The Royals Were Historically Clutch (Obviously)
The Royals — the 2015 MLB champion Royals — were historically clutch. If you paid even any attention at all, you probably don’t learn anything from that sentence. It’s so obviously true it might as well be left unsaid, like “that pizza was delicious” or “I wish we didn’t get lost.” Being clutch became the Royals’ whole thing, all the way through erasing a ninth-inning deficit in the last game they played. But, look: I’ve got data. I’m going to give it to you. You can’t stop me from doing this. By the time you’re reading this post, it’s already done.
Somewhere around the middle of September, I wrote a post for JABO talking about how the Cardinals had been historically clutch. For the Cardinals, the strength had been run prevention, above and beyond what one would’ve expected just from their already strong overall performance. Right here, I’m going to follow a similar template, since now that everything’s over, it feels worthwhile to consider the final numbers. The Cardinals, at one point, looked like the clutchiest team in the league. They ultimately surrendered their lead. (note: pretty unclutch of them) (second note: typical Royals)
At the center of this is our Clutch statistic. It’s not the easiest thing to explain in a sentence or two, but it works as a measure of good timing. Players and teams that do better than usual at important times will have strong Clutch scores. Players and teams that do worse than usual at important times will have weak Clutch scores. You can go to the leaderboards and look at player Clutch, and you can also look at team batting Clutch and team pitching Clutch, which you can further break down to starting Clutch and relieving Clutch.
In the table that follows, the top 10 overall most clutch teams since 1974, which is as far back as we provide information. I’m showing just the top 10 by overall Clutch, but just to provide more information, I’ve included the component Clutch scores so you can see how the various teams compare. In case this needs to be said, and making everything all the sillier: these are regular-season numbers, only. We don’t have team playoff Clutch. So as good as the Royals look here, this doesn’t even take into consideration their miracle title run. Onward:
| Team | Season | Batting | Starting | Relieving | Total |
| Angels | 2008 | 7.3 | 2.8 | 4.3 | 14.5 |
| Athletics | 2006 | 4.5 | 3.9 | 4.4 | 12.7 |
| Giants | 2012 | 5.2 | 2.4 | 5.0 | 12.6 |
| Royals | 2015 | 6.9 | 1.0 | 3.3 | 11.2 |
| Cardinals | 2015 | 0.4 | 4.7 | 6.0 | 11.1 |
| Diamondbacks | 2007 | 2.8 | 3.2 | 5.0 | 11.1 |
| Reds | 2004 | 6.3 | 3.1 | 1.5 | 10.9 |
| Orioles | 1977 | 6.2 | 4.7 | -0.2 | 10.7 |
| Padres | 1989 | 3.5 | -0.1 | 6.7 | 10.1 |
| White Sox | 1990 | 8.1 | -0.3 | 2.1 | 9.9 |
This covers 42 years of baseball. Most of them didn’t even include a work stoppage. This year’s Royals show up in fourth place, beating this year’s Cardinals by the narrowest of possible margins. The clutch strength, as seen here, was the Royals’ high-leverage hitting, but they did well in all categories. The Cardinals were a different kind of team — all their over-performance took place on the mound. The Royals spread it everywhere, even to the opposing defenders when October rolled around.
The first-place 2008 Angels didn’t win the World Series. The second-place 2006 Athletics didn’t win the World Series. The 2012 Giants did, so they might be the Royals’ main competition for being the most overall clutch team we have on record, but while those Giants might have the regular season in their favor, the Royals have what they just did in the playoffs, to tip the scales. The Giants trailed just one time in sweeping the World Series, and that was by one run in Game 4. On the other hand, those Giants erased a 2-0 NLDS deficit. They erased a 3-1 NLCS deficit. They had a good run. So did the Royals. No point in trying to see whose run was more good.
As a fan, this is how you want to see a clutch season end. The Royals’ success in the playoffs was just an extension of their success in the summer, so the fans will forever get to believe that these Royals had another gear. The fans will forever get to believe that these Royals were unique, which is what you always want to believe about your team, and since the numbers are all locked in there’s no changing this. There’s no slow-motion replay review to determine that the Royals slid into the championship and lost contact with it for a split-second. Cardinals fans don’t much care anymore that the team was clutch during the season, because it abandoned them when it mattered. It didn’t hold up. For the Royals, it held up, and it held up the whole way, until there was no time left for regression to the mean. It can be said, now: the Royals were an incredibly clutch baseball team. You can’t say they weren’t.
You can easily say they won’t be again. At least, you can easily argue it. I averaged the top 25 teams in all four Clutch categories, then I averaged those same teams the next season. Results:
Top 25, batting Clutch
- Year X: 6.4
- Year X+1: 1.7
Top 25, starting Clutch
- Year X: 4.2
- Year X+1: 0.6
Top 25, relieving Clutch
- Year X: 5.5
- Year X+1: 0.0
Top 25, overall Clutch
- Year X: 9.9
- Year X+1: -0.2
I guess you could say there’s a little evidence the hitting Clutch keeps up somewhat, but it’s limited. And the other stuff disappears. The top 25 most overall Clutch teams were league-average the following season, which very strongly suggests this is an ephemeral thing. In probably just a few weeks, we’ll get our first 2016 team projections, and the Royals will be projected to go something like 83-79 or whatever, and the argument will be “they won’t be that clutch again.” The argument will have statistical support. I doubt our pages will list the 2016 Royals as AL Central favorites. So goes the baseball math.
But every team is different, right? Every player is different, and every team is different, and maybe, just maybe, there’s something special about the Royals. Even if it turns out there isn’t, down the road, there was in 2015, or at least that’s what the records will show. In some ways, clutch performance is like a race against time — you want to win it all before the clutch runs out. Fans who want to believe in this stuff can never feel truly secure in their position until the players are hugging in the infield. The Royals players all hugged. The last word of the story’s been written.
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.
Of course, if you add up the Clutch of all the Royals pitchers, you get a total of -0.4 (or -2.1 if you leave out Jeremy Guthrie)
http://www.fangraphs.com/leaders.aspx?pos=all&stats=pit&lg=all&qual=0&type=3&season=2015&month=0&season1=2015&ind=0&team=7&rost=0&age=0&filter=&players=0&sort=13,d
So how did the Royals pitchers get a positive 4.3 Clutch in total? Simple – the best pitchers on that staff (the pen) got used in the highest leverage situations. That’s a component of how team-level Clutch is more than the sum of its parts.
As for the 2015 MLB leaders in batting Clutch, it was a repeat title:
http://www.fangraphs.com/leaders.aspx?pos=all&stats=bat&lg=all&qual=0&type=3&season=2014&month=0&season1=2014&ind=0&team=0,ts&rost=0&age=0&filter=&players=0&sort=11,d
There have been different analyses done showing that offensive clutch tends to favor left-handed hitters, high contact/low K guys, and tends to be worst for three-true outcome types. So while the Clutch stat does good job of describing the benefit of timing, nobody should assume that the value itself is 100% attributable to luck.
I believe the clutch stat only measures players against themselves and isn’t really a simple counting stat, so summing up the individual player scores doesn’t get you anything useful.
As much as Jeff cites the clutch stat, I think it’s about time for a more lengthy FGs post explaining what exactly it shows. The glossary isn’t terribly helpful. There seems to be a lot of confusion what this stat really means and how much bullpen strategy and composition can really influence it.
Absolutely true on the clutch stat. I was basically trying to illustrate the point you made – the pitchers’ total “clutch” wasn’t because they individually pitched better in high-leverage situations, just that the better pitchers were actually being used in high-leverage situations
This really shows up when you look at the MLB totals for pitcher clutch. Once teams starting stacking better pitchers in their bullpens in the late 1990’s or so, pitching clutch has always been positive.
http://www.fangraphs.com/leaders.aspx?pos=all&stats=pit&lg=all&qual=0&type=3&season=2015&month=0&season1=2000&ind=0&team=0,ss&rost=0&age=0&filter=&players=0
This comment is not directly related to the clutch question, but I guess this is as good a place as any to put it. Did anyone else wonder about Cain’s steal of second that started the 9th? He was not the tying run, so the only value in stealing in that situation was to prevent a DP. I haven’t run the value comparison, but it seems to me that the risk of getting caught in that situation might outweigh the value of preventing a DP.
The Mets didn’t catch a Royal stealing the whole series and I sure as heck don’t trust d’Arnaud to make a good throw to second. Let alone one that beats Cain. It also puts Cain in a good position to score if Hosmer singles (for the most part). If it were Salvy making the throw I might tend to agree with you but even then Cain is pretty fast. That is my common sense way of explaining it.
Basically it was a “no risk” steal. In most cases, down 2 runs and a runner on 1st with no outs, since you need 2 runs the risk isn’t worth the reward. But the Mets catcher hadn’t thrown out a base stealer all post season, and Cain is an excellent base runner. So in that case, since the ability to take 2nd base was essentially risk free provided he ran early not to put the hitter too far behind, it was worth it to ensure the DP was not a possibility.
The variables are great. “Clutch” might be 2 2-run HR, or merely putting the ball in play on 2 strikes, finding holes, interspersed between errors SB’s, and taking extra bases. The “home run or hitter-K” philosophy needs to be reevaluated. If HRs and batter-Ks are associated with more runs, which are associated with wins, perhaps it only finds significance when analyzed in isolation, compared only to itself. When compared to other philosophies, perhaps “royal-ball” is a superior phipsophy, or constructing your team according to the resources and limitations at hand.
Perhaps better than “clutch” would be “not-choke”. Last night Familia (again), Murphy (again), and Duda choked while the Royals stayed calm and did the job.
How do you propose to define choking? Familia got weak contact and might well have ended the game except for Duda’s poor throw. Is that choking whereas a scorched lineout with Hosmer being doubled off would have been not choking?
Not executing a routine play under extreme pressure. In a high leverage situation pressure is spread around. If we deny the existence of clutch then maybe what we’re dealing with is pitchers and defenders wobbling under pressure. Familia blew three saves, Murphy booted two grounders, Duda should have had Hosmer by ten feet.
I’d say that they are one and the same and we don’t need to deny the existence of clutch, just accept (as one commenter has been consistently saying) that it is a human factor that is unlikely to be measurable.
I put this in another thread but it is more pertinent here: I suspect that it’s largely a function of the confidence boost that is gained by a consistent set of guys, playing in a supportive structure, regularly being put in tight situations and coming up trumps, creating a positive feedback loop whereby guys come up loose whereas their opponents are full of nerves (soccer fans will recognize the German national teams success at penalty shootouts in this). The skill – to the extent that there is one – is on the side of management in keeping the bubble from bursting; confidence being an inherently fragile thing in something as emotional as sport.
Momentum, team spirit, relationship and team management. These are important and unquantifiable things more suitable to psychological than statistical analysis. They are also inherently transient — particularly in a sports context — so it makes sense that we’re not able to pick them up in the data.
I think Familia did about as well as he could in that situation. Sure, three Ks would have been ideal, but that’s a lot to ask. In most situations, Hosmer would not have tried to score, and has the Mets played it differently–with Wright covering third and Flores taking the ball–Hosmer wouldn’t have tried. As it is, Familia sets a record with three blown saves. The irony is that Hosmer’s tying run was not charged to him. He comes in, retires all three batters he faces, doesn’t allow a run of his own, yet blows a save.
He only allowed 1 earned run during the entire series.
Perhaps an alien race is trying to colonize the world and Dayton Moore has been creating human-hybrid baseball players who are immune to the black oil sentient virus.
I’ve been playing around the last few days with trying to figure out the possibility of teams having SUSTAINABLE clutch HITTING. For each year for each team, I take the teams top 10 hitters in terms of PAs, and for these hitters I take their high-leverage and low-leverage ABs and Hs over their entire CAREER, the hypothesis being that certain players have a capacity for clutch hitting, here measured in higher batting averages in HL situations than LL situations. Then I see whether the totals for the entire team is statistically significant in terms of different in BA between HL and LL. The 2014 Giants were significant at the p<.05 level (chi-sq=3.94), because six of their hitters (Pence, Posey, Crawford, Pagan, Morse, and Belt) have AVGs which we at least 0 points higher in HL situations than in LL situations and reasonable number of PAs in their careers, and only Blanco is somewhat lower (and only 17 points). Sandoval is the same, and Panik and Hicks have so few PAs in their careers that their numbers don't influence the total. The implication would be that this kind of clutch hitting would be sustainable as long as the players stay on the team. It would also account for why the Giants as a team had such good clutch hitting in 2012, since these were to a large extent the same players (except for Morse). Of course, even a sustainably clutch team is not going to be clutch every single year, there will be peaks and valleys, so 2012 and 2014 were peaks resulting in WS wins while 2013 was a valley. The 2015 Royals were not statistically significant in this respect, because they had only 3 players with large career differences between HL average and LL average (Hosmer, Perez, and Rios), while Escobar and Gordan have significant minuses.
Sustainable clutch hitting is literally impossible. Being a ‘clutch hitter’ is equivalent to saying that you don’t try unless it’s an important situation. There is no way to be a high level athlete without trying. Also your methodology is heavily flawed. Batting average is inflated in clutch scenarios due to sacrifices and defensive aggression. On top of that, you correlated a sample with itself (i.e. you took players that had a really high clutch score then checked if combined they had a high clutch score).
No, it’s saying that people can be capable of performing better in high-stress situations because these situations produce a chemical reaction which results in enhanced abilities, at the very least adrenaline but maybe other things as well. From an evolutionary standpoint, it would be odd if there WASN’T such a reaction. This chemical reaction can have different results in different people. There are presumably also differential reactions to high-stress situations in terms of panicking or overthinking which could produce anti-clutch behavior, otherwise know as choking. If you seriously think that doing between or worse in some situations is simply a matter of trying or not trying, it sounds like you’ve never participated in competitive sports yourself, because anyone who has participated in competitive sports knows this is reidiculous. Batting average in HL situations is identical to batting average in LL situations. If you look at the data in Fangraphs you’ll see that between 2006 and 2015 there is a difference of 2/10 of one point per year between HL average and LL average, and there is never more than a 3 point difference in any one year, so your claim that batting average is inflated in clutch scenarios is wrong. And you seem to have misunderstood my methodology–it’s based upon the CAREER statistics of ALL of the 10 players on a given team with the most PAs, not just players with a high clutch score.
The Royals 2015 Clutch statistics were radically affected by the performances of Cain and Infante, who batted 146 and 186 points higher in LL situations than in LL situations, even though previously Cain had been slightly worse in HL situations in his career while Infante had been much lower. The other Royals mostly more or less matched their career records in terms of HLAVG vs. LLAVG–Hosmer and Perez were very good but not unrealistically so (+31 and +29), Morales was pretty good (+21), Gordon was lousy (-40), Zobrist was also (-32, although overall he’s been only -8), and Escobar was below average (-8, although he’s generally been worse, at -27 overall). Moustakas and Rios were a good deal worse than their averages (-131 vs. -6 and -43 vs. +24), but this difference wasn’t nearly as much as the extreme pluses of Cain and Infante. So this really doesn’t look sustainable.