PITTSBURGH, PA - JULY 20: Joel Hanrahan #52 of the Pittsburgh Pirates celebrates with teammates after a 4-3 win over the Miami Marlins on July 20, 2012 at PNC Park in Pittsburgh, Pennsylvania. (Photo by Joe Sargent/Getty Images)
In last night's 4-3 victory over the Miami Marlins, Pirates' relief pitchers preserved a one run lead for four innings. Over the course of those four innings they stranded five baserunners. Twice Pirates' relievers extracted themselves from situations in which the Marlins run expectancy was very high. (Run expectancy is the average number of runs we would expect an average team to score in a given situation.)
In the seventh, Jared Hughes relieved Chris Resop with runners on first and second, one out. At that point the Miami's run expectancy was .85, and their probability of scoring at least one run was about 43 percent. The Marlins did not score. In the ninth, Miami had runners on first and third with one out. Their run expectancy was 1.13, and their probability of scoring at least one run was about 65 percent. Again the Marlins did not score.
Last night's performance by the Pirates' bullpen was not unique. It has been happening all year. While the offensive turnaround in the last month and half has been receiving deserved attention, it's been the bullpen's ability to shutdown innings and leave runners on base that explains much of the reason why the Pirates have a winning record. Indeed, to date, the Pirates bullpen has been shutting down scoring opportunities and stranding runners at a historically high rate.
Over the course of the next week I am going to look at performance of the Pirates bullpen. In particular I'm going to look at the significant contribution that Pirates' relievers have made to the team's winning record. By way of brief preview, here is some of what's ahead: as of July 21st, when one looks only at context sensitive statistics, the Pirates bullpen provides almost exactly the difference between the team having a .500 record and their current record of 52-40. (See below for an explanation of the difference between context-sensitive statistics and context neutral statistics.)
In this post I am going to focus on the Pirates' bullpen's effectiveness in shutting down scoring threats.
We will see that Pirates' relievers are on pace to post the fourth best ratio between shutdowns/meltdowns, and the second most shutdowns per game, since 1974.
What are Shutdowns/Meltdowns?
In May of 2010 Fangraphs.com introduced two new pitching statistics as substitutes for the flawed "Saves" and "Holds" metrics. The result was "Shutdowns" (SD) and "Meltdowns" (MD). SD and MD are based off of WPA (Win Probability Added), which I describe in detail here. Basically, WPA calculates how much each event in a baseball game adds or subtracts to the probability of winning the game. It then credits the change in probability to the players involved. Here is an example how WPA works taken directly from Fangraphs.com:
"In game 4 of the 2007 World Series, the Win Expectancy for the Rockies started out at 50%. When Jacoby Ellsbury doubled off Aaron Cook in the very first at-bat in the game, the Rockies WE declined to 44.2%. The difference or WPA was .058 wins (5.8%). Ellsbury was credited +.058 wins and Aaron Cook credited with -.058 wins."
The Shutdown and Meltdown metrics utilizes WPA in the following way: (again copied from Fangraphs)
"A Shutdown is when a reliever accumulates greater than or equal to 0.06 WPA in any individual game.
A Meltdown is when a reliever's WPA is less than or equal to -0.06 in any individual game."
In other words, a shutdown is defined as when a reliever adds six percent to the probability of his team winning. A meltdown is defined as a reliever costing his team a six percent chance of winning.
Since 1974 (first year for which SD and MD are calculated) the median ratio of shutdown/meltdown is 1.45 ('08 Rockies). The average ratio is 1.49. The worst full-season ratio belongs to the 1975 Angels (.583), and the best was posted by 1990 A's (3.22). (As I'm sure many of you will recall, in 1990 Oakland's closer Dennis Eckersley pitched 73.1 innings, had a .61 ERA, with only four walks and 73 strikeouts.)
As of July 21, the Pirates have a SD/MD ratio of 2.88. They have 92 shutdowns to 32 meltdowns. At their current pace, the Pirates would end the season ranked fourth best out of the 1078 bullpens for which we have data. Here are the top ten:
The individual ratios for the Pirates' relievers are:
In terms of league leaders, Joel Hanrahan has the second most shutdowns.
Current ratios for the National League
Shutdown Per Game
In Friday night’s game, three of the four Pirates relievers (Grilli, Hanrahan, and Hughes) recorded shutdown innings. This is a pattern we have witnessed all season long. Indeed, the 2012 Pirates are on pace to finish with the second most shutdowns per game since 1974. Moreover, this isn’t an isolated one-year phenomenon. They are continuing the pace set by the 2011 Pirates bullpen, which averaged the 7th most shutdown innings per game since the statistic has been calculated.
If the Pirates continue their current pace, they will be a historically top ten bullpen when it comes to both the number of shutdowns per game, and the ratio between shutdowns to meltdowns.
That's it for now, be sure to check back later this week as I continue to examine the extraordinary performance of the Pirates' bullpen this season.
Note on Contextual Statistics or Context-Sensitive Statistics: Rather than valuing all events equally, contextual statistics weigh events by their significance within the context of a game. The most important context-sensitive statistics are the Win Probability Added (WPA) family of metrics. Think about it this way: context-neutral statistics like batting average treat all singles the same. A batter gets credited with the same amount of value regardless of whether a player gets a hit with bases loaded in the 9th inning of a tie game, or the 4th inning of 12-0 game. Similarly, when calculating slugging percentage, a home run counts as four regardless of the situation. Contextual statistics, on the other hand, adjust the value of events based on their significance within the game. So, for example, since a walk with bases loaded in the 9th inning of a tie game is obviously more likely to affect the ultimate outcome of game than a double in the 9th inning of 12-0 game is, the walk is has a higher WPA score.