One of the most interesting developments in the evolution of baseball statistics is the increasingly precise quantification of things that were previously described only in qualitative terms. For example, Win Probability Added (WPA) gets us as close as we've ever been to putting a numeric value on exactly how "big" a home run was in the context of a single game. WPI/LI or Situational Wins is used to measure how well a player does the "little things" (my study of the Pirates and "little things" here). Finally, Leverage Index quantifies the relative "pressure" of every game situation. It is this last measure that I am interested in this study.
I became interested in studying how the Pirates have performed under pressure after last Monday night's extra inning loss to the Cincinnati Reds. According to Fangraphs.com measure of average game pressure (average Leverage Index or aLI), Monday's game was the most tension-filled of the season (aLI 2.03). The Pirates dramatic loss made me curious about the extent to which a lack of clutch hitting and pitching have played a role in their current slide (11-26 over their last 37 games).
The concept of "clutch" gets something of a bad rap in the baseball research community. This is largely because there is little evidence that players possess an ability to consistently perform above their average talent level in high-pressure situations. In other words, "clutchness" does not appear to be a repeatable skill and, therefore, is not very useful in making predictive evaluations of players.
However, studying performance in high leverage situations can be extremely useful in describing a player's value in a single season and explaining a team's wins and losses. This is largely because high pressure moments occur more often than we are probably aware. Across all of Major League Baseball, almost one-in-five plate appearances (19.2 percent) have been in high-leverage situations this year (Leverage Index of 1.5 or higher). Pirates hitters and pitchers have faced a slightly higher than league average rate of high leverage moments, 19.8 percent and 19.7 percent respectively.
At first glance, it is appears obvious that a lack of clutch hitting and pitching have played an important role in the Pirates free fall. As I noted in a post last week, since August 9th the Pirates have significantly underperformed their Pythagorean expectation. Moreover, during the same time period they have gone 1-9 in games decided by one-run, and 1-14 in games decided by two runs or less. Using more advanced baseball metrics we can dig a little deeper into this issue and add an additional way of looking at the role that "cluchiness" has played in the Pirates current slide .
High Average Leverage Games (i.e. tense or pressure-filled games)
The first way I looked at how the Pirates' have performed under pressure was to examine their record in the 20 highest aLI (average Leverage Index) games they've played. In other words, these are the 20 games that contain the most crucial situations, consistent tension and pressure moments. These are games in which the outcome swung in the balance throughout. I separated these games into three categories based on the high and low points of the Pirates season: April 5 - May 24, record 20-24; May 25 - August 8, record 43-23; August 9-September 17, record 11-25.
For the season, the Pirates have gone 12-8 in their 20 most high pressure games. Six of these occurred between April 5-May 24, and the Pirates record was 5-1. Eight occurred between May 25-August 9, and their record was 6-2. Six have occurred since August 9, and the Pirates record is 1-5. The Pirates have done poorly in high pressure games since August 9.
High Leverage Offensive Games
Next I looked at all games in which the aLI was "high" for the offense. (Each game begins with a Leverage Index of 1.0; High Leverage is a value over 1.5; Medium leverage is 0.7 to 1.5; Low leverage is less than 0.7). In other words, these are games in which the Pirates batters faced the highest average pressure.
For the season, the Pirates have played 20 games in which the average offensive leverage was over 1.5. Their record in these games is 9-11, and their offensive slashes are .227/.306/.364 .670 OPS.
In the table below I have separated these 20 games by the high and low points of the season mentioned above and added up their WPA (Win Probability Added) and RE24 (Run Expectancy) statistics.
WPA tells us how many games above or below average the Pirates won/loss due to their offensive performance in these high tension offensive games. It is the perfect statistic to use for this type of analysis because it is context sensitive. In other words, WPA weighs events by how much they contribute to the winning or losing of a game. Unlike context neutral statistics like batting average, which counts every hit the same regardless of the game situation, WPA weighs the values of hits relative to the context in which they occur. Here is an example how WPA works taken directly from Fangraphs.com.
"In game 4 of the 2007 World Series, the WE 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."
RE24 gives us an idea of the number of runs gained and lost. As defined by Fangraphs.com, "RE24 is the difference in run expectancy (RE) between the start of the play and the end of the play...(it) tells you how many runs a player contributed to his team. It's similar to WPA (except in runs), but unlike WPA it does not take into account the inning or score of the game. Therefore, it is a more context neutral statistic. It does however take into account how many runners are on base and how many outs are left in the inning."
Even though the Pirates offense performed horribly in the first 44 games of the season, they did get more "clutch" hits in games in which their batters faced high pressure. Timely hitting actually gained the team one additional win over average. Conversely, since August 9th, the Pirates have lost 7 of the 8 games in which their hitters faced high average pressure. Their lack of offense production in these games has cost the team about two and half wins.
High Leverage Pitching Game
Finally, I looked at games in which the average pressure on the Pirates’ pitchers was high. Again there are 20 games in which the average Leverage Index for the pitchers has been 1.5 or above. Their record is 16-4 in these games and they have posted a 2.78 ERA.
Breaking down the season by time period and WPA/RE24 yields the following results:
Pirates’ pitchers have consistently performed well in games in which the pressure has been on them. This is particularly true early in the season, where their performance in high pressure games gained the team 2.5 wins and saved 12.35 runs. Between 8/9-9/17, the pitchers have continued to perform well in high pressure games, but the Pirates have lost two of these games.
So What Have We Learned
Since August 9th, the Pirates have done noticeably worse in games that have had high average pressure. Perhaps more accurately, they have regressed from the remarkable pace they established through earlier in the season.
Moreover, the data shows us that the Pirates kept their head above water during the offensive catastrophe of April and May in large part because they got timely hitting in games that were particularly close and tense throughout.
In the coming days I will post the results of an analysis that looks at how the Pirates have performed in high leverage situations, not just high average leverage games. That is, I will look at how the team has fared in each and every high leverage situation they have faced throughout the season, and look to see if there has been a performance drop-off in the last two months. Included in that update, or one to follow, I will also present the data of how each player has performed in high leverage situations during the current 37 slide.