[EM] Better/Worse strategy with 0-info

[MIKE OSSIPOFF]

Say, for all you know, all the candidates are equally likely to win.

And say you're using the Better/Worse strategy, which doesn't use your 
utility estimates.

If your preferences are G>D>R, then vote for G.  No one better than G can 
win, but someone worse than G can win. Better/Worse says to vote for G.

If If Better/Worse says to vote for a candidate only if it's more likely 
that someone worser than him will win, rather than someone better, then vote 
only for G. Don't vote for D.

Of course one could say to flip a coin if it's equally likely that someone 
better or someone worse will win. In that case it's 50% whether or not you 
vote for D.

There now, Russ, I've shown you how to vote with Better/Worse when no 
candidate is more likely to win than any other candidate. That "formula" is 
"effective" in the sense that it maximizes your expectation, given the 
information that we've assumed that you have.

To use the Above-Mean strategy, which uses utility estimates, add up your 
utility ratings for the 3 candidates. Divide the result by 3. Vote for every 
candidate whose utility rating is greater than the result of that division. 
I hope that I've explained that slowly and carefully enough for you, Russ, 
because you haven't gotten it so far that there are strategies ("formulas") 
that are effective in maximizing expectation when you don't have frontrunner 
information or winnability information.

Mike Ossipoff

_________________________________________________________________
Is your PC infected? Get a FREE online computer virus scan from McAfee® 
Security. http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963