At this point, the majority of informed fans can accept that RBIs are not a good stat to use for comparing players. Judging a player based on a counting stat that is inherently dependent on the team around him is a flawed method for determining offensive prowess. In general, using a stat like wOBA (Weighted On-Base Average), wRC+ (Weighted Runs Created), or simply the Off score displayed on Fangraphs are better methods for judging how good a hitter actually is.
However, the ephemeral "clutch" label will still get thrown around, often with no way to judge the veracity of its assignment. If RBIs are an awful stat, how can we judge a player’s performance in run-producing situations (apart looking at numbers with men in scoring position, an often dubious practice due to sample size)? What we want is a context-dependent stat—as opposed the context-neutral wOBA and wRC+—that will not favor players who simply get more opportunities thanks to the players around them.
Enter RE24. This stat is based on the run expectancy of the 24 base-out states (e.g., no outs and nobody on, no outs and a runner on first, etc.) and judges the player on how he performs relative to league average in these situations. To explain this, consider the following example.
A player comes up to the plate with a runner on second base and nobody out. The run expectancy in this situation is 1.068, meaning that on average, 1.068 runs will be scored this inning. Now, let’s say the hitter singles and the run scores, so now there is a batter on first with nobody out. The new run expectancy is 0.831, but one run has scored already. So we do a simple calculation:
Final Run Expectancy – Initial Run Expectancy + Runs Scored = RE24
For this example, the batter has an RE24 of 1.237. The beauty of this stat is that it can also be negative for a given situation (for instance, if the batter popped out harmlessly). In this way, a hitter can lose RE24 if he doesn’t perform "in the clutch," whereas a batter can never lose an RBI.
Is it a perfect stat? Hardly. We should prefer context-neutral stats in almost all cases. However, RE24 offers the most accurate tool for distinguishing between hitters who have performed well "in the clutch" from those who have not, and it makes for an interesting lens through which to view the Mets. The table below presents the RE24 numbers for the Mets’ hitters in 2014:
|Matt den Dekker||-4.11||49||-0.084|
In case you missed it, Lucas Duda is awesome now. The Dude leads the Mets in both total RE24 and RE24/PA. Interestingly enough, Kirk Nieuwenhuis is close behind him in terms of pace, but 67 plate appearances is hardly a large enough sample to suggest that Nieuwenhuis is a truly awesome clutch hitter. Daniel Murphy, David Wright, and Curtis Granderson predictably round out the top four in terms of total RE24, but Granderson slips down the list quite a bit when normalizing for plate appearances.
On the other end of the spectrum, we can reaffirm that Chris Young and Ruben Tejeda are indeed god awful with the bat. Both are nestled amongst the roster dregs that have very few plate appearances. The fact that Young is still on this team continues to boggles the mind.
The struggles of Travis d’Arnaud early in the season definitely hurt him here, but he actually ranks second on the Mets over the last 30 days with 6.54 RE24. He’s returned to the majors a revitalized hitter, and looks the part of the mashing catching prospect we got in return for R.A. Dickey.
As a side note, let me point out that David Wright ranks ninth with 267.57 RE24 since 2006. Anyone who says he’s not a clutch hitter is flat-out wrong. It also goes to show how badly this season is going for him.
Just for a little fun, let’s look at the RE24 of our pitchers. All of the sample sizes are too small to mean anything here, but there are some laughs to be had:
Unsurprisingly, Bartolo Colon is by far the worst hitter by RE24, and probably any other metric you look at (unless there’s a metric for the number of times a batter losses his helmet). Conversely, Jeurys Familia, proud owner of a 1.000 batting average, is by far the Mets’ most clutch-hitting pitcher. We might have expected to see Jacob deGrom fare a bit better than this, but he still has a higher RE24 total than Ruben Tejeda, so maybe the Mets should play deGrom at shortstop on days he doesn’t pitch.
On a more serious note, it’s surprising to see Wheeler and Niese hit so poorly. Both at least look like they know what they’re doing at the plate, and Niese has historically been a decent-hitting pitcher. Not that this is what they’re paid to do, but maybe one or both can turn it around at some point.
Ultimately, there aren’t many shocking conclusions to draw from this data. Duda is awesome, Wright is having a down year, d’Arnaud has been great since his callup, and Young and Tejeda should not be starting players. At the very least, the next time an uninformed fan tells you David Wright is not clutch, you can throw this stat back in their face to prove them wrong.