Re: Mohawk v Hurricane record in Burma, on the details I counted in "Air War Burma" collision in combat, crashlanding, belly landing, etc on both sides as loss (as I do as standard). I just double checked it, 8 Type 1's (one a 'force land' but pilot died of wounds, I infer probable w/o), v. 7 Mohawks (one of which might have been an operational loss) in around 10 encounters.
But has anyone else studied statistics? Assume my counts are correct for the moment. Assume the number of kills and losses for both Mohawk and Hurricane v Type 1 is a flip of a fair coin, on average 1 heads (a kill) for every tails (a loss). 8 heads, 7 tails in 15 flips is then, obviously, quite a likely outcome if the underlying true probability is 1:1. But 12 heads and 55 tails in 67 flips OTOH is almost impossible to achieve by chance if the underlying probability is really 1:1, .000005%. The chance of even 24 or less heads (43 or more tails) in 67 flips, is only 1.3%, if the underlying probability is 1:1
By the same token if the real 'coin' is 4 times more likely to come up tails, ie 20% likely to come up heads, 12 and 55 is a pretty likely outcome, but 8 or more heads in 15 flips is only .05% likely.
So we can reject at a high level of confidence that the Hurricane and Mohawk chances of success were really the same but statistical noise, per se, made them appear that different. Even in that small a sample the difference is too big to be explained by statistical noise. OTOH we can't (and I didn't) say the Mohawk was 4 times as effective, because *some* of the difference could be explained by statistical noise. Systematic differences in the circumstances of Hurricane v Mohawk combats, not statistical noise per se, could explain the difference if it was the case, but I just don't see that in the book: it was basically similar operations v the same Japanese fighter units at the same time. Or, the Mowhawk pilots might have been better (though from the same AF, same time), etc... but it seems at that point we're grasping for straws to avoid the simplest conclusion from those results: the Hawk was at least as effective a fighter as the Hurricane. And that's what the larger sample in BoF also indicates v Bf109E (23:38 for Hawk, 74:151 for Hurricane, "Battle of France Then and Now"), but it's not AFAIK the general reputation of the Hawk.
Joe
But has anyone else studied statistics? Assume my counts are correct for the moment. Assume the number of kills and losses for both Mohawk and Hurricane v Type 1 is a flip of a fair coin, on average 1 heads (a kill) for every tails (a loss). 8 heads, 7 tails in 15 flips is then, obviously, quite a likely outcome if the underlying true probability is 1:1. But 12 heads and 55 tails in 67 flips OTOH is almost impossible to achieve by chance if the underlying probability is really 1:1, .000005%. The chance of even 24 or less heads (43 or more tails) in 67 flips, is only 1.3%, if the underlying probability is 1:1
By the same token if the real 'coin' is 4 times more likely to come up tails, ie 20% likely to come up heads, 12 and 55 is a pretty likely outcome, but 8 or more heads in 15 flips is only .05% likely.
So we can reject at a high level of confidence that the Hurricane and Mohawk chances of success were really the same but statistical noise, per se, made them appear that different. Even in that small a sample the difference is too big to be explained by statistical noise. OTOH we can't (and I didn't) say the Mohawk was 4 times as effective, because *some* of the difference could be explained by statistical noise. Systematic differences in the circumstances of Hurricane v Mohawk combats, not statistical noise per se, could explain the difference if it was the case, but I just don't see that in the book: it was basically similar operations v the same Japanese fighter units at the same time. Or, the Mowhawk pilots might have been better (though from the same AF, same time), etc... but it seems at that point we're grasping for straws to avoid the simplest conclusion from those results: the Hawk was at least as effective a fighter as the Hurricane. And that's what the larger sample in BoF also indicates v Bf109E (23:38 for Hawk, 74:151 for Hurricane, "Battle of France Then and Now"), but it's not AFAIK the general reputation of the Hawk.
Joe
Last edited: