Mets Hitters Couldn’t Be Less Clutch
Let’s face it: As the Mets go, there’s no shortage of things to worry about. The team overall remains in a decent position, but now there’s concern regarding two pitchers’ elbows. Meanwhile, Matt Harvey still doesn’t quite look like himself. David Wright is probably done for the year. And the lineup just isn’t producing runs. Injuries haven’t helped, and Michael Conforto’s collapse didn’t help, but the pitchers are getting so little margin of error. Things in New York are frequently tense. They’re tense today. It feels a little like last season, before the season turned beautiful.
I can’t say anything about Steven Matz. I can’t say anything about Noah Syndergaard. I can’t say much about the various injuries, or about Conforto’s chances of getting it going. I don’t know where the Mets are going to go, and their struggles have helped open the door for the Marlins. What I can say is this: Offensively speaking, the Mets have been impossibly unclutch. It shouldn’t continue like this. Of course, what’s done is done.
Just to provide a quick reminder of the Official Stathead Position on this: Mostly, we don’t believe in sustaining clutch performance, but we do believe in the reality of clutch performances. Obviously, there are important situations. Obviously, those situations end with results. Someone gets the better ones, and someone gets the worse ones. There most definitely have been clutch players and teams, in hindsight. Even if it wasn’t a real “talent,” events that happen can’t be taken back.
All right. We have a Clutch statistic here. Actually, even before I get to that, let me point out that, according to BaseRuns, the Mets should be averaging 4.07 runs scored per game. They’re actually at 3.58, which gives a difference of -0.49. That’s the biggest negative difference in the game, out of the 30 teams, and it works out to about 38 missing runs. And that would work out to about four missing wins. Something has gone horribly wrong. A recent buzzword has been “sequencing.” Call it whatever you like, but the Mets have shot themselves in the foot.
The Clutch statistic! Back to that. Clutch effectively tries to measure the difference between actual WPA and expected WPA. It captures offensive performance and then folds in the context. In April, Mets hitters ranked 29th in Clutch. In May, 22nd. In June, they’re 29th. Overall, they’re dead last, and it’s not very close. The Mets have been the least-clutch offense in the game, and they wouldn’t just stand out in 2016. We have this information stretching back to 1974, and, let me tell you something about their pace.
“Pace” is always misleading, because paces are un-regressed. The Mets have played just short of half a season, which is more than half a season away from being a full season. So, know that going in. But going back to 1974, I calculated every team-season’s Clutch score per 162 games. Then I adjusted those marks for the league average, since the average isn’t set to 0.0. The 10 lowest team ratings as of right now:
| Team | Season | G | Clutch/162 |
| Mets | 2016 | 77 | -10.9 |
| Padres | 1994 | 117 | -9.0 |
| Twins | 1978 | 162 | -8.5 |
| Indians | 2005 | 162 | -7.6 |
| Yankees | 1981 | 107 | -7.4 |
| Dodgers | 1989 | 160 | -7.2 |
| Angels | 1981 | 110 | -7.2 |
| Orioles | 2004 | 162 | -7.2 |
| Expos | 1983 | 163 | -6.9 |
| Giants | 2002 | 162 | -6.9 |
It’s the Mets, then a strike-shortened season, so you can see how sample sizes matter. But at this pace, the Mets would be the least-clutch offense on record by about two full-season wins. No, I can’t imagine they actually finish this low, but this gives you a sense of how they’ve performed. They’ve performed like a horribly unclutch baseball team, and that’s a surefire way to make observers upset. The offense hasn’t been good, but when you mix in bad timing, it makes the lineup seem atrocious. It makes everything seem worse than it is.
This would be a glimmer of good news: Unclutch lineups haven’t stayed unclutch lineups. I looked over the last five years, and I gathered each team’s first- and second-half Clutch numbers. I put each over 2,500 plate appearances just to give a common denominator. If it were easier to do, I would’ve gone with a bigger sample, but I think this is convincing enough:
Zero relationship. That’s almost a perfect random scatter. The Mets shouldn’t remain unclutch, and that’ll make the lineup seem stronger. Just last season, they didn’t have these issues. It seems like it’s just been miserable luck. But as you could imagine, it’s shown up most dramatically with runners in scoring position. Ken Davidoff just touched on this. The Mets have a .704 OPS, but with runners in scoring position, that drops to .607. That’s horrible! Horrible to what would be a historic extent.
Going now to the Play Index, we can look at how teams have done in the past in those situations. My stat of choice is tOPS+, which measures split performance vs. overall performance. By this measure, the worst team in recorded history with runners in scoring position is the 1962 Mets, who finished with a tOPS+ of 77. The 1987 Indians finished with a tOPS+ of 80. As I write this, this year’s Mets have a tOPS+ of 74. Another mark that would be the worst ever, if it were to keep up. Clearly, the points are related. The Mets have been historically unclutch, and it’s mostly because they’ve been historically bad with runners in scoring position. They shouldn’t stay like this going forward, but a lot of damage has already been done.
Thankfully for them, they remain in the race. And no one should be convinced that the Mets are bad. The Mets, on talent, are one of the better teams in the league, and though they’ll probably be without Wright, they could trade for an infielder. Or they could sign Yulieski Gurriel. Travis d’Arnaud should hit more than he has. The Mets basically have what they should need, and it would be absurd to write them off. It’s fine to be disappointed by the offense. The offense has been disappointing. It ought to improve. And then maybe all the concern should be reserved for Syndergaard and Matz. To me, those issues seem more likely to linger.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.
i’m sure losing 80% of games vs the nats the last 3 years and the embedded losing culture of the mets franchise (worst in all major professional sports) hasn’t already crushed all these players already
the nats’ very existence and success is going to kill all my spirit
even when the mets “dominated” the nats the games were almost always competitive. and the nats actually impacted the outcome of the mets’ seasons. just one more reason one is a blessed franchise and the other doomed to never be anything and lose even their most real fans like me
lol the nats are waaaay more cursed in the playoffs than we are. How are they a winning franchise when they had a huge collapse last year and never seem to make it out of the playoffs. I’m a huge mets fan too, but I’m sick of people perpetuating the losing culture myth. This decade has mostly been bad because we’ve mostly been bad. Luckily Sandy is turning that around and the future is bright. So buck up, and lets just enjoy the fact that the mets are goingto be competitive in the recent future without all the doom and gloom whenever anything goes wrong.
hey, but CARLOS BELTRAN tho
the only bad thing that ever happened to the mets
also this pretty normal lolmets snafu is almost too regular to even bother with at this point, but i guess it’s still not as franchise-crushing as that whole natinals thing that i know you guys really had to suffer through
i mean, i broke it down this way
-prospects who have not busted
-trades that haven’t gone against them
-signings that went really bad
-players who they let go who got much better after leaving
it’s pretty ridiculous. now they’ve been absurdly healthy this season, murph is putting up monster numbers, and their old guys are starting to look rejuvenated. guys like ross/roark way overperforming. harper hasn’t been that good since april ended either. it’s annoying
I realize you’re just another troll (and not even a particularly good or interesting one) but I find it hilarious that you can assert that the Mets have the worst “embedded losing culture” (whatever that is) in all major professional sports in a world where (a) the Mets went to the World Series just last year, and (b) the Seattle Mariners, Toronto Maple Leafs, and Cleveland Browns all exist as professional sports teams.
i don’t troll, i state facts
Uh huh.
A few bad decisions – can we call not re-signing Murph a mistake now? not having a legit back-up for Wright and his inevitable DL stint, De Aza, Asdrubal – have failed to make up for Cespedes unexpectedly returning. Couple that with Duda’s injury, d’Arnaud’s injury/collapse, Conforto’s struggles and Walker disappearing after a hot April.
Last year the elixir to an ice-cold offense was to bring in Kelly Johnson, Uribe and Cespedes, coupling with the return of Wright, d’Arnaud and the promotion of Conforto.
This year, Reyes could provide a start, theoretically. Granderson and Conforto or Nimmo would help. Getting Duda and/or d’Arnaud back healthy and hitting too. If they could make a trade for a big-time bat, maybe Wil Myers?
As for the un-clutch, I keep thinking of the Natural – “losing is a disease, as contagious as syphilis”
I disagree with Asdrubal, he has produced fine with some surprise home runs and has been impeccable defensively. Who did you want at short, tejada?
Cab just doesn’t belong in the 2 hole though, he is definitely feeling pressure from the injuries and is trying to do too much at the plate
I thought Cabrera was doing OK but has a negative WAR on the season. I would have brought the infield back intact – Duda, Murph, Flores, Tejada and Wright.
Are you confusing him with someone else? When I click on his player page, he has 0.9 WAR, and is on pace for about 2 WAR this year.
Flores and Tejada, on the other hand, both have negative WAR. But that goes against your idea of bringing both back, and forgetting about Cabrera.
maybe the Mets can pry away Jean Segura from the Dbacks. If they think they are sellers, this would be a good time to sell him, and the Mets need infield bats. Dback management may still think they are in contention though. Mathematically, yes, but realistically, no.
On second thought, never mind. I forgot the Mets signed Reyes, so they’re going in that direction and wouldn’t need Segura.
Some issues that are definitely contributing to the unclutchness are: Lack of speed, lack of ability to hit against shifts, inability to hit just about any curveball, a bunch of double plays and a team wide swing approach that makes zero two-strike adjustments and continues to swing for the fences.
Why would those things affect production in clutch situations more than they would affect production in unclutch situations? “Lack of ability to hit against shifts” in particular is a thing that suggests they should do better with runners on base (where shifts are constrained by the need to cover extra bases) than with the bases empty.
Lack of speed WOULD presumably contribute to the difficulty of scoring with RISP, as do double-plays. I’m just not sure how many of these criticisms are backed up by actual statistics.
anybody who bothered to watch the World Series knows the Mets aren’t clutch.
I’ve never been clear about whether this is a parody account.
Interestingly, though, the Mets current record (40-37) is only 1 win less than their BaseRuns predicted record (41-36). They’ve been unlucky on offense as Jeff notes (3.58 vs the predicted 4.07 per game), but they’ve been fortunate enough to give up fewer runs than BaseRuns predicts (3.53 versus the expected 3.79).
There cluster is awful though they have been lucky on pitching. https://thepowerrank.com/cluster-luck/
I don’t mean this in respect to this article, but as a community, I think Fangraphs writes off the idea that clutch outcomes can be a skill (or any outcome we can’t explain) far too often.
In order to do statistical analyses of At-Bats, you have to make the assumption that the data you’re collecting is uniform — sure we collect tons of contextual data about each AB and its outcome to try to stratify as much as possible — but ultimately, you still assume uniform conditions. Batter has the same approach every AB, pitcher has the same approach every AB, impact from baserunning threats, defensive alignment, etc, are all completely uniform, from the first inning to the ninth, regardless of a teams record, the chances for post-season play or any other factor that my bias their performance. It should just be intuitive that these factors mean ABs are not really that uniform, especially those ABs that we’re talking about when we’re talking about “clutch” hitting, but we still attempt to analyze these ABs to attempt to solve for unexplained variances between actual outcomes and expected outcomes as much as we can…..there is no flaw in that.
The flaw comes when we affirm the null, that no such skill exists, with that weak or potentially biased data. There’s no other practical modeling application where you’d make that jump, so why does it happen here?
WRT this article, I think its more accurate to say that Clutch is a skill, from the analysis in Tango Tiger’s The Book, but its not really an identifiable skill at the player level — so it more likely that maybe the Mets regress somewhat but likely have a banded range where their true Clutch (per the stat) rating sits, somewhere below zero, i.e you can’t put up historically bad Clutch numbers without having a somewhat low baseline Clutch.
Another thing that should be looked at is their baseruning, they don’t go first to third often which in turn makes it that much harder to score that guy on second because odds are if he’s not going first to third on an outfield hit then he’s not going second to home on the same hit
Actually, that’s not true. The length that a throw takes from the outfield to third is much shorter than the length from the outfield to home.
Sometimes the simple explanations actually do help explain things: the Mets rank middle-to-meh in OBP and SLG, but are dead last in the majors in batting average. In order to sequence hits with runners on base, you have to have hits to sequence (and not walks, that is).
Sure, but if you lose 100 points of OPS in a clutch situation, that’s either you being a team of chokers, or a very unusual random fluctuation. So while your argument would explain why a normal to high contact team should be more clutch than the Mets, it can’t come anywhere close to explaining this huge gap in success between low and high leverage situations. More walks in high leverage situations would be helpful to the Mets, though not as helpful as hits, as you point out.
Yes, I’m not trying to dispute the article’s basic point at all. Just want to suggest that a low-batting-average team (and a roster constructed with an overall deemphasis of batting average) might be a factor in how they got here.
To understand what’s going on with Clutch (and any WPA-related state), you have to do a lot of WPA calculations yourself to get a feel for how this works. From my own experience, a few things are clear:
(1) Teams with a lot of strikeouts are going to have a lower Clutch score than teams with the same batting average but fewer strikeouts. Even aside from the possibility of errors and fielder’s choices in which no one is retired, strikeouts in high-leverage situations will produce a huge minus whereas outs which advance a runner–especially but not only if this scores a run–will generally be about equal in terms of WPA.
(2) There are a number of situations in which there is a huge difference in WPA between walks and singles, specifically when there are runners in scoring position. This effect is enhanced if there is good team speed, so that runners can take an extra base. This will naturally result in a higher Clutch score for teams who hit a lot of singles and walk relatively little (the Royals recently being an obvious example), because situations which runners in scoring position will be higher-leverage than situations without runners in scoring position.
(3) As a result, teams with a lot of strikeouts, home runs, and walks but a relatively low batting average will do relatively poorly in terms of WPA and Clutch–the high number of walks will mean that they have a relatively large number of situations which are high-leverage, but in such situations the strikeouts and lack of singles will hurt WPA and Clutch a lot, the walks won’t help much, and although there will be a good number of multi-run homers, they won’t make up for the big minuses on outs.
(4) HOWEVER, individual home run hitters with low batting averages and a lot of strikeouts can nevertheless do relatively well in terms of WPA if they come up a lot with no one on base, especially if it’s with 2 outs, because (assuming that the score is close), they can get reasonably big plusses for their home runs, while when they get out the minuses are very small (unless it’s very late in a game which is tied or their team is behind by one run). This will not, however, help their Clutch score because these homers will come in low leverage situations. The Mets four big home run hitters (Cespedes, Walker, Granderson, and Conforto, who have hit 56 homers between them) are #1, #3, #4, and #13 in the majors in terms of the lowest pLI (minimum 220 PA), that is, coming to the plate in the least important situations, presumably because the Mets just don’t get many people on base, and have a combined Clutch score of -2.31, but their combined WPA is +1.35. Since the actual purpose of Clutch is to get the best WPA possible with the littlest general offensive production, this means that hitters with relatively bad Clutch scores can nevertheless make a positive contribution to their teams chances of winning provided that they don’t bat too much in high-leverage situations. The problem is that teams tend to bat home run hitters with low average in the order so as to maximize the possibility of a multi-run homer instead of putting them in the bottom of the lineup where their homers will help less but their outs will hardly hurt at all.
If you compare the top 5 Clutch hitting teams (Rangers, Reds, Giants, Phillies, and Yankees) with the 5 worst Clutch hitting teams (Mets, Astros, Cubs, Twins, and Rays), the top Clutch teams have a higher average ranking in (few) strikeouts (11th place vs. 26th place) and (less strongly) batting average (18th place vs. 23rd place), while the worst Clutch teams are better in walks (9th vs. 21st) and HRs (11th vs. 21st). The most striking correlation is between lots of Ks and bad Clutch–of the 9 worst clutch teams, 8 are 20th or worse in Ks (and the only exception is 16th), while of the 7 teams with the least Ks, 6 are in the top 8 in Clutch. A lot the exceptional data is from the Reds, Phillies, and Dodgers, who have terrible averages (27th, 28th, 29th) and aren’t good in Ks (13th, 19th, 24th) but somehow are pretty good at Clutch (2nd, 4th, and 6th)–in the case of the Phillies, this is presumably because they’re even worse in walks and HRs (29th and 26th).
This is good, and interesting, stuff. Although strikeouts generally advance no one and aren’t “productive” outs like sacrifices and fielder’s choices can be, that is probably somewhat offset by there being no possibility of grounding or lining into a double play.
It does offset it but to a relatively small extent. It only becomes a statistical factor if the there are ground ball relievers who are specifically brought in in dp situations (jeremy affeldt was like that). Also, teams with few strikeouts can avoid this by starting runners a lot. The giants even have a name for this, it’s called the ‘ground attack’, a runner gets on, the next few batters hit a lot of ground balls and almost never strike out, you do hitandruns with every batter.
I am not sure Mets are one of the better teams on talent alone. According to the article they should avg 4.07 rpg. Of course 38 less runs is a big difference, but it still has them 26th in rpg. Their talent happens to be all in one place, which is the rotation. They are talented in perhaps the most important area, but I just don’t think there is enough beyond the rotation. When I think of talented teams I think of teams that spread out their eggs into a few baskets, so they can survive a lost egg here or there. Mets have 1 basket with a ton of eggs and then a couple empty baskets. They can’t survive a pitching loss because even a non unlucky Mets O leaves no room for error, and they can’t survive an offensive slump since this is a bad O anyway you slice it.
There were more eggs in that basket. They just broke.