[EM] 02/01/02 - IRVing is superior in math, Steve Barney:

[Forest Simmons]

On Fri, 1 Feb 2002, Donald Davison wrote in part:

> 
> You are also correct to say that both IRVing and Top Two Runoff are
> mathematically identical for a three candidate election.
> 
> But IRVing is better in that it only needs one election.
> 

Top Two Runoff has an instant version also.  Let's call it ITTR for
Instant Top Two Runoff.

ITTR uses exactly the same ballots that IRV uses, but is easier for the
voter to fill out because the optimal strategy is easier.  In particular,
after the first choice, optimal strategy is to list (i.e. rank) the
remaining candidates in sincere preference order. 

ITTR is summable in data structures of size proportional to the square of
the number of candidates.  IRV is only exponentially summable.

Several reasons have been advanced recently on this EM list supporting the
thesis that IRV is more manipulable than ITTR when there are more than
three candidates.

No reasons, arguments, or examples to the contrary have surfaced so far.

Forest