Arbitration Breakdown: Craig Kimbrel

Over the next few months, I will be discussing some of the higher profile upcoming arbitration cases. I will rely partly on my arbitration model developed exclusively for MLB Trade Rumors, but will also break out some interesting comparables and determine where the model might be wrong.

In just over three years on the baseball diamond as a member of the Atlanta Braves, Craig Kimbrel has accomplished a wide variety of feats. He has led the National League in saves for three years in a row, with 139 saves in his young career. Kimbrel struck out a fantastic 381 hitters, which amounts to 43 percent of all hitters faced (more than twice the league average). He has also kept runs off the board like no one else, allowing a microscopic 1.39 ERA in his 227 1/3 innings pitched. Now, we can add one more trophy to his case here at MLB Trade Rumors: he is the namesake of the Kimbrel Rule, as well as the first player whose arbitration salary projection has ever been affected by it. Why do we need this Kimbrel Rule? Quite simply, Kimbrel broke my arbitration model.

The result that the model spit out was so outlandishly high that no panel would ever have awarded it to him. The model works by considering how panels and settling parties (teams and agents) have previously interpreted different statistics and factors into their decision, and provides a salary estimate based on these. This is often useful, for example, because we can see how many saves a closer would need to have to offset a higher ERA than a previous comparable. Similarly, for a hitter, we can see how much an MVP adds to salary and how many home runs a hitter would need to belt to make up the difference. However, there is no category in which Kimbrel failed to dominate the preceding closers who have reached their first year of arbitration.

Consider the following lists of maximum first-year salaries for closers in arbitration over the previous seven years:

Jonathan Papelbon ($6.25MM):

• Platform year—41 SV, 2.34 ERA, 69 1/3 IP, 77 SO
• Previous years—72 SV, 1.63 ERA, 160 2/3 IP, 193 SO

Bobby Jenks ($5.60MM):

• Platform year—30 SV, 2.63 ERA, 61 2/3 IP, 38 SO
• Previous years—87 SV, 3.26 ERA, 174 IP, 186 SO

John Axford ($5.00MM):

• Platform year—35 SV, 4.68 ERA, 69 1/3 IP, 93 SO
• Previous years—71 SV, 2.26 ERA, 139 1/3 IP, 171 SO

Brian Wilson ($4.46MM):

• Platform year: 38 SV, 2.74 ERA, 72 1/3 IP, 83 SO
• Previous years: 48 SV, 4.34 ERA, 116 IP, 108 SO

Now consider Kimbrel’s primary arbitration-relevant statistics:

Craig Kimbrel (unknown):

• Platform year—50 SV, 1.21 ERA, 67 IP, 98 SO
• Previous years—89 SV, 1.46 ERA, 160.1 IP, 283 SO

Now consider the records for each of these statistics by all closers (defined by having 15 platform-year saves or 50 pre-platform saves) put together before Kimbrel, which we will call Mutant Super-Closer:

Mutant Super-Closer:

• Platform year—41 SV, 1.71 ERA, 78 1/3 IP, 93 SO
• Previous years—87 SV, 1.63 ERA, 252 IP, 269 SO

Other than innings, Kimbrel actually beat the Mutant Super-Closer too. He beat every previous closer in every previous category.

Looking at the foursome of potential comparables above, it is clear that Papelbon is the only closer who even came close to achieving what Kimbrel ultimately has achieved, and he fell far short of Kimbrel’s accomplishments in every category.

Kimbrel had 50 saves in 2013, his platform season. The most any other pitcher has had going into his first year of arbitration in recent memory (which includes the last seven years, the data I have available) was Jonathan Papelbon, who had 41. No one had more career saves in recent memory than Bobby Jenks who had 117 going into arbitration for the first time, but Kimbrel has 139.

Kimbrel has a 1.39 career ERA, which is well ahead of any other closer reaching arbitration. The closest was Papelbon who had a 1.84 ERA going into his first year of arbitration in 2009, but his 2.34 platform year ERA fell far short of Kimbrel’s 1.21 mark for 2013.

Kimbrel also has all other closers beat on strikeouts too, tallying 381. The next highest for a closer going into his first year of arbitration was 362 strikeouts by Carlos Marmol leading into 2010, but Marmol’s 23 career saves at this point fall well short of Kimbrel’s 139.

In the end, there just weren’t any comparable pitchers for Kimbrel. There was no category where he fell short of another closer’s mark, and the accolades piled on to escalate his salary projection. This is why we invented “The Kimbrel Rule,” which is defined by limiting the maximum salary projection possible to exceed the previous record for his player type to $1MM (and similarly, the maximum raise for a non-first time eligible player is $1MM greater than the previous record raise as well). In this case, no one has ever earned more in their first year of arbitration as a closer than Jonathan Papelbon, when he earned $6.25MM in 2009, so Kimbrel is projected for $7.25MM despite the model itself predicting a salary well in excess of this amount.

This number was selected as a rough approximation of what teams, agents, and arbitration panels have historically decided on. In general, the model I use does a good job of approximating the end-result of their decision processes, but when faced with no historical precedent, there is often a settlement that avoids beating the previous record by too much. Therefore, we have made this specific rule a part of the model. It is not the first time that we have let the model “have eyes.” For players who do not play in a given season, we have observed that they so frequently get rewarded with their exact previous salary that this is now an explicit rule in the model. No player gets projected for a decrease in salary anymore because those that would have often received this floor anyway—their previous salary. Now, we have a rule for the ceiling for players.

We are eagerly awaiting the ultimate settlement in the Kimbrel case, because we have spent much of the last few months discussing the peculiar case of Mr. Craig M. Kimbrel. It is our suspicion that he will land much closer to the $7.25MM we have projected for him than the high number the model produced, which I might as well confess was actually $10.2MM. However, Kimbrel is not just the namesake of the rule; he is also the first test of the Kimbrel Rule. While a couple other players will have their salaries dampened by the Kimbrel Rule in 2014, the amount that it changed their salary projection was under $1MM in these cases, nothing that could shine a light on the theory. Going forward, this may be a clue about how to treat exceptional cases for us (and possibly, for the teams and players themselves). Of course, maybe the Braves will hurry up and settle with him before he breaks my computer.