24 November 2015

Chris Davis' Contract: Using A Comp Model

Yesterday, it was reported again that Peter Angelos is personally involved in talks concerning Chris Davis.  This was considered a given after he had publicly stated an interest in resigning Davis the first week of October.  Angelos has never lost a signing that he publicly acknowledged an interest in making.  With this in mind, I decided to go back and try to figure out just what might make sense.

Last week, FanGraphs put forward an article considering their top statistical comps for Chris Davis and looking forward to see where that population of comps wound up in order to illustrate what potential outcomes we might see in Davis.  This is not an uncommon way to determine what just might be in store for a player, but if can be a pretty dodgy comparison.  I had some concerns about the methodology Fangraphs used.  Davis notoriously is a late bloomer and career comps by age will throw in players who are more a product of long term accumulation as opposed to the short burst of performance Davis has seen.  Second, I think by focusing on wRC+ and ISO, it obscures specific aspects of Davis that get lost in that metric flattening performance to a single number.  Finally, I think the position pool is too large and should be restricted to similar players.

For my comp pool, I only looked at player seasons from age 26 through age 29 for first basemen, left fielders, and right fielders with a wRC+ above 120, an ISO more than 125% than league average, and an OBP less than 115% of league average.  That led to the following list:
Sammy Sosa 
Jay Buhner 
Frank Howard 
Lee May 
Cecil Fielder 
Adam LaRoche 
Kirk Gibson 
Jose Canseco 
George Scott 
Bobby Bonds 
Andres Galarraga 
Darryl Strawberry 
I used this population to project performances from age 30 through age 39, a ten year outlay.  To accommodate for survivor bias, performance for players who no longer were playing were accounted for by regressing performance to the mean with an age factor included.  Going back to the list, we see a similar player emerge.  These players largely depend on hard contact and not much else.  Some of these players developed a few seasons in their later career where they figured out how to walk for a couple seasons, but that skill was not typically found through out their careers.
2016 30 518 26 .263 .333 .482 2.2 15.4
2017 31 522 28 .265 .339 .498 2.5 18.4
2018 32 560 29 .271 .356 .501 3.2 24.7
2019 33 509 26 .261 .346 .481 2.4 19.4
2020 34 481 20 .248 .334 .438 1.5 12.8
2021 35 381 17 .243 .316 .438 0.9 8.0
2022 36 404 16 .237 .310 .418 0.6 5.6
2023 37 385 16 .222 .299 .405 0.2 2.0
2024 38 355 13 .211 .283 .379 -0.3 -3.1
2025 39 357 12 .200 .270 .356 -0.7 -7.6
Mean 203 .251 .333 .457 12.5 96
The mean player in this population remained a viable starter through age 34 (five seasons).  However, they performed at a first division level (above 3 WAR) for only one season.  That season was not age 30, but actually age 32 as several players tend to mature a bit later than normal in this group.  Unfortunately, we project that one season as the only season where Davis would be worth more than the expected 20+ MM he is expected to earn this off season.  It should also be noted that I compared this mean expectation with two proprietary models and mine was modestly optimistic.  The proprietary models were more in line with the FanGraphs mean.

Now, the mean result does not mean much without some concept of range.  To illustrate range, I divided this group into top and bottom halves and am reporting the averages of those groups to give some idea of what is a reasonable boom or bust.
Year Low Mean High
2016 1.3 2.2 3
2017 0.9 2.5 4.1
2018 1.5 3.2 5
2019 0.9 2.4 3.9
2020 0.7 1.5 2.2
2021 -0.1 0.9 1.8
2022 0.1 0.6 1.6
2023 -0.6 0.2 1.7
2024 -0.2 -0.3 0.3
2025 -0.5 -0.7 0.3
Above, we see the top half projection would be a first baseman who perform as a first division first baseman for four seasons before performing as a second division (1.5 - 2.9) first baseman for an additional four seasons.  Those first four years help keep the contract on value for several years after he no longer warrants a 20+ MM salary.  Of course, if his play collapses (a la 2014), then the Orioles would be stuck with an albatross.

Here is the value outlay:
5 6 7 8 9 10
High 141 157 172 190 192 196
Mean 91 99 104 106 103 96
Low 41 39 40 35 32 27
I would assume that you would rarely see a player signing for the mean because with player evaluation someone is likely to appreciate the player and expect better production than the mean.  That said, I would not want to push the club beyond paying for what Davis might be worth if he winds up being more like the players in the top half.  With that thought in place and with the assumption that he will command 25 MM a year, six years would be where I would stop (6/150).

Our BORAS model pegged Davis at five years and 104 MM, which would be a slight overpay given the comp player pool.  Based on the current chatter though, that sounds like a low ball offer.  This comp group appears to value Davis slightly more than other models of which I am aware.  In other words, all models that I know of think a deal of 6/150 or along those lines with be a loss of payroll flexibility.


Anonymous said...

The Orioles should NOT have let Frank Robinson, Don Baylor, or Eddie Murray go either.

Jon Shepherd said...

Baylor deal would have worked out smashingly if they respected what Reggie Jackson was.

Matt Kremnitzer said...

Yeah, you know, Frank Robinson, Eddie Murray, Chris Davis... all basically the same players.

Anonymous said...

The likelihood of this happening, based on past experience, is zero. I don't know why anybody is talking about it.
And that's OK, I can't think of any long-term contract that has been worth while. The exceptions are so rare that I don't know why anybody does these at all.
I think it would probably be much more productive to look at the anticipated nontenders and see what we can find in there,

Jon Shepherd said...

I think the reason why people are talking about it is that Peter Angelos has never personally and publicly advocated for signing a player who in the end signed elsewhere.

Salisbury Dave said...
This comment has been removed by the author.
Anonymous said...

What is the sample size for players that Angelos has advocated for?

Jon Shepherd said...

Not many. Based on memory...six or seven.

tony2302 said...

could it be possible that Pete decides to pay or overpay and still spend on other pieces? stranger things have happened. doubtful but possible

Jon Shepherd said...

I have no idea what Angelos is considering. Personally, I think the club is sold within five years. I do not know how that impacts his concept of payroll. It might mean he wants a healthier club on the field or it might be he wants assets that could be turned over in 2019 or so.

Anonymous said...

Matt, Davis does not have to be the same as the three HOFers to be worth keeping although you have no reason to believe he could not sustain his production. Baseball Reference has his "similars" through age 29 as guys like Richie Sexson, Cecil Fielder, Mark McGwire, Lee May, Willie Stargell. Obviously, the future is not predictable and Davis could just as easily be Bobby Bonilla or Albert Belle to the O's. More often than not, a player's success is at least somewhat tied to his environment. Players that stay with the teams that made them a success are more likely to build HOF credentials. Sometimes picking the right team makes a difference too. Teams build legacies through consistency once they build a good product. The Cardinals will always be a better "team" than the Dodgers are (now) with their willy-nilly team-building. The Orioles should be following the Cardinals' model.

Jon Shepherd said...

We could focus only on the successes or on the middling performances or on the wretched players. I think the above article discusses all of this.