/cdn.vox-cdn.com/uploads/chorus_image/image/27906069/181834018.0.jpg)
Pitchers are risky. If you run a major league team, it is impossible to avoid that risk no matter how hard you try. Every team tries to develop young arms and guide them through the minors to a productive major league career, but it doesn't always work out perfectly. When a team has a hole in their rotation and there is no can't-miss prospect ready to fill it, they have to turn to the free agent market. With fewer and fewer stars reaching free agency, this often means weighing the cost of significant risks against fairly limited upside.
More from our team sites
When the offseason began, the market for starting pitchers was thin. Now that 20 pitchers who project as full-time starters have signed major league deals, pickings are even slimmer. Masahiro Tanaka has the biggest deal so far this winter, and it is impossible to imagine anyone else topping the seven years and $155 million he got from the Yankees. He is the only pitcher to hit the market this season with true top-of the-rotation potential and his headline-grabbing deal reflects just how difficult it has become to land that level of talent on the free agent market. Teams are locking up the best starters early in their careers whenever they can and the pitchers who make up the bulk of the market now tend to be the guys with the most risk surrounding them.
The need for quality pitching sets a certain price for starters, and the risks inherent in pitchers are presumably included in this price. Individual players have unique risks that affect their price, and if we can better understand these risks, we should have a better idea of the dollar amounts teams are willing to pay for a player. A quick look at the free agent starters who have signed new deals this offseason seems to indicate that teams price certain types of risk differently from others. This sample is too small to give us an accurate gauge of how these risks are priced in the long run, but it does hint at some of the overall trends in play.
To look at the cost teams assign to risk, I needed some basic idea of the cost teams are paying per win for starting pitchers this offseason. There has been a fascinating debate recently about the best way to calculate dollars per win. The folks at Fangraphs, led by managing editor Dave Cameron, have been advocating a system based on projections, since teams are buying future performance. Cameron and company have put the cost at approximately $6 million per win right now. At Beyond the Boxscore, Lewie Pollis published the results of his own study, showing that when examining past data, teams actually end up paying just over $7 million per win. Though both studies have their merits, for the purpose of looking at risk assessment, I am following Cameron's projection-based technique and its baseline of $6 million per WAR. Using the 2014 Steamer projections, here is the $/WAR value for all of the starters who have signed new deals this offseason.
Name |
ERA |
GS |
IP |
FIP |
WAR |
Average Annual Salary |
$/WAR |
Hiroki Kuroda |
3.95 |
32 |
189 |
3.66 |
3.4 |
$15.00 |
$4.4 |
Scott Kazmir |
3.7 |
27 |
163 |
3.54 |
2.6 |
$11.00 |
$4.2 |
Ricky Nolasco |
4.36 |
30 |
192 |
3.97 |
2.4 |
$12.25 |
$5.1 |
Bartolo Colon |
3.78 |
29 |
182 |
3.54 |
2.2 |
$10.00 |
$4.5 |
Dan Haren |
3.54 |
30 |
173 |
3.55 |
2.2 |
$10.00 |
$4.5 |
Scott Feldman |
4.34 |
27 |
173 |
3.99 |
2.1 |
$10.00 |
$4.8 |
Josh Johnson |
3.6 |
25 |
144 |
3.31 |
2.1 |
$8.00 |
$3.8 |
Tim Lincecum |
3.9 |
31 |
182 |
3.54 |
1.9 |
$17.50 |
$9.2 |
Phil Hughes |
4.22 |
24 |
144 |
4.3 |
1.4 |
$8.00 |
$5.7 |
Mike Pelfrey |
4.66 |
25 |
144 |
4.36 |
1.4 |
$5.50 |
$3.9 |
Tim Hudson |
4.05 |
30 |
173 |
3.62 |
1.5 |
$11.50 |
$7.7 |
Edinson Volquez |
4.41 |
7 |
38 |
4.01 |
0.2 |
$5.00 |
$25.0 |
Matt Garza |
3.78 |
29 |
182 |
3.59 |
2.6 |
$12.50 |
$4.8 |
Masahiro Tanaka |
3.71 |
31 |
192 |
3.27 |
3.8 |
$22.14 |
$5.8 |
Bruce Chen |
4.38 |
17 |
99 |
4.61 |
0.6 |
$4.25 |
$7.1 |
Jason Vargas |
4.34 |
28 |
180 |
4.3 |
1.9 |
$8.00 |
$4.2 |
Ryan Vogelsong |
4.39 |
30 |
176 |
4.11 |
0.7 |
$5.00 |
$7.1 |
Jason Hammel |
4.26 |
17 |
96 |
4.05 |
1.1 |
$6.00 |
$5.5 |
Gavin Floyd |
3.68 |
8 |
48 |
3.59 |
0.7 |
$4.00 |
$5.7 |
Roberto Hernandez |
3.81 |
22 |
127 |
3.54 |
1.7 |
$4.50 |
$2.6 |
$/WAR Average |
$6.3 |
||||||
Median $/WAR |
$4.8 |
As you can see, the $/WAR average is right in line with Cameron's results, while the median is a bit lower. I have the average and the median dollars per win here to help deal with the impact a few outliers have on this small sample. Using these two numbers also gives us a nice range for predicting the annual value for the contracts of the remaining unsigned starters who are likely to sign a major league deal.
ERA |
GS |
IP |
FIP |
WAR |
Low |
High |
|
Ubaldo Jimenez |
4.03 |
30 |
192 |
3.79 |
2.5 |
$12.0 |
$15.7 |
Ervin Santana |
3.96 |
31 |
192 |
3.71 |
2.7 |
$12.9 |
$17.0 |
Bronson Arroyo |
4.49 |
27 |
153 |
4.25 |
1.2 |
$5.7 |
$7.5 |
A.J. Burnett |
3.63 |
32 |
189 |
3.27 |
3.4 |
$16.3 |
$21.4 |
Paul Maholm |
4.29 |
17 |
101 |
3.95 |
1 |
$4.8 |
$6.3 |
Joe Saunders |
4.57 |
32 |
198 |
4.15 |
1.4 |
$6.7 |
$8.8 |
Chris Capuano |
4.25 |
32 |
185 |
4.03 |
1.6 |
$7.7 |
$10.1 |
Aaron Harang |
4.58 |
25 |
149 |
4.49 |
0.7 |
$3.3 |
$4.4 |
Erik Bedard |
4.29 |
32 |
171 |
4.21 |
1.9 |
$9.1 |
$12.0 |
Of these numbers, a few are probably high due to some favorable projections from Steamer, but nothing feels wildly out of line. Still, there is a very wide range for some players. If Ervin Santana is looking at a four-year deal, for instance, the difference between his high and low is 16.4 million in total salary. This is where considering specific risks might be helpful.
I have selected three different types of risk and picked free agents signees that fit well to those categories to see if the results point to anything.These aren't projections for players just breaking in to the game, so they don't have unusual amounts of uncertainty surrounding them. Teams certainly aren't pulling up Steamer projections to set their offers, and there are many factors influencing these numbers, but one thing that the variations in $/WAR might indicate is the way teams value risk. For the most part, it is probably safe to assume that the deals free agents have signed are the highest bids they received, or at least a fair market value. Looking at players with easily identifiable types of risk attached to them and seeing which ones receive a lower-than-average $/WAR value should tell us something about the way teams perceive these risks Again, this is too small a sample to draw any strong conclusions, but the results can still be interesting and informative.
The first risk type is age. Getting old is rough. You have to get up to pee at night more, you wake up achy and just as tired as you were when you went to bed, and your baseball performance declines, sometimes rapidly. Three full-time starting pitchers signed this offseason are over 35:
Name |
ERA |
GS |
IP |
FIP |
WAR |
Average Annual Salary |
$/WAR |
Hiroki Kuroda |
3.95 |
32 |
189 |
3.66 |
3.4 |
$15.0 |
$4.4 |
Bartolo Colon |
3.78 |
29 |
182 |
3.54 |
2.2 |
$10.0 |
$4.5 |
Tim Hudson |
4.05 |
30 |
173 |
3.62 |
1.5 |
$11.5 |
$7.7 |
$/WAR Average |
5.5 |
||||||
Median $/WAR |
4.5 |
Hudson is the outlier here, which could be due to the Giants being the Giants or because he was fairly sought-after this offseason. Kuroda had the qualifying offer hanging over his head, of course, and I believe that is reflected here in the $/WAR value along with his age. I left 36-year old Bruce Chen off here because his low number of projected innings (96) deflates his WAR total below that of a full-time starter's workload. Including him puts the average $/WAR at $5.9 million, right in line with Cameron's data. Projecting Chen out as full-time starter keeps the numbers roughly the same as above, but the Panamanian still comes with a higher $/WAR than Kuroda and Colon.
A $500,000 to $1 million difference in $/WAR between the market at large and the market for the oldest pitchers available seems reasonable. The three charted players are all coming off strong seasons and fall at or near the top of the market, so the discounted rate does seem to connected to their age, not their ability or recent performance.
The second type of risk is injury history. Recent injuries would seem to be an even greater risk, but in this limited sample it is not priced much differently than just being old:
Name |
ERA |
GS |
IP |
FIP |
WAR |
Average Annual Salary |
$/WAR |
Josh Johnson |
3.6 |
25 |
144 |
3.31 |
2.1 |
$8.0 |
$3.8 |
Tim Hudson |
4.05 |
30 |
173 |
3.62 |
1.5 |
$11.5 |
$7.7 |
Gavin Floyd |
3.68 |
8 |
48 |
3.59 |
0.7 |
$4.00 |
$5.7 |
Matt Garza |
3.78 |
29 |
182 |
3.59 |
2.6 |
$12.50 |
$4.8 |
$/WAR Average |
5.5 |
||||||
Median $/WAR |
5.7 |
The discount is the same as the one older player's face, and the median is much higher even with Hudson in both groups. Josh Johnson has the one of the lowest $/WAR values of any free agent to sign this offseason, which is understandable. Johnson mixed being injured with being unbearably awful last season, putting his career at a crossroads. Gavin Floyd, on the other hand, is a unique case in this group and might be pushing the numbers up a bit. He is coming off Tommy John surgery, which gives him a low inning total projection. If he comes back in May as he is currently expected to, he will easily top these projections and his $/WAR will be much lower. Without him, the average drops to $5.1 million while the median stays the same. I don't see this as compelling reason to omit him, however, because injury risk is still the central issue for him. Garza's $/WAR might be the most indicative of the overall trend here, though his poor finish in Texas is also a factor. He followed the exact same path from Chicago to Texas as Ryan Dempster, with very similar results, but Dempster got $1.25 million more in average annual value. Dempster was also in the "old" category at the time but still came out ahead, so injury risk seems to be primarily responsible for Garza's lower than average $/WAR deal.
The most interesting type of risk comes from inconsistent performance. Projections are based on previous performance, so when that previous performance is all over the place, the projections are more difficult to trust. That opens things up to extremes of risk and upside, and if the 2014 free agents are any indication, teams want no part in that kind of drama.
Name |
ERA |
GS |
IP |
FIP |
WAR |
Average Annual Salary |
$/WAR |
Ricky Nolasco |
4.36 |
30 |
192 |
3.97 |
2.4 |
$12.3 |
$5.1 |
Dan Haren |
3.54 |
30 |
173 |
3.55 |
2.2 |
$10.0 |
$4.5 |
Scott Feldman |
4.34 |
27 |
173 |
3.99 |
2.1 |
$10.0 |
$4.8 |
Jason Hammel |
4.26 |
17 |
96 |
4.05 |
1.1 |
$6.00 |
$5.5 |
Phil Hughes |
4.22 |
24 |
144 |
4.3 |
1.4 |
$8.0 |
$5.7 |
$/WAR Average |
5.1 |
||||||
Median $/WAR |
5.1 |
There is no systematic method for establishing what marks an inconsistent performance, I just looked for notable variations in ERA+ over the past three to four seasons. Haren also has the added inconsistency of a terrible first half to the 2013 season followed by a strong second half. Roberto Hernandez also fits in this category, but his false-identity snafu also plays a role in his price and since he is the most extreme outlier on the low end of the $/WAR scale, I think it is best to leave him off.
It is a bit surprising that teams seem to be more wary of inconsistency than age or recent injuries. Three of the five players here project out as just above league average by WAR, so the natural assumption would be that they would be right there at the market rate. Instead they are well below it, though the median is not far off from that of the full sample.
These are extremely small samples so making any broad generalization from the dollar values we get here for these risks is dangerous. However, looking at the nine remaining free-agent starters, even this limited data might be helpful.
Steamer might love A.J. Burnett a little too much, but his age shouldn't keep him from landing a deal on par with Hiroki Kuroda's. And if he plays as well as he projects, he'd be worth it. Bronson Arroyo has been struggling to find offers this offseason, which isn't surprising given his projections. It is hard to image that he'd go for just $7-$8 million on a one-year deal, but that's what his projected value is even before accounting for age.
Both Ubaldo Jimenez and Ervin Santana fall firmly in the "inconsistent" category and with qualifying offers attached to them, I think they can both rule out the high side from the projected salaries above. The risk that comes with their unpredictability will probably matter a great deal more than the upside. Both players could be impact arms for the right contender next season, but with a draft pick added to the price, they might even fall below the lower end of the salary range above.
More data is needed before we can conclude that this is actually how teams price out the risk that comes with age, recent injuries and inconsistency, but I think the framework has potential. Even these small samples yield some useful insights and with years of free agent contracts in the data set, we might begin produce some reliable estimates on the true cost that comes with specific types of risk.