Catching up with the Astros’ Run Expectancy Leaders

If you have followed my previous articles, you probably know that I am partial to the statistic, RE24 (Run Expectancy 24). Since I have periodically followed the RE24 results for Astros players during the season, this is an opportune time to point out the Astros’ leading RE24 batters and relievers.

RE24 has been described as a middle ground between sabermetric stats and “traditional” stats because it takes into account both advanced stats as well as productive outs and strike outs and double plays. The advantage of RE24 is that it is a more accurate measure of the actual run value of a player’s performance. In my view, RE24 is a measure of production which is superior to the traditional RBI (runs batted in) statistic. That’s because RE24 encompasses more of the skills that the player employed to increase the probability of scoring a run.

This method measures the change in run expectancy based on the 24 base-out states. (The 24 base-out situations refer to the combination of runners on base and outs—for example, bases loaded with 2 outs is one of the 24 base out states, and has an average run expectancy of 0.736 runs scored.) The player’s actual change in run expectancy is compared to league average run expectancy for each of the base-out states. Unlike the normal advanced offense stats (linear weights measures) such as wRC+ or wOBA, RE24 takes into account the game situation and includes skills like advancing the runner and avoiding double plays.

RE24 Batting Leaders

RE24 is shown on Fangraphs’ leaderboard for win probability stats. The runs above average for the best Astros’ batters is shown below.

Fangraphs Win Probability Leaderboard

If you’ve been following them closely during the season, no surprises here. Yordan Alvarez is a big man, and he is head and shoulders above the other Astros’ batters. In fact, this is the exact same order for the top five Astros’ RE24 leaders which I published on May 31. Comparing their RE24 ranking to the more commonly used linear weights method, the most significant difference is that Altuve is No. 2 or 3 based on wRC+ or wRAA, but falls to No. 5 based on RE24.

Yordan’s 53 runs above average is No. 6 among the MLB RE24 leaders. Tucker’s, Bregman’s, McCormick’s, and Altuve’s ranks on the MLB leaderboard: 12, 19, 23, and 33.

The big dropoff in RE24 results occurs at No. 6 on the Astros’ list, Yainer Diaz, who is in low single digits, whereas the Astros ranking higher are all in double digit RE24 (i.e., above 20 runs above average).

Linear weight stats such as RE24 and wOBA are based on the league average run contribution of singles, doubles, triples, home runs, and bases on balls, without the context of the situation. Those linear weights stats are converted to wRAA (runs above average) based on generic estimates of the run values for those events. But RE24 is a more accurate estimate of the actual run values for the batting events because it reflects the run expectancy at the time.

By comparing wRAA to RE24, we can infer how much linear weights offensive measures over- or understate the actual run values for a player. This is due to both the base-out timing of run expectancy and the inclusion of events like productive outs, advancing runners, double plays, and strike outs in the RE24 measurement.

I have computed the percentage of “extra offensive contribution” based on RE24 percentage above/below wRAA:

Extra Run Percent Above wRAA

Alvarez +32% / Tucker +30% / Bregman +40% / McCormick +52% / Altuve -12% /

Diaz -70%

Interestingly, Alvarez, Tucker, Bregman and McCormick all have significantly higher run value than the linear weights estimates would indicate. However, for Altuve and Diaz, the linear weights estimates significantly exceed the actual run value based on RE24.

Without more analysis, all I can do is speculate on the factors that cause Altuve and Diaz to produce negative extra offensive contribution based on RE24. My guess is that plate discipline has something to do with it. The BB/K metric is a rough measure of plate discipline. Yainer Diaz has the worst BB/K rato on the team, and Altuve is ranked No. 6 on the team for this measure. The GDP measure (double plays per opportunity above/below average) may provide additional insight. Altuve and Diaz are ranked 14th and 15th (out of 17 roster spots) in avoiding GIDPs, based on their negative GDP value. Chas McCormick was the best Astros player at avoiding GIDPs. Alex Bregman was the worst at avoiding GIDPs, yet he produced significant extra offensive value on RE24—perhaps due to his top ranking in BB/K.

Relief Pitchers

RE24 can also be used to value pitchers. However, in my view, RE24 is more useful as a measure of relief pitching than starting pitching. Starting pitcher value is less “situational” than for relief pitchers. Relief pitchers frequently enter the game to deal with specific Base-Out situations. Unlike ERA, RE24 is not cluttered by inherited runner issues. And RE24 is superior to WHIP in measuring how the relief pitcher responded to the situatons he face.

The Astros’ relief pitcher rankings by RE24 are shown below.

Fangraphs leaderboard

Neris and Abreu were best, by far, in dealing with tough base-out situations. Stanek, Graveman, Pressly, and Martinez were all positive forces on a situational basis. The remainder of the bullpen was negative on a situational basis. RE24 confirms something we already knew about Bryan Abreu—he may be the Astros’ best relief pitcher.

Since it is possible, if not likely, that Neris will leave the Astros as a free agent, this may indicate that the Astros have a rather large hole to fill in the bullpen. Losing the No. 1 RE24 reliever is problematic. But RE24 only tells us what happened this season, and does not necessarily address likely pitching results next year. In particular, RE24 may not predict the liklihood of future regression. And advanced pitching stats point toward a significant liklihood of regression by Neris in the future. So, re-signing Neris may not be the answer.

But this would indicate that the Astros need to replace Neris’ strong situational performance with some combination of free agent acquistion and/or accelerated improvement in situational performance by other relievers such as Montero or Maton.

Also, seasonal RE24 performance can be strongly affected by just a few very bad relief appearances. This explains why Montero is ranked last in RE24 among Astros relievers, with a high -10 run value. For example, if the time frame is limited to August forward, Montero exhibits a positive RE24 value. (Note that Pressly’s run value is negative during that same time period.)

I hope that this review of 2023 RE24 performance has helped your understanding of how player performance relates to run expectancy.