Recent Posts

Archives

Links


« ESPN: Tigers sign Miguel Cabrera to 8-year Deal | Main | New Citi Field interior pictures »

Building a Bullpen, Part III

By Jason E | March 24, 2008

A successful bullpen depends not only on its pitchers, but also on the manager using them effectively. Willie Randolph has drawn criticism for his bullpen management, as has every other major league manager. In part three of our series on the bullpen, we focus on Randolph and whether the criticism is justified.

Bullpen management boils down to two important tasks. First, a manager must put each of his relievers in the best position to succeed. To do so, he must make use of batter-pitcher history, the splits of both pitcher and hitter, and gut instinct. But to keep pitchers effective, the manager must also keep them fresh. The success of managers in these aspects is extremely difficult to quantify. Last year, Billy Wagner’s ERA was two runs higher in the second half than in the first. Was he tired or simply due to regress? If he was worn down, was there anything Willie Randolph could have done to prevent that? Any answer to these questions is entirely conjecture, as the only evidence that can be offered is anecdotal.

One thing we can examine is how well Randolph has made use of platoon splits. It is a matter of simple common sense that Pedro Feliciano should face a greater percentage of left-handed batters than Joe Smith does. Feliciano held lefties to a .483 OPS last year, while Smith was clobbered to the tune of an .853 OPS. As expected, Feliciano faced 41% lefties and Smith only 28%. Plotted, it looks like this:

Feliciano is conveniently on the far left, Smith on the far right. That’s all the x-axis means, better against lefties on the left, better against righties on the right.

So what if we looked at the same thing for every significant reliever in Randolph’s tenure?

Dots all over the place.

The far left dot is Scott Schoeneweis in 2007 and the far right dot is Braden Looper in 2005. What we’d expect here is a downward trend moving from left to right; that is, pitchers like Schoeneweis facing a high percentage of left-handed batters and Looper facing a much lower percentage.

Obviously, that is not the case. In fact, there’s essentially no correlation between these two variables for Mets relievers from 2005-2007. It’s actually trending upward slightly. Does this mean Willie is a terrible manager? Well, maybe, but consider that opposing managers aren’t going to send many lefties up there against Billy Wagner. We need some context. So here’s the trend line taken from the above chart of Randolph’s bullpen use compared to all of baseball last season.

As you can see, other managers generally obtain favorable matchups for their relievers while Randolph appears to do the opposite. Where they particularly outshine Randolph is in the use of lefty specialists. (Note the upward curve of the yellow line on the far left.) We shouldn’t be too harsh on Willie here, his sample size of left-handed specialists is all of… 1. Scott Schoeneweis in 2007. Still, Schoeneweis faced only 40.8% left-handed batters, meaning that in 6 out of 10 at-bats, he was likely to get clobbered.

Even though it’s over three years, there’s still not much of a sample for Randolph. So can we conclude anything? I think we can.

Conclusions
The numbers back up something a lot of Mets fans probably noticed. In a few cases, Willie has left relievers in roles that it had long been obvious they weren’t suited for. For most of last season, he insisted Schoeneweis could get both lefties and righties out. In 2005, he left Braden Looper in the closer’s role. The problem with that wasn’t so much that Looper was bad, it was that he was brought in for the 9th inning regardless of who was coming to bat for the other team. Also, in the ninth opposing managers will pinch-hit more liberally, meaning that a pitcher’s splits get exploited the most when you put them in the closer’s role. Looper happened to have the most extreme splits on the team and the results were ugly.

Aside from those two examples, Randolph has done fine putting his relievers in a position to succeed. The problem is that those two cases were so utterly mismanaged that they make his overall results, seen in the graph, appear as if he’s paying no attention to Left/Right splits whatsoever.

As we noted in the beginning, there is a lot to keeping pitchers effective that we cannot measure. But what we can measure indicates that Randolph has made a few mistakes. Usually this involves putting too much trust in one of his pitchers to get the job done even when the situation makes that unlikely. In terms of what all this means for the 2008 bullpen, we’ll assume he’s learned his lesson on Schoeneweis (fingers crossed). The other pitcher who may become a problem is Jorge Sosa. Sosa has been destroyed by left-handed batters throughout his career and last year was no exception. However, he did come through against lefties in a few critical spots last season. We fear this may lead Randolph to put the ball in his hands again, with disastrous results.

In the next part of our bullpen series, we will continue to scrutinize Willie Randolph’s bullpen management with a look at whether he’s leveraging his relievers properly.

Topics: AnyBodyButBengie, Bullpen |

One Response to “Building a Bullpen, Part III”

  1. MetsBlog.com » amLinks: SNY, Facial Hair and Bullpens Says:
    March 24th, 2008 at 7:47 am

    […] at Mets Refugees, Jason E continues his explanation of what it takes to build a […]

Comments

You must be logged in to post a comment.