Pairwise Ranking Bonus Points

[Petunia \'95]

Does anyone remember what the Pairwise Bonus points were figured out to be?
I thought it was either .005 .003 .001 or .003 .002 .001.
I thought it was figured out based on the teams which made the NCAAs 2 years ago.
Anyone remember?

[jy3]

my knowledge is that no one really knows for sure and no set figure has been given. one of the secret aspects of the selection process. i would suggest predicting seeding based on no bonus, and the two you have listed

[Greg Berge]

It is not published, which leads some to postulate that they figure out what result they wish to get, then reverse-engineer the weights to achieve it.

[Newman]

In theory, if it was one set-in-stone set of values, it could be calculated after a finite and probably small number of years of observing the selections. But there's no reason to believe that it's a fixed set of numbers from year to year. Since it's secret, it can be changed on a whim.

[Greg Berge]

There's no reason to believe they can even do the arithmetic correctly.  Trying to detect a pattern implies there's a pattern to detect.

[Jeff Hopkins '82]

Add to that the fact that they can always move a team within a band for attendance reasons.

It'll take several years before a pattern develops, and by then they'll change the rules again.