The Loveable Losers wrote:J35J wrote:markj11 wrote:How come you do not consider W for RP?
Thanks for the spread sheet.
I did in my last years spreadsheet but figured that was probably the biggest crap shoot to try to predict so I figured I would move guys up or down manually if a player normally gets alot of wins or if they normally don't. By all means add that into the formula if you want but I figured I would go without this year.
Here's my article on projecting wins. The methodology used in there is by no means professional but I gave it as thorough a handling as my non-professional stats background (read: I don't have the first clue how to do a regression analysis) allowed. It seems fairly accurate over a large data set (say all of the closers in a given season) but there's such a chance for wild fluctuations in the small data sets of closers that it can seem that there's no rhyme or reason. Wins do seem to follow predictable patterns when you look at closers as a group though...it's just that an individual closer may or may not end up anywhere near his projected wins.
That is superb. Interesting to revisit some of the old 'Middle Relief Strategy' threads to see what one could do with that kind of information. It seems like an elite MR in 5x5 could give you the ratios, without harming the team as much in wins as might have been believed.