Tuesday, July 28, 2015

What Batter Had The Best September Since 1914?

I used the Baseball Reference Play Index to call up all the cases of 90+ PAs that had an OPS+ (in this case called sOPS+) of at least 250 for September (actually, Sept/Oct). As far as I can tell, that is OPS adjusted for the league average and park effects (UPDATE 8-3-2015: BR informs me that sOPS+ is not park adjusted-I tried to do so myself and the new rankings are at the end of the post-details tomorrow if I have time).

But after compiling that list, I tried to make an adjustment for how close the pennant race was (or wild card race). That is discussed below.

Here is the list:


Rk
Player
Year
sOPS+

Rk
Player
Year
sOPS+
1
1942
348

25
1981
262
2
2001
341

26
1932
261
3
1940
323

27
1947
261
4
1941
297

28
1933
260
5
1967
294

29
1969
260
6
1925
294

30
1949
260
7
1992
291

31
2000
259
8
1920
283

32
1938
259
9
1946
283

33
1923
257
10
2000
283

34
1959
257
11
1951
282

35
1935
257
12
1944
275

36
1966
256
13
1928
274

37
1955
255
14
2004
272

38
1968
254
15
1927
272

39
1998
254
16
1949
269

40
1923
253
17
1933
269

41
1995
252
18
1929
266

42
1972
252
19
1946
266

43
2011
252
20
1939
265

44
1924
251
21
1944
264

45
2005
251
22
1965
264

46
2002
250
23
1954
263

47
1967
250
24
1917
262





 
So Ted Williams has the best at 348. But were the Red Sox in a tight pennant race in 1942? Not exactly. They came in 2nd, 9 games behind the Yankees. They entered September 8 games out and were never closer than 7.5 games (and that was only for one day).

So I decided to look at only guys who were on teams that entered Sept. no more than 5 games out (or ahead), were no more than 5 games out (or ahead) on Sept. 15 and finished no more than 5 games out (or ahead). That actually left Bobby Thomson of 1951 out, so then I also included any team that was no more than 5 games out (or ahead) with 5 games left.

To adjust sOPS+, I divided by the average number of games out (or ahead) per each day in Sept. and/or Oct. a team played a game. So if a team played 25 games and they were exactly 1 game out each time, they got a "closeness" score of 25/25 = 1 (a game ahead was counted the same as a game behind).

I used the average because if I am dividing, then a guy on a team that played 30 games and was 1 game out each time would have a score of 30 and it seems unfair to divide his sOPS+ by a bigger number than the guy mentioned above.

I did not take playing time into account. Maybe I should, but I am not sure what the best way would be to do so. Also, for post 1994 years, if the wild card was in play, I used games ahead or behind for that (but it got tricky since sometimes a team would be closer to the wild card leader one day but a couple of days later they were closer to the first place team in their division-I always used the lower number).


Rk
Player
Year
sOPS+
Closeness
sOPS+/Close
1
1967
294
0.426
690.3
2
1967
250
0.367
681.8
3
2004
272
0.536
507.7
4
1940
323
1.362
237.1
5
2001
341
1.519
224.6
6
2000
259
1.328
195.1
7
1920
283
1.593
177.7
8
1949
260
1.466
177.4
9
1966
256
1.682
152.2
10
2002
250
1.750
142.9
11
1959
257
1.804
142.4
12
1951
282
3.911
72.1

Of course, Yaz and Killebrew went up against each other in 1967. Yaz played for the pennant winning Red Sox and Killebrew played for the 2nd place Twins. I gave alot of details on that season and how well these two guys did in a recent post called A Case For The 1967 American League Being The Best Season Ever 

Park adjusted ranking


Harmon Killebrew 635.24
Carl Yastrzemski 623.73
Barry Bonds 506.05
Barry Bonds 241.56
Hank Greenberg 207.79
Jason Giambi 202.62
Babe Ruth 175.94
Stan Musial 167.88
Ron Fairly 167.07
Barry Bonds 150.69
Willie Mays 148.23
Bobby Thomson 71.65

I divided sOPS+ by the park factor and multiplied by 100. I used a three year average for park factors. If a team changed it's park, then I used only 2 years, as in the case of Mays. The park factor was adjusted if a player did not play exactly half of his Sept. games at home. For some guys, it was not close to an even split. Greenberg played 22 games at home and 5 on the road. Then I divided by the closeness factor.

My closeness factor might be too linear. If one guy had 1 and another 2, that says that the first guy faced pressure that was twice as strong as the second guy. But those two cases are, in a way, pretty close. If I use the square root of the closeness factor, so that pressure increases but at a diminishing rate, here are the leaders:

Carl Yastrzemski    407.06
Harmon Killebrew    384.65
Barry Bonds    370.39
Barry Bonds    297.67
Hank Greenberg    242.51
Jason Giambi    233.46
Babe Ruth    222.03
Ron Fairly    216.66
Stan Musial    203.23
Barry Bonds    199.35
Willie Mays    199.12
Bobby Thomson    141.69

Now, Yastrzemski is a bit ahead of Harmon Killebrew. Either way, those two guys are 1-2.

No comments: