[EM] Alternative counting method for RP

[Richard Moore]

Roy One wrote:
> This is fairly preliminary, but interesting.
> 
> It occurred to me that the table constructed for RP is, effectively,
> a Borda count. Obviously, Borda and RP are not the same, and the
> example on
> http://www.fortunecity.com/meltingpot/harrow/124/path/intro.html
> yields a different ranking if interpreted as a Borda count, rather
> than for RP.
> 
> But it also occurred to me that RP's method of resolving Smith sets
> was to evaluate the losers first. That may or may not equate to a
> form of elimination, but the thought appealed to me, so I gave it a
> shot on the data there.
> 
> Method: Compute the sum of each row. Eliminate the row and column
> corresponding to the lowest sum, repeat with remaining figures.
> 
> Result: same as RP: E was eliminated, then D, then C, then B, leaving
> A. (i.e., ranking is A>B>C>D>E)
> 
> This could fairly be called RP-E, I think, and it doesn't determine a
> winner before eliminating anyone.

Wouldn't it be more fair to call it Borda-E? Summing the 
rows in the pairwise matrix is Borda.

Richard