A very fast Stan Musial
I have wasted the entire day, and I mean the ENTIRE day, doing a study that didn’t work out, but I thought I would tell you about it for whatever the information is worth.
Occasionally, once or twice a year, I wake up in the morning with a fully-formed idea for a study which was not there the previous night, and I don’t know where it came from, but. . . .there it is. This morning was one of those. My idea was, growing out of the speed assessment by Game Lines that I did yesterday, to study the potential impact of speed on a player of a known quality. I chose Stan Musial, because I love Stan Musial and he seems special to me, but also because his skills are close enough to common that I thought it would not be too difficult to do what I was doing. Musial and George Brett and Hank Aaron have "common" skills, except at a higher level, as opposed to Mantle and Ted Williams and Joe Morgan, who have pretty unusual skills that would be difficult to replicate.
So my idea was, taking the 277,000 and some Game Lines in my file, to create 10 Stan Musials. In yesterday’s study I used 5 levels of speed, but that was a simplification; 10 is more natural. 10 Speed is the fastest players, Willie Davis and Willie Wilson and Byron Buxton; 1 is Ernie Lombardi and Steve Balboni and the aging Albert Pujols. I thought that I could sort the data so that I could create a "player" who had stats in the exact same proportion as Stan Musial, but out of game lines by "10" runners, and a player of the same exact stats, but created out of players who were 9 runners, 8, 7, 6, etc. I thought that I knew how to do that in a time-efficient manner.
I didn’t.
My idea of how to do that didn’t work, so I tried a second approach, and that didn’t work, either, and I tried a third approach, and that eventually worked, but it took me like 6-8 hours to create just the fastest Stan Musial clone. And then it doesn’t tell us anything about how speed influences runs scored by players because, in order to make it work, I had to scuttle the "control" on batting order position, so about 47% of the players represented in the group were leadoff men, so of course he scored more runs than expected but also drove in less. Since I know that is mostly a batting order position effect, there is no real information resulting from the study.
I did succeed in creating a group of 5,587 games which have the following numbers:
G
|
AB
|
R
|
H
|
2B
|
3B
|
HR
|
RBI
|
BB
|
IBB
|
SO
|
5587
|
20258
|
4004
|
6702
|
1339
|
327
|
877
|
2999
|
2952
|
98
|
1285
|
HBP
|
SH
|
SF
|
XI
|
ROE
|
GDP
|
SB
|
CS
|
AVG
|
OBP
|
SLG
|
98
|
67
|
98
|
0
|
306
|
449
|
144
|
131
|
.331
|
.417
|
.559
|
Numbers which are in the exact same proportion as Stan Musial’s career numbers, except for Runs Scored and RBI. This is kind of fun, I guess, because it would allow me to create an essentially infinite number of Stan Musial seasons. That seems like fun, but, you know. …I had work to do. It was a wasted day. There are a lot of those in the research business.