Is Ryan Worth It?

I tried to defend him. I really did. And then, J.P. Ricciardi took all the good faith and signed B.J. Ryan to the craziest contract of all time. Five years, $47 million. What was he thinking?

Actually, that’s a good question. What was he thinking? Ricciardi is supposed to be a smart man. I know he has smart people working for him. So what happened? When did the Blue Jays front office decide that it would be smart to make Ryan the highest paid closer of all time? Didn’t they read this study at Baseball Prospectus? For those that don’t click on that link, let me summarize: signing closers in their late 20s or early 30s (Ryan is 29) is a bad idea. Period. The end.

But maybe the Jays are onto something. Maybe every other front office read this study (I know, yeah right), but Toronto did its own study and came to its own conclusions. Maybe. In that case, it makes sense for me to do my own study, I guess, and see what I can come up with.

Here’s my methodology: Using the Lahman Database, I found every pitcher who had a 20 save season before 2001 (so there would be at least five years of data to look at). I then compared them to Ryan using my own version of similarity scores, employing a statistical technique known as z-scores. Basically, I looked for the players who were most similar to Ryan in four categories: Strikeouts per Batter Faced, Walks per Batter Faced, Home Runs per Batter Faced, and Batters Faced. Thus, I found the pitchers who were most similar to Ryan in terms of fielding independent numbers and in terms of how many innings they pitched (which removed guys that may have had arm trouble due to high use).

Then, using the ages 27-30 as my cutoff, I compiled a list of the ten most similar pitchers to Ryan, at a similar age. The pitcher seasons that were used as comps for Ryan’s 2005 are listed at the bottom of this article. Here’s a comparison of the average comparable season used and Ryan’s actual 2005 numbers:

Player    IP    ERA+    RSAA    RSAA2
Comp      77     168      14       16
Ryan    70.1     170      15       15 

RSAA is runs saved above average and RSAA2 is adjusted for the expected park-adjusted ERA in Toronto next season. This adjusts for the fact that our comparable pitchers had an overall league ERA lower than what the ERA is now. Anyway, the point here is that the most similar season’s to Ryan’s 2005 and Ryan’s 2005 performance are indeed very similar. Thus, it’s not much of a stretch to use the averages of the next five seasons to see just how well Ryan might be expected to do. In fact, that’s what I did next.

Season        IP     ERA+    RSAA  RSAA2
One           60     117       4       4
Two           57     133       6       7
Three         60     128       6       7
Four          51     176      10      11
Five          34     113       2       2

Let me note that throughout the study, ERA was weighted by innings, so that the performance of a guy who threw three innings is not counted as equal to the performance of a guy who threw 73 innings. You can see that the road here is a little bumpy in terms of runs saved: After an expected decline of 11 runs next season, we would actually expect Ryan to get somewhat better the next three seasons in a row. But you can see here that his expected innings would continually drop, as his comparables get injured or fall out of the league (four out of the ten were not pitching by year five), and that it is unlikely that Ryan will ever post as good an ERA as he did last season. That’s expected. But how much would the above performance be worth?

I need to add an extra step to find that out. Closers pitch in higher leverage situations than most pitchers, so we need to adjust for that. A shutout ninth is worth a lot more than a shutout first. But by how much? I used Baseball Prospectus’ Leverage to find out. I adjusted BP’s Leverage so that the league average was equal to one (otherwise, the average is about 1.11), and then found and averaged (again, weighted by innings) the Leverages of every pitcher who had at least 30 saves in 2005. These, in my estimation, are “true closers,” and by looking at them we should be able to tell how much more valuable a closer’s innings are. The answer was 1.6

That is, every inning a closer pitches is worth 1.6 innings thrown by any old pitcher. B.J. Ryan’s adjusted Leverage for 2005, by the way, was 1.42.

Anyway, knowing that, I can calculate adjusted runs saved. And then, using a methodology similar to the one used in my Burnett article, I can find Ryan’s expected value over the next five years. Using Dave Studeman’s excellent past research, I found that the average run saved by a reliever is worth about $284,000. The average pitcher signed to a multi-year contract on the free agent market is worth about $4.7 million. That sounds about equal to a contract you would expect an average closer (say, Danys Baez) to get.

But there’s one more thing to adjust for, inflation. Every year, the average major league salary increases by about 10%. That’s held true for the past five, fifteen, twenty-five, thirty-five years. We need to adjust for that. $10 million today is equivalent to about $6 million five years ago. Five years down the line, it’ll be worth about $16 million. So, adjusting for that, here’s the last table, of how much Ryan is worth based on my simple projection system:

Season        IP    ERA+    RSAA   RSAA2     Adj   Value
Comp          77     168      14      16      26    12.1
One           60     117       4       4       6     6.9
Two           57     133       6       7      11     8.8
Three         60     128       6       7      11     9.4
Four          51     176      10      11      18    12.0
Five          34     113       2       2       3     8.5
Total                                               45.6

So Ryan’s comparables were actually worth about $12 million in their year similar to Ryan’s 2005. But remember, the Jays are paying for future production, not what Ryan has done in the past. And over the next five years, we expect Ryan to be worth roughly $46 million, almost exactly equivalent to the contract he’s getting from the Blue Jays. Unbelievably, this looks like a good deal, even with the declining innings and ERA and all.

Who else is surprised?

References & Resources
The ten most similar seasons to Ryan’s 2005 are:

Duane Ward – 1993
Tom Henke – 1986
Trevor Hoffman – 1996
Rob Dibble – 1992
Jeff Russell – 1989
Mel Rojas – 1996
Goose Gossage – 1981
Lee Smith – 1987
Robb Nen – 1998
Billy Wagner – 1999

Using Recurrent Neural Networks to Predict Player Performance
Technology is rapidly advancing possibilities in decision-making.

Print This Post

Comments are closed.