Calculating the Smith set?

[Mike Ossipoff]

Yes, that algorithm is good, but you could reduce the number of pairwise
comparisons needed by adding a rule to admit an alternative to the Smith
set if it isn't beaten by more alternatives than some member of the Smith
set. 

So an alternative gets into the Smith set either by beating or tying a
member of the Smith set, or by not being beaten by more alternatives than
some member of the Smith set. The use of that 2nd way is optional, of course.

And by the way, I've just realized that Bruce's ordering of the alternatives
accoring to how many they beat, instead of how many they're beaten by, 
works ok too. Either way. I wrote it in terms of how many they're beaten
by.


Mike






--