Bayesian analysis of combat odds - or what the hell happened?

[Taillebois]

I've got the games (well six of them), but I still don't get combat outcome.

Irrespective of the thousands of calculations, if you can calculate them, why can't you have a prior probability?

[Ebbingford]

Which General would know before a battle anyway?
All you can do is have your troops rested, supplied and in the best position. If you have done your scouting you should know the basic make up of the opposing force. You should be able to gauge from all this whether offering battle is worth it, if not retreat.
I for one don't want to know what the outcome is probably going to be before the event. I'd rather have a battle where I can weigh up all the pros and cons as mentioned above, you have then got the unexpected to factor in and FoW.
Anything can happen, and as we learn from history it often did. I like the present system.

[Pocus]

I've got the games (well six of them), but I still don't get combat outcome.

Irrespective of the thousands of calculations, if you can calculate them, why can't you have a prior probability?

Because a combat is made of thousands of micro events, involving same number of random dices. Calculating a probability from that would mean the entire battle process is 'simulated' before giving you the unique number you seek. An impossibility it would be for such a small team, and something major even for a big one. A combat can involves hundreds of sub-units, each taking part in perhaps 15+ micro-events, like being committed to battle, passing morale check, passing initiative check, hitting the enemy (which one? again, dice roll), etc.

[Taillebois]

Napoleon thought he had nine chances in ten at Waterloo didn't he?

Anyway, thanks for the replies. I should stop dreaming things up at midnight after drinking the best part of a bottle of port (non vintage, but nice all the same).