The Marlins Bullpen In Shutdowns and Meltdowns

Lately, the Miami Marlins have been involved in a handful of late-inning close games. More than a few of those late-inning affairs involved the Fish blowing leads or maintaining them via strong performances by the bullpen. In that respect, the Marlins have been one of the most exciting teams in baseball this season. In fact, the team ranks fourth in average play Leverage Index (LI, a measure of how important a plate appearance or situation is compared to an average PA, which is set at 1.0) for their hitting PA and second in pitcher average LI, In short, the Marlins have kept things exciting, and sadly, part of that has been the bullpen's doing.

Here at Fish Stripes, we strive to use statistics that better capture the world of baseball, and in no area is that needed more than in the world of reliever analysis. The save statistic has dominated that scene for a long time, and it is such a poor stat that people need to see something that better reflects the game. To that end, Tom Tango devised an interesting statistic that is based on the Win Probability Added (WPA) stat that we use often in our game threads and for determining Ichthyomancy winners. FanGraphs' Steve Slowinski explains the rules, which are much simpler than the rules of the save.

Using Win Probability Added (WPA), it’s easy to tell exactly how much a specific player contributed to their team’s odds of winning on a game-by-game basis. In short, if a player increased his team’s win probability by 6% (0.06 WPA), then they get a Shutdown. If a player made his team 6% more likely to lose (-0.06), they get a Meltdown.

The +/- 6% cutoff puts SDs and MDs on a similar scale as saves and holds, meaning 40 shutdowns is roughly as impressive as 40 saves or 40 holds. Dominant closers or set-up men will typically have 35 to 40+ shutdowns and a handful of meltdowns.

Meanwhile, meltdowns are more common than blown saves, and they can happen to both closers and non-closers alike. The worst relievers will rack up around 10 to 15 meltdowns in a season.

The concept is really simple: if you help your team win by more than six percent of a win, or 0.06 WPA, you get a Shutdown. Do the opposite, and you get a Meltdown. Given the work of the bullpen this season thus far, it would be interesting to compare how the Marlins are faring as a team this year versus last year's similar bullpen.

Last year, the Marlins bullpen did well as a unit in terms of run prevention; they were seventh in baseball in ERA (3.44) and eighth in FIP (3.62), so both their pure run prevention and their peripheral numbers were solid, if unspectacular for a relief crew. The team certainly fared better than the year before, and it was a credit to the Marlins having invested in the bullpen and having more than a few balls bounce their way as well. This season, the team is 16th in ERA (3.89), though they are also eighth in FIP again (3.45). This likely means that, despite the frustrations of the season thus far, the Fish are actually about the same in terms of performance from last year to this year.

But the frustrations of each of the last two years are very clearly shown in the Shutdowns and Meltdowns statistics. Last season, the Marlins managed 127 Shutdowns, which was tied for 12th in baseball. However, true to the nature of the team's Leo-Coaster closer, the club also managed 81 Meltdowns, which was the fourth-highest total in the league. This year, the numbers are similar, with the Fish ranking 10th in Shutdowns with 28 and third in Meltdowns with 20.

As a result of the high Meltdown counts in each of these two seasons, the team's Shutdown Rates (SD / (SD+MD)) have been lower than the league average in each season. Last season, the Marlins were 25th in Shutdown Rate, with a rate of 61 percent versus the league average of almost 65 percent. This season, the Marlins have actually done better in SD Rate relative to the rest of the league; they rank 20th in the league with a 58 percent mark versus a league average of 63 percent. In either case, however, the Marlins were below average, and that lends directly to the fan frustration that has been seen thus far with the bullpen.

Individuals At Fault

It is still quite early in the season, but we already can see the relievers who are potentially at fault for the Marlins' bullpen "struggles" despite decent peripheral statistics. Here's the Fish leaderboard in the early going.

As you can tell, a number of the Marlins' relievers have at least done a league average job at holding leads and performing. There is no surprise that Steve Cishek is leading the way, as Cishek holds a 1.65 ERA and 2.49 FIP to start the season. Similarly, it is good to see Randy Choate up there on the list, as it shows that Choate has been used mostly correctly by Ozzie Guillen; had he been used inappropriately, you know he would have a much worse rate than 66 percent and a much worse FIP than 2.43. Ryan Webb is also a pleasant surprise, as he has posted some good numbers while holding a 20 percent strikeout rate, the highest of his career thus far.

Of course, the bad is also quite obvious. Mujica has struggled this season and his numbers are down in all of his peripherals. Heath Bell took yesterday's tied game and left the ballgame having given up two runs and potentially the game had the Marlins not bailed themselves out against the New York Mets' pen. Bell's struggles have been well publicized for his problems, and given the horrific peripheral numbers we are also seeing from him, it is very likely that he issue is not just something that will regress fully back to the mean. Dunn has been awful all season and was demoted for a short period of time by the Marlins.

Let's look at last season's crew by comparison. Here are the top seven in Shutdowns.

Ironically, this is the polar opposite of what Marlins fans may have expected. It was in fact the middle of the bullpen that struggled for the Fish in 2011, at least in terms of holding leads and not blowing games. Dunn and Webb were particularly terrible at preventing problems for the later innings, while Mujica and the man formerly known as Leo Nunez, Juan Oviedo, actually held leads just fine. In fact, Oviedo's SD and MD numbers were comparable to better relievers like Craig Kimbrel (35 SD, 10 MD), Sean Marshall (32 SD, 11 MD), and (gasp) 2011's Heath Bell (31 SD, 8 MD).

Word of Warning

Shutdowns and Meltdowns better credit relievers of all kinds, not just closers, for their performance in the later innings. However, there are definite limitations in the stat in that, like saves and holds, they are not very predictive of future performance. It would be much better for you to look at peripheral ERA predictors such as FIP and SIERA to try and guess at future performance rather than use counting stats like these.

It is quite telling also that the Marlins made a good number of executive decisions based on the results of counting stats like these. The top two pitchers in Shutdowns were brought back by the Marlins this season, one of whom (Oviedo / Nunez) was brought back solely because he was "the closer" and had racked up saves and, consequently, Shutdowns. One suspects the Marlins would not have considered a pitcher with an ERA of 4.00 to be worth $6 million without the saves, and in fact the team let go Burke Badenhop (4.10 ERA, 2.95 FIP) mostly because he did not perform well in terms of Shutdowns / Meltdowns performance. It is another strike in the mounting evidence that, in terms of evaluating relievers, the Marlins front office is still stuck in the past.