested it did not reach Tokyo until April 1958, when the Washington Document Center shipped the bulk of its holdings back to Japan. In October 1955, the Japanese Defense Agency had opened a war history office, and when the crates arrived three years later, that office held something no memoirist had possessed and no American historian had been able to use. It had the operational paperwork of the Imperial Japanese Navy in the original, alongside the men still alive to explain it. The result came out between 1966 and 1980: 102 volumes.
One man, one week of fighting, two accounts seven years apart. The 1945 interrogation transcript already shows a commander who could tell one unit’s reporting from another’s. Asked what he had reported to Tokyo after the Formosa fighting, Fukudome answered that the night attacks had gone a good deal better than the daylight ones, and that his crews claimed three carriers and several other vessels sunk or damaged. Had he attached any qualifying note?
The reports came in from a captain in Kyushu, he said, and all he could do was gather them up and pass them along. On what the general staff made of the claim, he had no idea. Pressed on the overall effect of the American strikes, he allowed that the results may not have matched what the crews reported at the time, and then said he had felt considerable success had been attained. He also placed the American attack on Formosa on October 14, when it had come on the twelfth.
The 1952 article tells a different story. In that version, Kuno’s report arrives and Fukudome sees through it on site, knowing at once that the figures were far too high. He explains that he had already made a statistical study of how much aircrew overclaim, and that the study told him the true figure was about a third of what was coming in. He describes his own reconnaissance aircraft establishing on the sixteenth that the damage was slight, and says he was therefore certain the Philippines were about to be invaded and set about preparing for it.
The battle is dated down to the hour. The loss figures have hardened too. The 1945 answer gave a range. The article gives one flat number, tied to an exact count of sorties flown.
The later account is the tidier of the two, better organized, more exact, and it places Fukudome’s own understanding several days earlier. In 1945, he believed in the victory, passed the claim along, and got the date of his own battle wrong by two days. In 1952, he saw through it at once, had the mathematics to prove it, and knew what was coming next. The drift has a direction.
Across the Japanese naval memoirs of the early 1950s, wherever a contemporary file survives to check them against, the corrections tend to run the same way. The writer’s private doubts move earlier. His objections get louder. Warnings that appear nowhere in the signal traffic of the week become clear and firm in retrospect, and the errors belong to the system, to headquarters, to the other service.
The memoirs came out in English or went into it within a year or two, published by an American naval press, and they remain in print and cheap. The 102 volumes assembled from the paperwork exist in Japanese, and only a handful have ever been translated. An American reader who wants the Japanese side of the Pacific War can have it tonight in his own language, from the men who lost it.
The files that contradict them sit in a reading room in Tokyo, in a language he does not have.