Bang for their buck
For a while now, I’ve been working on a series of studies to evaluate the performance of baseball executives—primarily general managers. The idea of this particular study is “payroll efficiency,” which is just like it sounds.
To begin, I came up with a statistic called Pay+, which compares a team’s payroll to the major league average. The 2011 Rays had a payroll of $41 million, and the average team was at just under $93 million. So the Rays’ payroll was 44 percent of the league average; therefore, their “Pay+” was 44. The Red Sox’ payroll was roughly $162 million—74 percent above the league average—so their Pay+ is 174. A score of 100, of course, represents a league-average payroll.
Next we figure a team’s expected wins. So, sticking with the ’11 Rays … no team actually wins just 44 percent of the league average; that’d be a 36-126 record. I came up with this formula, which works really well:
(Pay+ minus 100), divided by four
Plus decisions divided by two
Which gives us “expected wins.” Do the math for the Rays ((44-100)/4)+(162/2) and you get 67 expected wins. The Rays actually won 91 games, so they’re +24 wins, tied with Arizona as the most efficient team in baseball.
So that’s the idea. Can the formula be improved? Absolutely. But it seems to work pretty well.
I have complete payroll data for each season of the free agent era, and I’ve identified general managers for each team in that era, so it’s easy enough to figure out who the most (and least) “efficient” GMs have been. Following are some of the extremes, by way of illustration.
Most wins above expectation in a season
A lot of the same GMs show up multiple times near the top of the list—Billy Beane, Andrew Friedman, Hank Peters. The 1993-94 Expos had different GMs, but both teams are in the top 10:
Rank Year Team General Manager Pay+ ExpW ActW Wins+ 1 2001 A's Billy Beane 53 69 102 33 2 2002 A's Billy Beane 59 71 103 32 3 2001 Mariners Pat Gillick 119 86 116 30 4 2008 Rays Andrew Freidman 49 68 97 29 5 1994 Expos Kevin Malone 57 46 74 28 6 1979 Orioles Hank Peters 89 77 102 25 6 1988 A's Sandy Alderson 94 79 104 25 6 1993 Expos Dan Duquette 54 69 94 25 7 1980 Orioles Hank Peters 81 76 100 24 7 2002 Twins Terry Ryan 60 70 94 24 7 2010 Padres Jed Hoyer 42 66 90 24 7 2011 Diamondbacks Kevin Towers 58 70 94 24 7 2011 Rays Andrew Freidman 44 67 91 24
Fewest wins above expectation in a season
The amazing thing about the bottom of this list is how frequently the Mets show up. And each Mets team has a different GM: Frank Cashen (1983), Al Harazin (1993), Steve Phillips (2003), Jim Duquette (2004), and Omar Minaya (2009). All of them are more than 20 wins below expectation. Here are the bottom 10:
Rank Year Team General Manager Pay+ ExpW ActW Wins+ 998 2003 Mets Steve Phillips 168 97 66 -31 998 2004 Diamondbacks Joe Garagiola Jr. 103 82 51 -31 998 2003 Tigers Dave Dombrowski 71 74 43 -31 993 2009 Mets Omar Minaya 168 98 70 -28 993 1993 Mets Al Harazin 124 87 59 -28 993 2008 Mariners Bill Bavasi 131 89 61 -28 993 1983 Mets Frank Cashen 158 96 68 -28 992 1992 Dodgers Fred Claire 135 90 63 -27 992 1988 Braves Bobby Cox 102 81 54 -27 991 1985 Braves John Mullen 140 91 66 -25
Most wins above expectation in a career
Beane totally dominates this category. However, over the past five years, he’s been a much more modest +14 (so, just three wins above expectation per year). Jim Campbell’s numbers only cover the last eight years of his long career. I was particularly surprised to see longtime Twins owner (and de facto GM) Calvin Griffith score so well:
Rank General Manager Years Pay+ ExpW ActW Wins+ 1 Billy Beane 14 67 1,017 1,206 189 2 Terry Ryan 13 62 919 1,023 104 3 John Schuerholz 26 128 2,251 2,348 97 4 Hank Peters 16 94 1,240 1,321 81 5 Andrew Friedman 6 53 416 495 79 6 Jim Campbell 8 69 559 637 78 7 Pat Gillick 27 110 2,204 2,276 72 8 Calvin Griffith 9 49 586 653 67 8 Joe Burke 6 95 449 516 67 10 Larry Beinfest 6 55 419 485 66
Fewest wins above expectation in a career
Okay, I need to explain something here: my win expectation formula works for every team except for the Cashman Yankees.
Those Yankee payrolls have been double the MLB average—well, actually, way more than double. Cashman’s average Pay+ is 224; that is, his average team is paid 124 percent more than the league average. The 2005 Yankees had a Pay+ of 284 (184 percent above average). It’s absurd, and it produces completely unrealistic results. Those 2005 Yankees are expected to win 127 games. This year’s Yankees are expected to win 111. That’s silly; no team realistically can go into the season expecting to win more than about 105 games.
To fix this problem, I imposed a cap on the system, the “Cashman cap,” which maxes out win expectation at 105. Even with the cap, Cashman scores at -100 for his career, but at least it brings him down to a somewhat realistic level. Anyway, I don’t think Cashman can truly be evaluated in the same way as other GMs, in terms of payroll efficiency.
With that one caveat, the bottom 10:
Rank General Manager Years Pay+ ExpW ActW Wins+ 146 Brian Cashman 13 224 1,355 1,255 -100 145 Bill Bavasi 11 111 889 793 -96 144 Buzzie Bavasi 8 136 693 604 -89 143 Bobby Cox 5 105 409 323 -86 142 Jim Hendry 9 132 801 733 -68 141 Steve Phillips 6 152 563 502 -61 140 Chuck LaMar 8 63 573 518 -55 139 Lou Gorman 12 114 1,014 963 -51 138 Harding Peterson 10 119 827 778 -49 137 Al Harazin 2 132 178 131 -47
Leaving aside the Cashman Yankees, here are the only other teams with at least 100 expected wins (in other words, the highest non-Cashman payrolls in the free agency era). Cashman’s teams own the top 11 spots.
Rank Year Team General Manager Pay+ ExpW ActW Wins+ 12 1978 Yankees Cedric Tallis 188 104 100 -4 13 2011 Phillies Ruben Amaro Jr. 186 103 102 -1 14 1977 Phillies Paul Owens 186 103 101 -2 15 1977 Yankees Gabe Paul 185 102 100 -2 15 1996 Yankees Bob Watson 183 102 92 -10 15 1997 Yankees Bob Watson 182 102 96 -6 15 1998 Orioles Pat Gillick 182 102 79 -23 15 2004 Red Sox Theo Epstein 184 102 98 -4 21 1998 Yankees Bob Watson 182 101 114 13 21 2010 Red Sox Theo Epstein 179 101 89 -12 23 1976 Red Sox Dick O'Connell 175 100 83 -17 23 1979 Phillies Paul Owens 174 100 84 -16 23 1982 Angels Buzzie Bavasi 176 100 93 -7 23 2011 Red Sox Theo Epstein 174 100 90 -10
On the opposite end, the teams with the fewest expected wins (non-strike seasons only) are the ones you’d expect: the 1999 and 2006-08 Marlins, the 2003 and ’07 Rays, the 1999-2000 Twins, and the ’98-’99 Expos. All these teams were expected to lose around 100 games.
Basically, the system doesn’t expect any team to win or lose many more than 100. I think that’s about right—teams can obviously be more extreme than that, but going into a given season, nobody is projected to be that more extreme than about 100-62 or 62-100.
Payroll efficiency is a blunt instrument. It doesn’t consider individual transactions, or which GM was responsible for acquiring which players, or a host of other factors. But the system does help identify which GMs have generally gotten the biggest bang for their buck. It passes the “smell test,” too: Billy Beane, Andrew Friedman, John Schuerholz—these are well-respected GMs, and they score very well by this method. On the flip side, Bill Bavasi, Jim Hendry, and Steve Phillips are generally regarded as poor GMs, and the system confirms this. This study also makes me curious to learn more about Hank Peters and Jim Campbell—for instance, should they be considered for the Hall of Fame?
In future articles, I’ll continue my attempt to systematically evaluate baseball executives.
Matthew
this is an interesting read and concept. I would be curious to see if we could somehow peel out the amateur draft effect on a team. Since Billy Beane became GM of the A’s after the 1997 season, can we attribute the success of the A’s teams in 2001 and 2002 to his reign as GM? Much praise went to Beane for his stable of pitchers and the 98 to 00 drafts show how he acquired these core pitchers, but Er Chavez, J Giambi and T Hudson were all in place before he took over as GM.
Name W L ERA
Zito, Barry 145 124 3.91
Mulder, Mark 103 60 4.18
Harden, Rich 59 38 3.76
Adkins, Jon 5 5 4.54
Lehr, Justin 9 6 5.34
In general, how could we seperate the actual moves of a GM from the original resources he was given upon arrival?
Mark
What I like most about this is seeing that big spending teams have existed for quite a while (i.e. 1977 Phillies 86% above average payroll). Would be interested to see a historical look at this.
Interesting work Matthew. It’s too bad you didn’t post it a couple days earlier. Maybe it would have saved Terry Ryan’s job.
A few thoughts – there’s nothing to account for inherited contracts. How many GMs come into a job and are hamstrung by bad contracts from the previous GM? One entry that jumped out at me was Dave Dombrowski’s 2003 season with the Tigers. It was his first year, and his predecessor, Randy Smith, threw around a lot of bad money (i.e. Bobby Higginson). I’m sure there are others on the negative list in the same situation.
I think the only solution would be to tie a player’s contract to the GM that signed him. That would certainly be a lot of hard research, though.
Also, I would say that Cashman’s results are no more unrealistic than the ability to spend 124% above the league average!! Your formula only reinforces how out of whack baseball economics is.
Mark and Jim, you’re both absolutely right: my system doesn’t account for which GMs acquired which players, so it’s going to be particularly unenlightening for GMs early in their tenures. Also, it’s not clear how much “credit” a GM could get for amateur acquisitions.
Anyway, with regard to Beane in particular, out of curiosity, I looked at the 2002 Moneyball A’s. The only guys on that team who weren’t acquired on Beane’s watch were Tim Hudson, Miguel Tejada, Eric Chavez, Ramon Hernandez, and Adam Piatt. It’s a small group of players, but of course three of those guys were among the team’s biggest stars.
I also looked at the 2001 A’s. Pat Gillick acquired most of the position players (excepting Edgar Martinez, Carlos Guillen, David Bell, and Dan Wilson) and the core of the bullpen. But the only starting pitcher Gillick acquired was Aaron Sele; the rest were Woody Woodward acquisitions.
Anyway, to truly evaluate a GM, we have to dig much, much deeper than simply payroll and wins. I view payroll efficiency as a quantifiable starting point, but only a starting point for discussion.
A follow-up to my previous comment:
I looked at all the players on the 2002 A’s who were were worth at least 0.5 rWAR. (For pitchers, I only considered their pitching, not their offensive WAR.) A total of 20 players qualify, and they accumulated a 47.6 rWAR. Of that total, Beane acquired 30.9 rWAR (65%), and Alderson acquired 16.7 rWAR (35%). So although Alderson only acquired 4 of the 20 key players, they were all really important players: Tim Hudson (6.6 rWAR), Miguel Tejada (5.2), Eric Chavez (3.6), and Ramon Hernandez (1.3).