QUOTE(wKkaY @ Jul 15 2009, 01:37 PM)
If this is like most toy problems, it shouldn't be a 15-variable problem. Assuming that pigs within the same class have same probability of winning, then there's only one variable involved - the bias in favor of the strong pigs vs the weak pigs.
The problem can be simplified by classifying the pigs into {Strong,Weak} x {Odd,Even}. And the top three positions as {Even,Odd} x {Even,Odd} x {Even, Odd} - pick only the outcomes that result in an even sum.
Then the rest of the problem can be solved with a probability tree. Okay, that would be how i'd do it coz i'm stupid - but it will work!

That's right, except that he's weighted each strong pig with different probabilities, which makes it more of a pain instead
Edit: I forgot an added simplification step:
Suppose pigs 1, 3, 5 get the top 3 positions. We don't really care about the arrangement of the other 12 pigs thereafter, so we can compute the number of possible permutations with pigs 1, 3, 5 being in the top 3 to be 12! * 3! (3! for the various arrangements of pigs 1, 3, 5)
This post has been edited by bgeh: Jul 15 2009, 02:00 PM