Saturday, October 19, 2013

Say Goodbye to Jim Johnson

Jim Johnson
Jim Johnson celebrates one of his 101 saves between 2012 and 2013. He shouldn't see any more in black and orange.
The Baltimore Orioles and their fans have seen two very different Jim Johnsons over the last two seasons. In 2012, Johnson was lights out, finishing a club record 63 games and recording a club record 51 saves (out of 54 total opportunities). In 2013, Johnson recorded 50 saves, seemingly duplicating his 2012 output. As any Orioles who watched his games knows, the 2013 version of the closer was prone to the big inning, blowing 9 heartbreaking saves that the Orioles totally could have used at the end of the year as they finished 5.5 games behind the playoff-bound Tampa Bay Rays. So what should the Orioles be looking at? And should Jim Johnson return to Baltimore as an Oriole in 2014 and beyond?

First things first: the Baltimore Orioles have expressed faith in their closer and have announced their intention to tender a contract to Jim Johnson. With that out of the way, let's look at whether Johnson deserves the hefty raise he's expected to receive in his final year of arbitration.

Further, I am fully aware of and support the argument against designated closers. Often, a game's highest-leverage situations in which a reliever would be used occur before the 9th inning and would be, logically, better fits for the pitcher designated as the best and most reliable member of the bullpen. Since apparently the only 30 people in the world that are still blind to this pretty sound reasoning work in baseball and in a position to judge closers, my argument is based on the antiquated statistics used by teams to evaluate pitchers in traditional closers: saves and blown saves.

In 2013, the Baltimore media ignored average save rates. Reporting, for the most part, fell along two tracks: "Jim Johnson sucks!" or "Jim Johnson just needs to work some things out!" Both of these storylines contributed so much to the conversation. Even some "The Orioles can't go 2-for-11 with RISP and expect to win that game, we can't blame Johnson," stories cropped up, which I would usually agree with except Baltimore was winning after the 7th inning in all three games against the Diamondbacks (all ended up being losses) and therefore seemed to do just fine with scoring runs. You can't push 10 across the plate every game; it doesn't work like that. Sometimes you score 1 run and sometimes you score 6 and you can't and shouldn't let a guy who isn't doing his job off the hook because you don't score more than league average more than 50% of the time. Averages don't work like that.

Jim Johnson's 2012, which featured 52 saves in 55 opportunities, was an impressive, club-record setting campaign. But total number of saves are a function of total number of opportunities: a closer given 20 save opportunities can't have any more than 20 saves, and a closer given 100 opportunities will blow 52 saves out of the water. So, as an equalizer, we can look at save rates. Jim Johnson earned a 94.5% save rate in 2012! A great save rate! Except that's only a little better than league average. The Orioles won 98.7% of games in which they led after 8 innings last year, just above league average.

Since this data not qualified by score differential, I would assume that teams leading after 8 with a 3-or-fewer run lead would win slightly less than 95% of games; therefore, JJ's save rate was pretty good. But it's important to remember that most of those games would have been won anyway. The probability not scoring in a given inning is between 66% and 78%, depending upon the team's average runs per game. So off the bat, we can say that a major league pitcher shouldn't be allowing 1+ runs any more than 22%-34% of the time. Any pitcher in the league with an infinite number of save opportunities would post an overall save rate of at least 66%, and probably closer to 80% for the average closer, since closers are typically better pitchers than run-of-the-mill relievers. In fact, only Heath Bell posted a save rate of less than 70%.

Based off of a very limited study done with a very limited sample size, the average save rate in one-run games for closers who saved 20+ games is approximately 75%, which seems accurate because it falls just above the expected floor for save rates, accounting for random chance and quality of pitchers.

While acknowledging that the league winning percentage when leading after 8 innings is influenced by active closers, if you believe that closers are more talented pitchers, we can pretty accurately say that most closers pitch in a situation in which most professional pitchers would be able to not lose a game. If we knock a few percentage points off of the league winning percentage when leading in the 8th inning because closers were pitching (a very non-scientific adjustment for talent) and a few more because some of those games were not save situations, it would mean that in 2012, his best year ever and one that he has little history of showing is the norm and not an aberration, he was at most a handful of percentage points better than league average. Johnson was one of 10 pitchers to post a save rate of 90% or better in 2012, and actually posted just the third best in baseball.

Jim Johnson is bummed out by all of this. In 2013, Johnson went 50-for-59, good for an 84.7% save rate. The most up-to-date information shows that Johnson was 9/16 in one-run save opportunities. By my previous guesstimate, that puts him at a 56.25% save rate in one-run games. That's well below league average in save percentage overall and is an anchor in the the statistical tail end that drags down the 75% save rate in one-run games, assuming that trend holds. In fact, you would expect a league-average pitcher to blow, at most, 34% of an infinite number of one-run save opportunities based on the probability of scoring a single run in an inning. Jim Johnson blew 43.75% of them.

Now, obviously, Johnson is working with a limited sample size and would likely perform better than a 72% or 75% save rate in an infinite number of one-run save opportunities if afforded the opportunity, but in 2013, he was below league average in this flawed statistic.

Obligatory rundown of statistics: Jim Johnson failed the eye test in 2013, as most Orioles fans noticed that his sinker stayed straight and up in games he blew (and many he didn't), but he pitched twice as many double plays (GDP) as he did in 2012 (17 to 8). GDPs are also a function of baserunners, so it may be that he simply had more opportunities to pitch GDPs in 2013. Johnson also posted a better strikeout percentage in 2013 (19.5% to 15.2%), much better than his career average. This contributed to his big drop in balls in play in 2013, when only 70% of plate appearances resulted in a play in the field (compared to 77% in 2012). His walk rate was up slightly (6.2% to 5.6%), but still fell below his career average of 6.9%. Where he fell off was in line drive percentage (LD%), or the percentage of balls put in play that are well-hit. In 2012, Johnson put up a LD% of 13%, the best percentage of his career, but not significantly better than his 15% in 2011. In 2013, his LD% was 21%, similar to 2010's 25% and 2009's 18%. His rate of extra base hits went up a full percentage point (4.8% in 2013, up from 3.7% in 2012), likely a result of the increased number of line drives given up. He also gave up home run more often in 2013 (1.7% to 1.1%).

This is all to say that Johnson's numbers in 2013 were not entirely unexpected. He was not significantly worse in many statistical categories other than LD%, which is admittedly a very important metric. His horrible 2013 performance and his great 2012 performance were natural variation in the career of a closer, an incredibly volatile position that is forced to live on incredibly small sample sizes. If 2012 was a little longer, Johnson probably would have blown a few more saves. Actually, for the Orioles, 2012 was longer for the first time since 1997, and Johnson blew one save and lost another tie game in the ALCS. And if 2013 was longer, he probably would have blown a higher percentage of saves than in 2012, but not at the miserable rate he did over 162 games.

And herein lies the difficulty in paying closers big money. Johnson is expected to receive a raise from his 2013 salary of $6.5 million through arbitration, possibly reaching $9 million. Closers are volatile and subject to an awful lot of random chance, making it risky to reward one with big money, especially one past 30 and with no history of sustained success in LD%. Applying wins above replacement (WAR) to the situation would be difficult since relief pitchers work with such a limited sample size, but if you must, puts Johnson's 2013 at 1.5 WAR, and his 2012 at 2.4 WAR.

Effectively, the Orioles paid Johnson $4.3 million per win added to the team, for whatever weight that carries for pitchers in specialized, limited roles. That figure is much higher than the estimated value of a win in the New York market, and Baltimore is a significantly smaller market than New York. Jim Johnson was vastly overpaid for his contribution to the team in 2013, a mistake that should not be repeated in 2014 by a team with a limited budget compared to its divisional competitors.

Beyond that, Johnson is, like most relief pitchers not named Mariano Rivera, a big ball of variance rather than a rock-solid, lock-down closer. He can't be relied on to repeat 2012 (or 2013 for that matter) and instead, the Orioles should hunt for a bargain at the closer role that manager Buck Showalter seems to love or make moves based on matchups and leverage in the later innings the way most sabergeeks would love.

At Camden Yards after the 2012 All Star Game, I told a friend that it was time to trade Johnson because his value had peaked and his role was limited and glorified for what he actually contributed. Obviously, the Orioles did not follow my advice. That doesn't mean that they should cling to a pitcher that has set team records related more to opportunity than ability, and definitely not a pitcher in an out-of-date role that is subject to a great deal of random chance, and most certainly not one in line to possibly make $9 million. That money should be used to grab a player the team desperately needs, like a starting pitcher, that will have a bigger impact on more games.

My argument against Jim Johnson isn't that I expect him to be as bad in 2014 as he was in 2013; in fact, I'm very sure that he will improve on this past season, mostly because there's nowhere else for him to go but up. I believe that retaining Johnson's services for another season represents a volatile, unpredictable investment that will likely produce at a rate consistent with any pitcher in a similar role available on the market at the cost of an investment offering much higher rewards.

No comments:

Post a Comment