This concept - we have no records means nothing happened - is an assumption - and not a reasonable one either.
I agree. At the same time, just because the records were all destroyed does not mean we can assume that every wild assed claim by Sakai, or Fuchida, or whomever, is necessarily "good" information by virtue of being the only information from the power. Fuchida's book about Midway, for exaample, has been for western scholars the definitive treatment of the battle until Parshall and Tulley's book. What Japanese sources are available make Fuchida out to be a self-serving liar -- the ultimate fanboy whitewashing the Japanese navy's record to make them seem the victims of fate. Sakai's account (Zero Pilot) has also been demonstrated false on numerous occasions, as have late war accounts of single Japanese pilots downing numerous F6Fs in a single mission. All are now known to be false.
Spence's torpedo analysis is probably incomplete but at least he's doing the sort of "Well, what really happened?" analyses that need to be done to improve the history (or at least games that purport to be about things that may reasonably construed to logically follow from ahistorical choices). His "opposition" vis the torpedo debate remains wedded to unsubstantiated assertions and smary rhetoric as a form of rebuttal.
The truth is that most shots fired as torpedoes at most *combat* vessels missed when they were deplpoyed by aircraft or surface combat vessels. The only ships with decent torpedo *hit* rates were submarines (although the US early war dud rate obviated that considerably).
For example the mean IJN Type 93 hit rate was, statistically speaking, zero percent hits, as were the median and the mode. In a night surface action at ranges less than 9,000 yards the mean hit rate was about 12%. By 1943 the USN was doing as well with their slower, shorter-ranged torpedoes (largely because radar gave USN ships bearing, range, and speed, before their victims were being shot at). But if you really wanted to *regularly* hit the target at which you shot, you had to be less than 2000 yards from the target and the erstwhile victim had to be unaware that they were probably being engaged. The only ships really capable of achieving that sort of success were submarines.
Think about it: if a torpedo attack at sea is successful - you might not even hear about it at all.
A submarine torpedo attack at night ... maybe. Otherwise no. Any aerial attack occurring against a radar bearing USN ship would have been reported by radio. It's not logical or reasonable to assume that there is some statistically significant body of "successful Japanese aerial torpedo attacks against combat vessels." The overwhelming preponderance of "missing, presumed lost" ships were Japanese ships, and most of the Allied and Japanese ships so lost were submarines. There are a very very few notable exceptions in surfaqce combat, but not enough to warrant much skepticism over Spence's conclusions.
If you want to refute his argument you'll have to do something both useful and substantive and add to the database using credible sources, rather than attempting a sort of rebuttal by rumour.
Show me a fellow who rejects statistical analysis a priori and I'll show you a fellow who has no knowledge of statistics.
Didn't we have this conversation already?
