by Keith Law
In The Book, a 2007 tome from Tom Tango, Mitchel Lichtman, and Andrew Dolphin, the authors took a different approach to the question, figuring an “expected” clutch performance from hitters by adjusting for certain variables that might otherwise obscure a clutch effect in the statistics. They also found no evidence for the skill, and concluded that, “for all practical purposes, a player can be expected to hit equally well in the clutch as he would be expected to do in an ordinary situation.” (They also conducted the most extensive study I’ve seen on whether pitchers exhibited the “clutch” skill, specifically relief pitchers, and again found no evidence of such an effect.)
In fact, the only serious analytical work in the last forty years to claim that the clutch hitter was a real thing, a skill distinct from the basic hitting skill, was done by the Elias Sports Bureau as part of their annual Baseball Analyst books in the 1980s. These books didn’t just argue that the clutch hitting skill was real, but attacked those who argued otherwise. Unfortunately for Elias, their own lists of the best clutch hitters (using their own homegrown definition of “clutch”) varied so much from year to year that an analysis published in 1989 in the Baseball Research Journal by Harold Brooks, now a meteorologist for the National Oceanic and Atmospheric Administration, found that “the conclusion that the Elias definition of clutch hitting is irrelevant is inescapable. Clutch hitting, as presently defined, is a mirage at best.”
Indeed, many of the hitters most often labeled as “clutch” today actually aren’t. Miguel Cabrera, for example, has hit .321/.399/.562 in his career (through the end of the 2016 season); he’s actually been slightly worse than that in “late and close” situations, .298/.393/.514. He’s probably faced more tough relievers in those situations than inferior hitters would, so I would not argue that he’s somehow unclutch; I think we’re looking at a great hitter who’s still a great hitter in those situations we’d ordinarily define* as clutch.
David Ortiz, who in 2016 had one of the great swan-song seasons in baseball history, is called “clutch” so often he might as well make that his middle name. But Ortiz, a great hitter in all situations, isn’t anything more than that in situations we’d typically call “clutch:”
Even in the postseason, where Ortiz’s reputation became something of a caricature in 2013 (even as he went 2-for-22 in the ALCS win over Detroit), he’s still just the same hitter he always is, hitting .289/.404/.543 in 369 career playoff plate appearances, or a little more than half a season’s worth of playing time, getting pitched around a bit more in October than in the regular season.
And how about Derek Jeter, the player whose very nickname, “Captain Clutch,” invokes the term?
Nope. Even the Captain can’t transcend the basic limits of human ability: if you’re a good hitter, you’re a good clutch hitter, and if you’re a good clutch hitter, you were just a good hitter to begin with.
For many readers, this is a hard conclusion to accept. We all think we know about performing under pressure and then impute the same travails to big-league ballplayers. In my opinion, however, any player who can’t maintain his skill set when playing in a high-pressure or “clutch” situation won’t get very far in baseball. To reach the majors, a player has to succeed at the multiple lower levels of baseball, whether it’s high school, college, the minors, or foreign leagues. A talented player who can’t handle a high-pressure situation would be flushed out of pro ball early, which is one possible explanation for the fact that we don’t see such players in the majors. (Another is that this type of player doesn’t exist, period.) Many of the myths that fans and writers hold so dear, like the clutch hitter who can elevate his game when it counts—speaking of which, what the hell was he doing the rest of the time? Damn slacker—might actually hold true when we’re dealing with lower levels of baseball. But when it comes to the majors, the Clutch Hitter isn’t just dead; he was never alive in the first place.
The myth of lineup protection has some intuitive appeal, even though it doesn’t fully hold up under further analysis. A good hitter is “protected” in the lineup by having another good hitter batting right behind him so that a pitcher can’t simply pitch around the first hitter (or in some cases walk him intentionally) to get to the weaker hitter in the on-deck circle. Since setting the lineup is one of the most tangible things that a manager does to affect the team’s offense, understanding where lineup protection breaks down is key to figuring out who should bat where in the order.
Lineup protection does exist, at least in those lesser environments I mentioned above. The high school superstar who’s going to be a first-round pick might find himself intentionally walked four times in a game, because whoever’s hitting behind him is a tiny fraction of the hitter he is. The team is worse off for the free baserunners, because that superstar might be a .500 or .600 hitter with a slugging percentage near 1.000—that is, he had a better-than-even chance of reaching first base anyway, and was probably going to hit the ball hard enough to advance runners on base by a couple of spots. Walking him intentionally, while rather unsportsmanlike, could be the right move for the opposing coach. (It is, however, grossly unfair to the scouts who may have flown or driven for hours just to see that hitter swing the bat.)
In the major leagues, however, lineup protection does not exist. Again, people have looked for it, and found no evidence of its existence. Now, it’s possible that the shape of a hitter’s production will change if there’s a huge drop-off to whoever’s hitting behind him, but the overall value of his production will remain about the same.
Lineup protection has stuck around as a prevalent belief because it’s founded on a bit of common sense. If a great hitter has a not-so-great hitter behind him, opposing pitchers may try to pitch around him or walk him to get to the inferior guy on deck.
Tom Tango wrote a passage in the aforementioned The Book about lineup protection, finding, as you might expect, no tangible evidence that it exists, when he looked at thousands of MLB plate appearances where it might have manifested itself. Hitters who were less “protected” were moderately more likely to walk, but their performance in nonwalk plate appearances was unaffected by the lack of protection behind them. Since walking a major-league hitter is, in most cases, not a positive move for the team in the field, the lack of protection wasn’t hurting the offensive clubs in those situations—they got some free baserunners out of it. The lack of protection would become an issue if the next hitter (or sequence of hitters) was so bad that the probability of that free baserunner scoring was close to zero—for example, walking the eighth-place hitter on a National League team to get the pitcher up to the plate. That effect exists, as you can often see with mediocre eighth-place hitters who post surprising walk rates amid terrible offensive stat lines but can’t maintain the “patience” when hitting elsewhere in the lineup. Seeing the protection issue crop up when the pitcher is on deck may contribute to the endurance of the myth that protection exists elsewhere in the lineup, too.
Thus, lineup protection might change the shape of a hitter’s production, but there’s no evidence that protection (compared to a lack of protection) helps a hitter’s overall production. The shape of it might matter to fantasy baseball players who don’t care about walks, and it might matter to a manager trying to construct a lineup that will maximize the team’s run-scoring potential. These results, which confirmed previous, smaller studies on the question of lineup protection, also should serve as a deterrent to opposing managers considering the strategy of “pitching around” a hitter: you’re more likely to walk him, but aren’t any more likely to get him out just by trying to avoid giving him something to hit.
Speaking of intentional walks—let’s talk about the supposed benefit of those, which is also another myth. While the intentional walk isn’t a mythical creature like the clutch hitter or woo-woo like lineup protection, it is a wildly overused strategy in baseball that feeds off the belief in lineup protection and a general misunderstanding of run expectancy. The good news, at lea
st, is that intentional walks have come down over the last forty years; the AL level is about half of its late-1960s peak, while the NL is down about 80 percent. Intentional walks are boring; you might argue they’re cowardly; but most of all they’re just bad strategy.
Let’s look, once again, at the run expectancy table:
Now, bear in mind that these stats cover all batters in all situations, so we’re looking at an average over tens of thousands of at bats here. Barry Bonds’s career is in there somewhere, and so is Doug Flynn’s. (Flynn hit .238/.266/.294 over a ten-year, 4,000-PA major league career. You can find a worse hitter, but I wouldn’t recommend it.) And the table will change depending on who’s at the plate. I’ll circle back to this in just a moment.
Major-league teams issued 951 intentional walks in 2015, and 95 percent of them came in one of three situations: man on second (47 percent), men on second and third (30 percent), and man on third (18 percent).* So we can examine the boost that adding a trailing runner on first gives in each of those situations, which turns out to be 0.3637 with no outs, 0.2432 with one out, and 0.1207 with two outs. That’s the probability of any runner on first scoring depending on the number of outs, regardless of who else is on base. Putting a runner on via the intentional walk with no outs is like handing the opponent a card that says he gets to roll one six-sided die, and if the roll comes up 5 or 6, he gets an extra run.
There are some valid reasons for an intentional walk, however.
A manager may also choose to intentionally walk one hitter to get a more favorable platoon matchup with the next hitter. If the pitcher is a left-handed reliever with a large platoon split—meaning that he’s much more effective against lefties than against righties—then walking the right-handed hitter to get to the left-handed one could be sensible. MLB teams now have mountains of data on hitter tendencies as well, so a manager could try to match up his sinkerballer against a hitter who puts a lot of batted balls on the ground, hoping for a double play. These decisions, however, should no longer be a question of a manager’s “gut,” but informed choices that consider whether the team is better off putting an additional runner on base to gain some sort of advantage, preferably a substantial one, against the next hitter. You’re always worse off in the average case if you allow an additional baserunner, so the default argument should be not to intentionally walk a hitter; to argue in favor, you must have some solid evidence that doing so still reduces the probability that you’ll allow one or more runs in the inning.
But many intentional walks are nothing more than handing the opposition that aforementioned free shot at a bonus run. If the intentional walk does not give you a more favorable platoon matchup, or doesn’t at least bring a substantially worse hitter up to the plate, it’s probably a terrible idea. When you hear about walking a guy to set up a double play, it’s probably a terrible idea. When you hear about walking a guy intentionally because he’s “hot,” that’s almost certainly a terrible idea. And when you hear about how you can’t “let that guy beat you,” well, I’m pretty sure the goal is to not let anyone beat you, so that’s a terrible idea, too. Terrible ideas abound in baseball. It would be good to have fewer of them.
While we’re on the subject of intentional walks and lineup protection, let’s talk about the actual lineup itself. If you’ve spent any time watching or playing baseball, you know you’re supposed to bat your best hitter third, so that the first two guys can get on base for him and he can drive them in. This is also wrong, for reasons that should have been obvious from the start. Your best hitter should hit second, and we’re starting to see a few teams figure that out—although unlike some of the other changes I’ve described so far in this book, this particular adjustment is taking a while to trickle through.
Here’s the crux of the argument: Every spot in the lineup gets about 2.5 percent more plate appearances over the course of a full season than the spot after it, which amounts to another trip to the plate every eight or nine games. That means moving your third-place hitter up to the two hole in the lineup will get him something like 18 more PA over the course of a season. It’s a marginal gain, but it’s essentially all upside: why wouldn’t you want your best hitter to come to the plate 18 more times over the course of the season?
Lineup construction itself has been heavily studied and found to not make that much of a difference; if you compared an optimal lineup to an average one for a major-league team, the gain in runs over the course of a single season will likely be only about 10–15. (That figure, like many in this section, also comes from The Book, by Lichtman, Tango, and Dolphin.) The impact all around is small, but any advantage you can capture with a move that is essentially costless is worth capturing.
There is a bit of randomness involved in how often and when the hitter you’ve moved up in the lineup gets those extra plate appearances, but one potential benefit, albeit an unpredictable one, is decreasing the probability that your best hitter ends up standing in the on-deck circle to end the game. Baseball-Reference’s Play Index found 363 games in 2015 where one team ended the game with 38 plate appearances—four times through the order plus the first two batters getting a fifth turn each, so the number-three hitter was left in the on-deck circle; those teams were 177-186 in those games. There were 1,492 games that went beyond the 38th plate appearance, and as you might expect the teams’ records improve with each additional PA:
PA
W/L Record
37
207-246
38
177-186
39
194-142
40
178-117
41
144-81
. . .
. . .
47
47-12
48
31-10
That’s 363 games where the number-three hitter was left in the on-deck circle, and 453 where he was left “in the hole,” all games in which the teams on aggregate had losing records. If your best hitter is indeed hitting third in the lineup, bumping him up a spot at least gets him to the plate in more of these situations, about 23 times per team in 2016 (363 PA/30 teams), and allows him to at least extend the game. It’s an easy, obvious switch to make.
You may well ask why I wouldn’t go further and simply suggest that the best hitter lead off, with the team’s remaining hitters sorted in reverse order of offensive ability. The answer is that the leadoff spot in the lineup comes up far too often with the bases empty, and thus moving the best hitter to that spot gains you another 20 or so PA but at the cost of a lot of run-scoring opportunities from other baserunners. The Book showed that the leadoff spot has 64 percent of its PA come with the bases empty, far more than any other spot in the lineup, none of which comes in above 56 percent. So your leadoff hitter should be someone who gets on base at a high clip but doesn’t have much power—speed is nice, but entirely optional—while your number-two hitter should be the best overall bat on the club, preferably someone who gets on base and also hits for power.
While this particular myth isn’t really dying, just fading in fits and starts, we are seeing one change in lineup construction around the National League, and this one is based in some actual evidence. Pitchers nearly always hit ninth in NL lineups (and in the AL prior to the 1973 introduction of the designated hitter) until the last few seasons, when some NL managers have shifted the pitcher to the eighth spot and used the final place in the lineup as a sort of “second leadoff” position, placing a batter there with solid on-base skills so that the leadoff and number-two hitters are more likely to have someone else on base for them. I’d rather just see pitchers batting eliminated entirely—in 2015, MLB pitchers hit .132/.160/.170, dropped 488 sacrifice bunts, and struck out 42 percent of the time they didn’t bunt—but if you’ve got to have one in your lineup, considering batting him eighth instead.
The productive out is a new yet horribly misguided statistic that briefly came into vogue as a way to reward hitters for making outs that advanced runne
rs, a bit of baseball pseudoscience that sounds good because it has a veneer of validity covering an empty core. Born of a longtime bias against hitter strikeouts—the truth is a strikeout is only very slightly worse than most field outs, and of course it’s better than hitting into a double play—the productive out ascribes far too much value to a field out, because all field outs reduce the team’s “run expectancy,” how many runs, on average, the team could expect to score in that inning based on the number of outs and the men on base.
The fundamental problem with the productive-out stat is that there is never a situation where you would prefer an out to a hit or a walk. That is, the expected run production of the inning always goes up when you put more men on base, and goes down when any hitter makes an out. Making outs is bad, and getting on base is good. Even if the hitter walks and that man on third base doesn’t score right away, the probability of him scoring hasn’t gone down—the next hitter gets a crack at it—and the expected total runs scored for the inning goes up.
Part of the appeal of the productive out is the explanation that there are outs that are clearly “productive”—namely a walk-off out, one that pushes across the winning run in the bottom of the ninth or in extra innings. But in those situations, the out is only as valuable as a hit, not more valuable; the upside of the hit is capped because of the game state, not because the out is suddenly worth more. You can’t win the game twice, and there are no extra points for technical merit or artistic impression if you win on a hit rather than a sacrifice fly. A win is a win, so an out that scores the go-ahead run is worth as much as a hit that does the same thing—but it’s not worth more, because it can’t be.