Unless one is concerned about using Ran's bandwidth, why are folks bothered about where the discussion goes? Self-discipline and selective perception are not being unduly challenged here. Don't like the subject or the poster? Don't click on the thread or page down quickly. Simple. No?
But anyway, such systems are pretty much plug-n-chug. Built into all the usual stats software and such.
Thanks Brent. Are there easily modifiable canned models that would allow me to do this? Say I had 10 raters that would give me their individual rankings. Is there an existing program that would give me a composite "best fit" list of these? I'd like to play with, preferably if it is in Excel.
Lou,
I did extensive modeling using a program called S-Plus, which is now available as a web tool, but then was a desktop app. I could do things like cluster analyses of weekly football picks in a few minutes, with graphs, or with coding the ranking type statistics you are interested in.
The free version is R,
http://en.wikipedia.org/wiki/R_(programming_language), which has some kind of Excel plug-in. My experience with S-plus was that with reading the guides, you could get to the routine that would spit out a number or result that had the right structure. Not being a statistcian, being confident that result isn't being perverted by some bias you don't know to look for, that's the tough part.
Dave