This past offseason, we projected the salaries for 155 arbitration-eligible players who received one-year contracts. The results were significantly better than I expected. In my first article on the projections, I estimated that we would be within $700K for about half of players, but we actually were within $200K for half of players. Our projected salary was within 5% of actual salary for 29% of players, within 10% of actual salary for 55% of players, and within 20% of actual salary for 81% of players. In fact, there were only four players who received a salary that was more than $1MM away from their projected salary.
However, within these aggregate numbers are a mixture of many very accurate projections, and quite a few that were way off. Pitchers were either very easy to project or very hard to project. When pitchers matched up very well with historical comparables, they fit squarely into categories. However, some pitchers proved to be a new breed with weak sets of comparables. As a result, of the closest ten projections, nine were pitchers, but of the worst ten projections, eight were pitchers.
When there was more precedence for a player’s performance, projecting his salary was much easier. The reason that so many relief pitchers were among the best projections was that they have very defined roles, and they are paid according to their role. Closers, set-up men, middle relievers, and long relievers all tend to get similar salaries as other such relievers in their service class and role have received in the past. Our projections were within $25K of actual salaries for relievers such as Craig Breslow, Brad Ziegler, Daniel Bard, Bill Bray, Edward Mujica, and Burke Badenhop. Each of these guys had defined roles and matched up nicely with historical comparables in similar roles.
Salaries are also very predictable for players who miss all or most of the previous season. These players almost never get big raises, and almost no players ever get pay cuts—so these are often players who get the same salary as the previous season. So, this year it wasn’t surprising when Manny Parra and Dallas Braden were rewarded another go around at their 2011 salaries of $1.2MM and $3.35MM. Next year, it won’t be surprising when Joey Devine and Brian Wilson get repeats of their 2012 salaries (if they are tendered contracts) after coming back from Tommy John Surgery.
Defined back-up hitters’ salaries can be pretty predictable as well. As with relievers, players with roles that are comparable to several other players in recent history make for quick agreements between players and teams. Jeff Baker, Emmanuel Burriss, Wilson Valdez, and Chris Denorfia all had salaries within $35K of our projected estimates.
Not all projections were so easy. One subgroup of pitchers where we may have overestimated salaries is swingmen, or pitchers who were converted from reliever to starter, or vice versa, during the season. Andrew Miller only received $1.04MM despite a $1.6MM projection, and Jesse Litsch received $975K after we estimated $1.3MM. There may be room for improvement by correctly modeling pitchers like these going forward. Broken service time can really take a chunk out of a player's salary too, especially if it’s in an atypical way. Also somewhat of a swingman, Jerome Williams settled for $800K after being projected for $1.4MM.
Two of our four biggest misses came on pitchers who were eligible for arbitration for the first time, but were coming off large salaries they received as part of a Major League contract signed as amateur. These pitchers are Rick Porcello, for whom we overestimated his expected salary by $1.1MM, and David Price, for whom we overestimated his salary by $2.55MM.
Porcello was an interesting case because his numbers were pretty standard for a healthy, solid, but not elite, starting pitcher. Pitchers like those typically get salaries in the $3.0-3.5MM range, so Porcello’s salary wasn’t surprising. However, he already earned $1.536MM in 2011 as part of his original contract signed out of high school, so we projected him for $4.2MM. His 2011 salary seems to have been irrelevant in the discussion about his 2012 salary.
Price was coming off a $1.25MM salary in 2011, and with a 19-win season in 2010 and 224.1 IP in 2011, he seemed primed to get a nice raise. However, as I attempted to model the effect of his 2011 salary, I overshot. It seems like Price may have given in a little early in accepting a $4.35MM deal, though, because Tim Lincecum was the only pitcher in the previous five years before Price with a career ERA under 3.70 (Price’s was 3.38), at least 40 career wins (Price had 41), and over 200 IP in his platform season (Price had 224.1). Jered Weaver was given $4.265MM, the largest one-year deal for a starting pitcher his first time through arbitration in that timespan, and he had a career ERA that was 0.34 higher than Price had, while having fewer innings, though Weaver did have more wins (51) than Price. However, it seems reasonable to guess that Price should have landed closer to Weaver than to Lincecum. I will look for ways to better incorporate pre-arbitration salaries going forward.
The most surprising big miss was Melky Cabrera. We expected that he would receive a nice raise from $1.25MM to $4.4MM in 2012. That would have been a raise as high as all but 14 position players over the previous five years. However, Melky Cabrera and the Giants agreed on a $6MM salary for 2012. There were only six position players to get raises that large in the last five years. They were Jose Bautista in 2011, Josh Hamilton in 2011, Carlos Pena in 2008, Matt Holliday in 2008, Ryan Howard in 2009, and Rickie Weeks in 2011. Those players had anywhere from 29 to 54 home runs in their platform year; Cabrera only had 18. Only Rickie Weeks (a leadoff hitter) had fewer than 121 RBI. Cabrera had 87. Among the players who had raises larger than our estimated $3.15MM estimate, none of them had more than 25 home runs either. In this case, I think this one might just be a case of the Giants were out-bargained by the aces at ACES. I’m not sure that he would have gotten a raise anywhere near that large if the Giants had held out and taken Cabrera’s case to a hearing (however, the sides wouldn't necessarily have argued 'raise' for Cabrera and others with broken service time).
Overall, the first year of these projections went very well. However, the projections were not so perfect that there is not still some room for improvement. Going forward, we will make sure to take a better look at swingmen, and other pitchers who had multiple roles in their platform season. We will also see if there is some way to tell when a large salary before arbitration is going to affect a player’s salary when he is eligible for the first time. There also may be a way to find a class of hitters where projections are as cut and dry as they often are for relievers with defined roles, so we will look for this as well. As players are just starting to accumulate their statistics for the 2012 season, we are already preparing to evaluate what those statistics will mean for their bottom lines in 2013.