[EM] Apologies and corrections re: Borda and IRV

Paul Kislanko kislanko at airmail.net
Mon Jan 26 12:44:48 PST 2004

Mea culpa for sloppy thinking/writing.

The other day I suggested (quite incorrectly) that IRV was equivalent to Borda, when the only conclusion I should've drawn from a specific analysis was quite different than that statement. I can only blame laziness on a slow Saturday afternoon for that error.

It turns out that the algorithm I was analyzing only had a superficial resemblance to IRV, and I had labeled it IRV without checking to see what it really was. I still don't know which it was but will not compound my error by trying to match it to a set of abbreviations. What the algorithm I was looking at had in common with IRV is that of "transferable votes", and the only conclusion that can be drawn from the analysis is that the process of transferring votes upward in a ranking is the same as increasing the Borda Count for remaining candidates (by moving eliminated candidates to the end of the list). 

So, my original comment was appropriate to the process used in IRV iterations being the same as the process used to assign Borda Counts at any particular stage in an iterative method, and my observation should've been that iteration in an STV process is identical to that used to identify Borda Counts at a given stage. 

The first counterexample was unconvincing, since there was a majority winner from the beginning and therefore no iterations in IRV. The second counterexample pointed out the mistake in my original assertion - IRV eliminated the alternative in the first round that would have been relevant in my analysis of round 2. So whatever I was looking at was not "IRV."

Thanks very much to the ones who helped me identify the error. 

Paul Kislanko

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20040126/e367d3e5/attachment-0002.htm>

More information about the Election-Methods mailing list