clock menu more-arrow no yes mobile

Filed under:

Red Reposter: Sabermetric Saturday

Key to a good bullpen: post lots of Shutdowns.  And try to avoid the meltdowns.  The Reds have had their share of meltdowns this week...
Key to a good bullpen: post lots of Shutdowns. And try to avoid the meltdowns. The Reds have had their share of meltdowns this week...

After taking last weekend off for finals, Sabermetric Saturday returns!

Shutdowns & Meltdowns | FanGraphs Baseball
Perhaps the biggest sabermetric news this week was the development (by Tango and others at The Book Blog) and establishment on Fangraphs of a new statistic!

Here's the deal. We all hate saves, right? Saves are supposed to be counting stats that give us an idea of how many times an ace reliever shut down the opposition, right? But that's not really what they tell us. How many times, for example, have you seen an ace reliever come in to pitch an easy 9th inning with a three run lead to get a "save." But when a setup reliever comes in and shuts down the rally in the 8th inning, a key moment when the game outcome hangs in balance, at best he might hope to get a "hold," (and holds have their own problems...). Furthermore, there's considerable evidence to suggest that the mere existence of the save statistic has caused managers to use their bullpens less effectively, saving their closer for a "save" situation when the most important moment in a ballgame might instead be happening in the 7th or 8th inning!

Enter Shutdowns and Meltdowns. Shutdowns is a stat that essentially can replace saves and holds. It is defined as an instance in which a reliever accumulates more than 0.06 WPA in a game. This will occur at times when a reliever pitches through an important part of a game for a team that is currently in the lead. Pitching a scoreless 9th in a close game (1-2 run lead) does count as a shutdown. But so does shutting down a bases-loaded jam in the 7th of a close game. The number of shutdowns per season will be roughly the same as the number of saves + holds.

Meltdowns, on the other hand, can be used as a replacement for blown saves. They occur any time a reliever gets a WPA of -0.06 or worse in a game. This will happen in most blown saves, but they also would occur if you give up a run in a tie ballgame in the 9th inning. Here are the 2010 Reds reliever shutdown and meltdown totals through entering Friday's games (format below is Name: Shutdowns-Meltdowns):

Francisco Cordero: 10-3 (including two meltdowns this week)
Arthur Rhodes: 6-1
Nick Masset: 5-3
Daniel Ray Herrera: 4-4
Micah Owings: 3-1
Mike Lincoln: 3-2
Logan Ondrusek: 2-3
Carlos Fisher: 0-1

Here's why I like this stat: Cordero received a shutdown for all but one of his saves: the 3-run save against the Dodgers on 4/22.  So most saves are still relevant, just not the absurdly easy ones.  But Cordero also gets credit for two key appearances in non-save situations (4/8 and 5/3), both times when he held the line late in a tie ballgame.  Ultimately, I think shutdowns are doing what we always wanted saves to do for us.  

So let's use it moving forward as much as we can in place of saves.  You can always find current and past leaders in shutdowns and meltdowns at Fangraphs in their Win Probability section of their pitcher pages.  Here, for example, are the 2009 Reds.


Soria and Leverage Index - Royals Review
Jeff Zimmerman gives us a neat visualization of the efficiency of how the Royals have used their bullpen. I love this. I'm going to put this together for the Reds relievers too, if someone doesn't beat me to it.

The one caveat I'll throw out here: it's probably not possible to be perfectly efficient with your use of your closer, always using him at the most important possible instances.  Obviously, you never really know what's going to happen later in a game, and so it's hard to know when the time is "right" to use your closer early.  Furthermore, you have to have your closer ready and warmed up in that exact time when he is needed. Sometimes, a situation becomes bad too quickly to have your main gun ready to go. Furthermore, you need to be careful about when you get your closer warming up--you can't get him up 3-4 times per game, or you could stress his arm and/or make him tired. But that said, surely it can be done better than the typical manager does it in modern baseball. In fact, we know it can be done better, because often managers do it better when in the playoffs.


Ryan Howard's Extension | THE BOOK--Playing The Percentages In Baseball
Tango has my favorite look at the Ryan Howard extension. My favorite part is at the end, where looks at the distribution of possible values rather than just point estimates of what Howard's salary should be worth. He finds that there's roughly a 50/50 chance that the Phillies break even with the deal....compared to a 50% chance that it's a disaster. Essentially, the contract is paying for a best-case projection for Howard.  Good job by Howard and his agent....but I think this deal is likely to hurt Philadelphia pretty badly by its end...


Baseball Prospectus | Ahead in the Count: The Source of the AL's Superiority
Outstanding article by Matt Swartz investigating the causes of the American League surperiority. His conclusion surprised me: it's not about free agent spending. The biggest cause seems to instead be spending by AL teams (and especially AL East & AL Central teams) on the draft and on international free agents. The effect is huge: if they played in the same league, you'd expect the typical AL team to win 6-8 games more than the typical NL team (here's some of my own work on this same issue).

This is one of the reasons that I've been most happy to see the Reds amping up their draft/latino signing budget over the last several years, and is something that they will have to continue to expand if they want to get competitive.

Baseball Prospectus | Baseball Therapy: Why Are Games So Long?

Pizza Cutter ends his time at Baseball Prospectus with a regression-based look at why games are so long. He finds that the single largest predictor of game length is....number of pitches thrown! And that's after removing the confound of extra innings by including "number of between-inning breaks" in his model. Number of pitches explained 80+% of the variance in game length by itself. Adding in number of mid-inning pitching changes and the number of throws to first added another 5%. Other variables that were significant accounted for little additional explanation.

Unfortunately, number of pitches thrown gets to the heart of what makes baseball, I don't think we're going to solve that one. We might shave some minutes here and there by instituting other rules, but the 3.5-4 hour games may not be a solve-able problem. And yes, I'm assuming that Morgan Ensberg is right that we'll never get rid of advertisement breaks between innings that slow things down.


Baseball Prospectus | Ahead in the Count: The No Turnover Standings

This is an older article, but I missed it. Matt takes a look at how teams might have shaken down last season if all of the players had stayed with them that drafted or originally signed them rather than being traded, released, or otherwise transferred to other teams. The Reds come out 64-98, second-worst in baseball. This is a big part of why the Reds are where they are: poor drafting. I think that has largely turned around as of 2004 or so (the Homer Bailey pick), but there's a lengthly legacy effect that we'll be dealing with for several more years down the road. The good thing about this is that if we're mad about the team, we can still blame Jim Bowden!


How the top-100 position players in baseball are distributed among the 30 teams. Based on the... - Beyond the Box Score

Great graphical depiction by Peter. Yankees have four of the top-20 hitting talents in baseball. Reds have two in the top 100.


Updated Projection for hitters -

Sean comes to my rescue and posts updated CHONE hitter projections for 2010. I love that updated in-season projections are now the standard practice. Sal Baxamusa gave us in-season Marcels (do it yourself, anyway) in 2008, last year we had in-season ZiPS, and now this season we're getting in-season CHONE (at least for hitters) and in-season Oliver. Why do we need in-season projections? They are our best objective estimates of the true talent of a player, i.e. what we can expect from them moving forward.