@Hitman: I like your approach, but I think the numbers are too high. I would tune the parameters that (assuming an equal weather distribution) you get a failure rate of approx. 20-25% up to June '40 and then less until mid '42. I think the failure rate dropped considerably after the invasion of Norway. I would also avoid a failure rate of 100%.
Maybe let's try to first agree upon the weather-independent probability on which the weather part is added?
|