As Demorep said, IRV fails to use most of the information in the ballots. When working on an important project or job, would you take irrevocable action based on only a small fraction of the available information? That's what IRV does when it eliminates candidates based only on looking at one rank position. Pairwise-count methods look at every voter's every pairwise preference. Approval, whether voting is strategic or sincere, elects a candidate who is, overall, rated high by voters. Both of those 2 methods do better than just looking at one rank position. Having said something good about pairwise-count methods, I must add the disclaimer that, under the conditions existing in actual elections, it makes all the difference how circular ties are solved. Most pairwise-count methods, like most rank methods in general, don't even come close to Approval's merit, in regards to the standards that are important to most of us. For example, the methods that measure defeats by margin-of-defeat do quite poorly by the lesser-of-2-evils and majority-rule standards. Simulations have shown the Margins methods doing as well as the defeat-support methods even when there's much truncation. But those simulations add up all the utility, including that of truncators. If many voters are randomly chosen to truncate, then there'll by many who fail to support a needed compromise. Their utility loss, and the corresponding utility gain by people whose majority-defeated candidate wins as a result, tends to obscure the difference between Margins & defeat-support. In fact the randomness of the truncation itself fails to bring out the differences in benefit for voters who need a compromise and support that compromise. For those reasons the only reasonable way to compare Margins & defeat-support is via examples such as the ones that I posted some time ago, for truncation and for order-reversal. Mike Ossipoff _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp
