On Thu, 2004-12-16 at 20:15 -0800, Frank Reichert wrote: > Good evening Lowell! > > I'm dealing with this issue here on the "Newsnet" side, and don't > know for sure if it will really make it's way across, but I hope > it will! > > Lowell C. Savage wrote to Steve Thompson... > > > I'm afraid I'll have to take issue with you on a couple of points. First, I > > suspect that you are making a lot more out of Cheney's remarks that they > > deserve. It was certainly a poor choice of words, but at the same time, > > fewer candidates makes it a lot easier to design, print and tally ballots > > and a lot easier to deal with security for candidates and set up debates and > > all kinds of other things. > > So? What's the problem for YOU, dealing with conflicts in shere > numbers in terms of ballots cast? I'm certainly not ready to > make a 'big deal' over it here in Idaho, however, it still is a > hell'a long shot, isn't it? > Imagine, over 12,000 votes cast, and suddenly there is an 'exact > tie' vote between Steve Elgar and Steve Anderson! Exactly, an > exact tie!? Hoexactly do probabilities of such a thing happen? > You count the odds of THAT happening! Talley the probable result > and send 'em on to me, I'm interested.
Give the expected ratings/ranking of each candidate and calculation of the likelihood of a tie is a simple process from there. The odds or probability of a tie happening between is dead even if the candidates are ranked equally. Without using any ranking or expectation, or any other external data, the odds or probability is dead even a tie will result form an even number of ballots. The larger the "sample size" the less likely a tie is to happen. If you have 12,000 people, and assume no predisposition (IOW each voter is randomly selecting between the two) than the chances of a tie are actually very high. That's because the odds of a voter choosing A are the same as a voter choosing B: 1 in 2. IIRC, to determine the probability of a tie between two people in a two-way election you start by assume a tie, then modify base don other factors. It's that predisposition combined with other factors that are needed to determine the probability of a tie. I believe that adding candidates above 2 increase the odds of a tie, particularly if the "additional" options are assumed to shrink the pie by drawing away "partial supporters". Essentially they decrease the sample size under that scenario. An in reality, IIRC, there are something like dozens to hundreds of ties each even numbered election year. Seems I hear about a tie in Idaho at least every other year. Usually it goes to a coin toss. -- Random Fortune of the moment: An English judge, growing weary of the barrister's long-winded summation, leaned over the bench and remarked, "I've heard your arguments, Sir Geoffrey, and I'm none the wiser!" Sir Geoffrey responded, "That may be, Milord, but at least you're better informed!" _______________________________________________ Libnw mailing list [EMAIL PROTECTED] List info and subscriber options: http://immosys.com/mailman/listinfo/libnw Archives: http://immosys.com/mailman//pipermail/libnw
