Expected Player Performance

A few things were apparent to me when I thought about how I could breakdown the ASPs -- 1) I invested heavily in a bullpen, 2) I invested relatively heavily in a fair, but not outstanding starting staff, and 3) in my quest to balance my lineup, I drafted a bunch of players with high platoon ratings. In this last strategy, because I took a lot of pitching early, I was left with trying to get a solid lineup despite most of the top tier players already having been taken. As such, the best players remainining had deficiencies, some of which included high platoon ratings. So though I meant to balance my lineup with the best remaining players, this strategy had an unbalanced effect, as my team finished in the bottom third in runs scored.

The best breakdown I thought would be to review the platoon ratings. But after a lot of thought, I concluded that it was much too hard to quantify the ill effects of the high ratings -- if the high ratings even had a negative effect on my team's output. Here are the stats (in terms of On-Base Plus Slugging, with each group arranged by PA's) that lead me to question the effect of platoon ratings:

                        MLB   SL   DIFF
High Ratings (5,4)
Clark       (L SA-5)    966   910  - 56
Drew        (L SA-5)    880   927  + 47
Hansen      (L PR-5)    985   689  -296

Fairly High Ratings (3,2)
ARod        (R SA-3)   1026   818  -206
Griffey     (L SA-3)    943   956  + 13
Hammonds    (R SA-3)    924   836  - 88
Boone       (R SA-2)    827   729  - 98
Fabregas    (L SA-2)    693   753  + 60

Low Ratings (1,0)
Posada      (B SA-0)    944   791  -153
Justice     (L SA-0)    961   773  -188
LCastillo   (B SA-0)    806   632  -174
Lowell      (R SA-0)    818   654  -156
Berkman     (B SA-0)    949   759  -190
Floyd       (L SA-1)    902   730  -172
Reese       (R SA-0)    705   498  -207

In the group of guys with the highest platoon ratings, 2 of the 3 had about the same OPS in the SL as they did in MLB -- the third player, Dave Hansen, was notably worse in the SL, but he did have relatively few plate appearances compared to the other two. In the middle group of 5, the 4 with the most plate appearances averaged about 95 points below what they did in MLB (the 5th, Jorge Fabregas, had very few PA's to go with his 60 point increase in the SL). In the group with the lowest platoon ratings, all of the 7 players had a significantly worse OPS, ranging from 153 points below to 207 points below -- if you choose to ignore Pokey Reese and his relatively few plate appearances, the poor end of the range only decreases to 190 and the mean is lowered to the still-very-high 170 point differential.

A better topic I thought after reviewing these statistics would be to decide what can be expected from a player in the SL relative to his MLB numbers. Facing baseball's top pitchers, instead of the occasional Micah Bowie's of MLB, few players will improve upon or even match their MLB totals in the SL (only 2 of the 15 ASPs matched or increased their MLB totals). So when going into a draft, analyzing players, constructing an SL team, and playing games, I think it would be great for a GM and a Manager to know what the expected decrease in performance is for an SL player.

From the outset, I'll admit that I doubt I'll be able to fully and accurately quantify these numbers, but this is one of the most important parts of player analysis come draft time, so it's extraordinarily relevant to the Draft Notes and worth addressing, no matter how feebly I address it.

This topic was talked about after the Greg-o-rama, as Ken marvelled at Andruw Jones's SL 20001 numbers and Greg responded with compliants about Travis Fryman's performance for the Splinters. Here are there final numbers, plus some of Fryman's numbers from previous seasons:

On-Base Plus Slugging   MLB   SL   DIFF
Andruw - SL 2001        907   848  - 59
Travis - SL 2001        908   681  -207

Travis - SL 1999        844   836  -  8 (just 144 SL PA's)
Travis - SL 1998        766   547  -219
Travis - SL 1997        766   627  -139
Travis - SL 1996        756   521  -235

Andruw was pretty close to his MLB performance at the end of the year, but he did not exceed his OPS, and Fryman, as he has done several times in his SL career, performed far below what he did in the SL. Based on these 2001 numbers and Ken and Greg's reaction to their performances, I estimate that the expected OPS of a player is somewhere between Andruw's 59 points below his MLB, and somewhere above Travis's negative 207.

I'll break it down a little further, estimating that Travis was far below expected and Andruw slightly above expected. From my OPS comparison of the ASPs, I would estimate that all of the players from Posada on down, players who finished below -153, had disappoiting seasons. Of the other guys with negative SL OPS, (Clark -56, Hammonds -88, Boone -98), I would agree with the Andruw-based estimate that Clark slightly exceeded my expectations. Hammonds I thought did well, but he didn't grossly exceed anything. Boone, because of his relatively few PA's, I don't have much of a feeling about. So the expected number is definitely below Andruw and his -59, probably below Hammonds's -88, maybe below Boone's -98, definitely higher than Posada's -153, and even probably higher than what Travis did in SL '97, when he finished -139.

I could look at everyone in our history, but I have a self-imposed deadline for the Draft Notes. Instead, I'll look at a group of players who span our history and try to narrow the number down further from there. The group I chose are the players with the top individual OPS seasons in our history (list currently available along with debug code at http://www16.brinkster.com/kklein/sli/ptg.asp, and later available on the SL Player Register). The numbers for this group are:

Player           Year  MLB   SL  DIFF
McGwire, Mark    1997 1197 1257  + 60
Delgado, Carlos  2001 1134 1135  +  1
McGwire, Mark    1999 1222 1104  -118
Walker, Larry    1998 1177 1099  - 78
McGwire, Mark    2000 1121 1075  - 46
Bonds, Barry     1997 1076 1069  -  7
Alfonzo, Edgardo 2001  967 1068  +101
Jones, Chipper   2000 1074 1061  - 13
Sheffield, Gary  1997 1089 1039  - 50
Bonds, Barry     2001 1128 1028  -100
Helton, Todd     2001 1161 1027  -134
Martinez, Edgar  1997 1059 1004  - 55
Griffey, Ken     1999  976  995  + 19
Sosa, Sammy      2001 1040  994  - 46
Walker, Larry    1996  988  992  +  4
Ramirez, Manny   2000 1105  990  -115
Thomas, Frank    1997 1085  985  -100
Piazza, Mike     1998 1069  981  - 88
Belle, Albert    1997 1033  974  - 59
Bagwell, Jeff    2000 1045  971  - 74
Walker, Larry    2000 1168  970  -198
Giambi, Jason    2001 1123  960  -163
Martinez, Edgar  2000 1001  956  - 45
Griffey, Ken     2001  943  956  + 13
Lankford, Ray    1998  996  953  - 43

Narrowing the list, the players in this list who fall in the Andruw to Posada range are (arranged by difference):

Player           Year  MLB   SL  DIFF
Belle, Albert    1997 1033  974  - 59
Bagwell, Jeff    2000 1045  971  - 74
Walker, Larry    1998 1177 1099  - 78
Piazza, Mike     1998 1069  981  - 88
Bonds, Barry     2001 1128 1028  -100
Thomas, Frank    1997 1085  985  -100
Ramirez, Manny   2000 1105  990  -115
McGwire, Mark    1999 1222 1104  -118
Helton, Todd     2001 1161 1027  -134

Do these numbers tell us anything about expected performance? Just about all of these guys had great SL's, because if they didn't, they wouldn't be included in the greater list of best OPS in league history. Is it possible that all of them had the year's their managers expected? Maybe, but McGwire in '99, for one, is a guy who I know was expected to do a lot based on what he did in MLB, and while he did do very well in the SL, his numbers were not up to expectations.

So is -118 the limit of disappointment? The number below it is Helton '01 and his -134. I think Graham was generally happy with most of what Helton did, but I'm sure he expected some more of his 108 extra-base hits would be home runs. In that sense, his OPS was reduced and because of that Graham was disappointed in Helton, in spite of a 1027 OPS.

The three numbers ahead of McGwire's "disappointing" -118 are Manny's -115 in 2000, and a -100 from Bonds '01 and Thomas '97. I can't pretend to remember what Jim Frank felt about Thomas and his performance in '97, but to that point it was the 6th highest OPS ever. I'm not even certain how Steve felt about Bonds's '01 performance, but I would guess he had few complaints -- Bonds was the 3rd position player taken, and he had the 3rd highest OPS in the league. Whether or not Manny met expectations in 2000 is even more of a mystery. The stats I think Graham didn't like were his moderately low batting average (.255), his high HBP numbers (22), and his relatively few games (147). Despite being the 3rd player taken, Manny finished just 9th in MVP voting -- but in his defense, he had the 3rd best OPS in the league, trailing McGwire and Chipper, the two players taken ahead of him, and he did have tremendous power, evidenced by his 92 extra-base hits in 542 at bats. If I had to guess, I'd say Graham was much more disappointed in his 2000 team than he was with Manny's -115, especially Brian Giles's -239 differential and Jose Lima's blow up.

So it may be a simple study looking at very few cases, but -- to make sure I meet my deadline -- I'll conclude that the expected performance in the SL is close to a -100 differential in OPS versus MLB, probably up to -115, but likely not to exceed that number by much. I welcome someone to do a more definitive study.

Back to Draft Notes