home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: rec.sport.baseball
- Path: sparky!uunet!boulder!ucsu!spot.Colorado.EDU!franjion
- From: franjion@spot.Colorado.EDU (John Franjione)
- Subject: Re: Cecil signs : -( :-( :-(
- Message-ID: <franjion.727660460@spot.Colorado.EDU>
- Sender: news@ucsu.Colorado.EDU (USENET News System)
- Nntp-Posting-Host: spot.colorado.edu
- Organization: University of Colorado, Boulder
- References: <2123@blue.cis.pitt.edu> <1993Jan19.125019@IASTATE.EDU><GRABINER.93Jan20140019 @kovalevskaia.harvard.edu><1993Jan21.025125.26861@athena.mit.edu> <GRABINER.93Jan21141053@germain.harvard.edu>
- Date: Thu, 21 Jan 1993 23:54:20 GMT
- Lines: 40
-
- grabiner@math.harvard.edu (David Grabiner) writes:
-
- >I don't believe this is true, though. The value of a player to your
- >team does not depend on standard deviations; it depends on the number of
- >games he helps your team win and lose. If you trade a shortstop who
- >slugs 50 points better than average and an average first baseman for an
- >average shortstop and a first baseman who slugs 50 points better than
- >average, your team will still score the same number of runs. Since you
- >break even on the trade, you must have traded players of the same value.
-
- I think I disagree with this. Yes, a player contributes the same
- amount offensively regardless of his position (and also regardless of
- the distribution of performances at that position). But his value, in
- terms of what you would be willing to pay in order to acquire that
- player, should depend on the distribution.
-
- [Note: the following example might not be very realistic]
-
- I'm going to put players' performances in terms of runs relative to an
- average player at that position. Say that zero (0) runs is an average
- performance for that position. Let's say that among firstbasemen, the
- "standard deviation" of the distribution of performances is 50 runs
- (over the course of a season). In contrast, suppose that the StD of the
- distribution for shortstops is only 10 runs. That is, more shortstops
- are closer to average than firstbasemen.
-
- Let's look at two teams
- A B
- ----------------------------
- 1b +50 0
- ss 0 +50
-
- In terms of contributing to wins, both teams are identical. However, I
- think that team B is "worth" more. One simple way to measure worth is
- to normalize runs by the StD at that position. Thus, team A is worth
- 1, but team B is worth 5. Intuitively, this makes sense to me too.
- Because the StD for shortstops is lower than for firstbasemen, it would
- be more difficult to replace a +50 shortstop than a +50 firstbaseman.
-
- John Franjione
-