Quote:
Originally Posted by Mannyw
I'm the original poster. I've followed this thread and given it a lot of thouoght. I looked up Sabermetrics in relations and found something very interesting called Speed Score. Below is the information copied from Wikipedia:
https://en.wikipedia.org/wiki/Speed_Score
So I do think it is possible to more accurately calculate player speed. I played APBA Baseball growing up and have messed with their Window version some too. They have somehow calculated player speed for years, seemingly very accurately. In their system a player is ranked 1-20, with a slow player being 1-6, average 7-14 and fast 15-20. Only the most elite runners reach 20, Rickey Henderson, Tim Raines come to mind.
Their speed ranking is entirely separate from a runners SB success. In speed a player speed is subtracted from outfielder's arm rating and then that sum is subtracted from a number representing ball depth in outfield.
In base stealing the runner has a 0-36 base stealing rating. Catcher have a -4 to +6 arm rating and pitchers have a 0 to +3 pitch-off move rating. So a +4 catcher and a +1 pitcher would have a sum of +5. This +5 is subtracted from the base stealing number, say 30 of a possible 36. So 30 minus 5 gives the runner 25 out of 36 chances to successfully steal the base.
Just so thoughts, it would take work but runner speed accuracy is possible, even for non modern era players. |
While I agree that Speed Score would be an interesting way to calculate speed for historical players, I think it ought to be noted that for the guys people are citing - Roberto Clemente and Willie Mays - that way of calculating their speed would likely give them slightly above average but certainly not elite speed ratings.
Well... I'm going to test that. Here are some speed scores by players who may or may not have been considered fast in the early 60s (using 1962):
Clemente: 7.82
Frank Robinson: 6.77
Mays: 7.88
Tommy Davis: 7.98
Hank Aaron: 7.74
Lou Brock: 9.58
Frank Howard: 5.6
Maury Wills: 10.66
I find it interesting that it the algorithm managed to suss out Lou Brock so well (Brock, then playing for the Cubs, was only 18/24 for steals) but jeez, if Wills is an 80/80 and Brock is about a 75/80, Clemente and Mays are pretty far off... maybe 55-60/80 for both? And that's probably being generous... Frank Howard is pretty close to a 20, maybe a 25, but there's more difference by speed score between Clemente and Wills than there is between Clemente and Howard.