AN Book Club, Topic 2 - The Myth of Clutch Hitting

[EDITOR'S NOTE: Here is the second installment of bbg's excellent diary pieces that deal with the origination of the A's current philosophy. - Blez]

In case you missed the intro to the book club.

Just so we're clear, this is not me taking a baseball book and trying to fit it to the A's philosophy. This is me taking the original book that shaped the A's philosophy and telling you about it. Here's some background on Eric Walker, lest you think that any of these ideas are mine, or that they have not been a part of the overall A's philosophy, since the early 80s (which surprised me a great deal!)

...Lewis's second false impression [in Moneyball], according to Schwartz [a senior writer at Baseball America, author of a book called The Numbers Game], is that Beane introduced the mathematical approach to the As. The honor for that, according to Schwartz, goes to an NPR sports reporter called Eric Walker, way back in 1981. Schwartz tells the whole story.
In the 1970s, Walker, a former aerospace engineer, had started doing some radio reporting for the NPR affiliate KQED in San Francisco. A chance visit to a San Francisco Giants game was all it took for him to see through his engineer's eyes what few others appeared to have noticed: the supreme importance of walks and of a batter not getting out. Walker started to talk about his observation, and some ideas to capitalize on it, in his daily five-minute morning baseball report on NPR.
He also took his ideas to the Giants, but they never bought into them, so a few months later he took his pitch across the Bay to the Oakland As. By then, he had written up his ideas in a little book called The Sinister First Baseman. The As' legal counsel, Sandy Alderson, had heard Walker on NPR, and had just read his book, and as a result Walker was very well received by the club. The A's hired him as a consultant, and started to implement his ideas. Among the decisions they made by following Walker's creed were:

-In June 1984, they drafted slugger Mark McGwire tenth overall rather than two more speed-oriented players, Shane Mack and Oddibe McDowell.
-In 1986, they let go slugger Dave Kingman (35 homers and 94 RBIs the previous season) because he rarely walked, having a low OBP of .258. They signed Reggie Jackson (OBP .381) as new designated hitter.
-In 1987 they traded Alfredo Griffin, who drew few walks, for pitcher Bob Welch.
-In 1988 they signed Don Baylor, a power hitter who frequently got on base by being hit by the ball.
-In 1989, they acquired Rickey Henderson, who had a super record of both homers and walks, and Ken Phelps, another player with a high OBP, both from the Yankees.
-In 1990 they acquired another good walker, Harold Baines.

By concentrating on OBP, the A's became a highly successful team, winning four division titles and three American League pennants from 1988 to 1992. Then they hired Billy Beane, and started to indoctrinate him in their ways.

Trust me, by the time we get to the diaries about OBPs, the myth of choosing defense over offense, the useless statistic of errors, the myth of "intangibles", and predicting Joe Morgan's reaction to Moneyball way back in 1982, you will recognize the Oakland A's.

I know everyone has a preconceived notion of `clutch hitting'. I invite you to put aside any opinion and just read the entire diary before dismissing it. There is some math to work through, but if you still believe in `clutch hitting', you'll be surprised at the outcome.

From The Sinister First Baseman by Eric Walker, "Of Black Cats and Calculators":

"Coming through in the clutch" has always been held out as one of the clearest hallmarks of a quality ball team or individual for that matter. This is one of a large number of baseball superstitions, most of which arise from a common source, a failure to grasp the elements of random-phenomenon patterns ("stochastic theory," the science types call it).

Don't misunderstand me; there's a reason why the front office of baseball management may prefer to hire people who have little invested in the game of baseball to crunch some of these numbers. Again, hitting in baseball compares more to a series of Texas Hold `Em than it does to performance in a sport like football; in reality, that's not a bad comparison. No one feels the need to blame the poker player when the River card is crap; all the player can really do is make the right bets, go with the right cards, set everything up correctly. Similarly, a player can have a great at-bat, give himself every chance to get a hit, try to hit the ball as well as he can, but sometimes just has to take comfort in knowing that over the course of enough games, the hits will even out.

I think people want baseball to be like football; if a player just tries a little harder, that they can be the best, the superstar, and get the big hit at the right time. I'm going to define what this diary is about, simply to stave off comments about Chavez's lack of heart or seeming inability to hit for power when he is the only slugger on a team. I'm talking about clutch hitting here, not non-clutch hitting. You'll find players in every sport, including baseball this time, who do better in certain roles, and are set up to try too hard in pressure-filled situations. You can have a bad at-bat. If you have enough of them, you'll try too hard when you shouldn't. Baseball is not a game of robots; it's a game of human beings. BUT by the same token, it is also a game very dependent on random patterns, and you can't hold human baseball players to a physically impossible standard.

 Specifically, what I'm talking about is this: Is there a quantifiable statistic that shows that a certain player (or team) has a tendency to hit really well in the "big" situations (i.e. men on base or October)? The Book says no. Emphatically no.  

When you say that a player is a "clutch hitter", this is what you are saying. You are agreeing to two conditions: that a) he somehow possesses the ability to control during which at-bats he is going to get a hit and  b) that you can prove this by showing me that his numbers with runners on base or during October are statistically higher than his regular career numbers in a reasonable sample size. By the way, neither of these conditions hold true.

To put it plainly, over the course of a season, for a .333 career hitter, you can reasonably expect a hit to come in roughly 200 of his 600 ABs. Depending on how much power a player has, a percentage of those will be extra-base hits. Those are his stats, spread out over a season, over a career. Players do not go to the plate saying, "I know I hit .333, so over the next few plate appearances, I need to decide wisely when to get my hit." If players could hit on command, they would get a hit just about every time they got a good pitch. Want more proof? Walker pulls data from the large sample size of the history of Major League Baseball and proves there is neither a clutch hitting team, nor a clutch hitting player.

In certain respects, a baseball player is like a die. Nowadays, dice are very precisely controlled so as to be as identical as it is humanly possible to make them, but once upon a time they were merely slightly polished bits of bone; various early dice would give roughly comparable results, but no two were quite alike. A ball player's trips to the plate are very like a roll of such a die. If, let us say, a 1 or 2 comes up, it's a hit, a 3 is a walk, a 4 is a strikeout, a 5 is an infield ground-out, and a six is an outfield flyout. (These assignments don't quite approximate real percentages, and they aren't meant to; it's just a simile, and I'm sure you get the idea.) Over a few hundred rolls of such a die, one would normally get a fairly mixed succession of results, with overall results tending toward that particular slightly-irregular die's norms; but one also would see shorter-term runs of eccentric luck-an unusually high incidence of 4's, or lots of 2's, or very few 6's and so on. No sane person would say such a die has "found the groove" or is "tanking it", but when the batter's results do the same...well you know the story.

Ball players' long-term results inevitably tend toward true representations of their overall ability; short-term results inevitably fluctuate considerably from dumb, blind luck.

So what Walker is saying is that a player's career statistics are already set for the at-bat, just like the die. All that's left is to roll it. You can make sure it's on a smooth surface to get the maximum effect, you can make sure it doesn't fall off the table and ruin the roll, but other than that, that's all the control you have. Similarly, a player can have the best possible at-bat, but ultimately, the final control of the hit is out of their hands.  

So how do you build a team? You get the guy with the numbers, not the "clutch" hitter. You take A-rod over Jeter every day and twice on Sundays. And as for the A's? You can't afford the guy with the huge numbers, so you have to get the guy with the batting eye and the high OBP, or any other statistic currently undervalued on the market. But you don't search for the nonexistent player who is average the rest of the season, but somehow can hit during the big moments.

Take the best hitter you can imagine. He's a career .350 hitter. The bases are loaded. It's October. He's your go-to guy. He's your superstar. The pitch is thrown. The baseball fan in you knows he's a clutch hitter. You just know that he's going to get a hit. But you're a practical baseball fan and you know the odds. Do you realize that it is absolutely more likely for this hitter to make an out than to get a hit? Somewhere inside us, we know this. Yet, weirdly, at the same time, we don't. Again, I'm not talking about him having an awful at-bat. I'm talking about him seeing every pitch, fouling some back, taking pitches that he knows he can't handle, really concentrating. Do you realize that even if he sets himself up with the greatest at-bat ever, it is STILL likely that he is going to fail to get a hit? And when he does, the comments start to fly. "No heart, no passion, he's not good in the clutch, he folds under pressure, etc etc". The truth is (and no one wants to hear this) one of his hits just did not come during that at-bat. Period.

To sum up a rather complicated math introduction, I'll just say that in order to prove that a hitter is "clutch", we need to have enough of a sample size to show that the phenomenon is not just taking place over a few at-bats, and we need to have a significant enough spike in percentage overall to prove there is a real difference. Walker did the calculations, first for a team:

A commonly adduced statistic for clutch performance is batting average with men on base versus overall average. Let's assume this team had 5600 at-bats in the season...let's also assume their overall team seasonal batting average was .250, again a reasonable norm. If we assume that fully half the time (a conservatively optimistic assumption) there is someone on base, that means 2800 such at-bats. Sparing you the actual calculations, their batting average with runners on base could then be as much as 16 points higher (.266) without being demonstrably meaningful; it would have to be fully 25 points higher (.275) to be conclusively significant. On a team basis, such performance would be extraordinary indeed! In fact, it just doesn't happen. With individual players, the case is even clearer. Remember that the result depends on the sample size. Consider a player who batted .270 overall in 600 at-bats. How high would his average with runners on base have to be to show that he is indeed a clutch performer? Again, we give the benefit and generously assume runners on fully half the time or 300 at-bats' worth. His average would then have to be .321 to even be possibly statistically significant. To be conclusively so, it would have to be an awesome .347. On a career basis (to provide some perspective), our .270 hitter after five very full years to play might have 3000 at-bats; if we again assume runners on base half the time, the key figures would be .293 for his average to even be possibly, and .304 for it to be conclusively, demonstrative of any ability to hit better with men on base. If his stats fail the lower number, we can reasonably assume that it was mere blind chance that the results from that particular selection of 1500 at-bats. To put it another way, we could have taken his at-bats on even-numbered days of the month, or on Tuesdays, Thursdays, and Saturdays, and have been equally likely to find 428 hits out of the 1500 at-bats so selected. Nobody, however, would refer to a great Tuesday-Thursday-Saturday hitter (I hope). Yet to most observers a .285 five-year average with runners on base for a .270 overall hitter would seem meaningful. Sill not convinced? Let's look a little deeper. Now we need to handle a bit more arithmetic, bordering on algebra (hello? hello? anyone still there?). Skipping the very lengthy why of it, it turns out that there's a formula that relates a team's batting performance closely quite closely to its runs scored. Now "classical" baseball theory (an elegant name for that superstition) says that this should be impossible. A formula can't "know" which hits or walks came with runners in scoring position or with no one on at all. It has no way to know if a team is a "good clutch-hitting team" or a bunch of sad sacks. And in fact, the derivation of the formula implicitly assumes that hits are purely random in distribution over all possible circumstances, clutch and otherwise. E pur si muove, as Galileo said. Yet it works. The median error size is about 2 ¼ percent (about 16 runs per season in the National League, 17 in the higher-scoring American League), and the average error (combining plus and minus errors) is virtually zero; errors over 5 percent are rare.

R = (H+BB) x (TB) x (4040) / (AB-H) x (AB + BB)

The fact such a formula exists and works necessarily confounds utterly any belief in "clutch" performance. That the superstition has survived so long is evidence of how, as the saying goes, a man hears what he wants to hear, and disregards the rest. Pinch-hit homeruns are memorable, like crystal-ball predictions that come true; pinch-hit strikeouts, like unfulfilled prophecies, are dull, easily and quickly forgotten.

And no one, but no one seems to believe in the Law of Averages. When the breaks go our way, why that's just the way things happen; but when they go against us, we rail bitterly at the unfairness of it all. It's all fair, brother, it's all fair, and we have to take it as a piece, the bad with the good, the strikeouts with the hits.

And as for clutch playoff performances, let's take our cue from the New York papers. Here's what we `know' from this October: Derek Jeter is a clutch hitter and A-Rod is an October choke. Let's look at everyone's favorite `clutch hitter'. If he's the greatest playoff hitter and clutch performer ever, why does he make an out at all in October? And you can't say, "Statistics show that he can't go 45 for 45 in the playoffs". You can't bend math around what you want to believe. How is it possible that Jeter somehow knows when to hit a base hit and when to ground out? When it's important, he gets a base hit, but when it doesn't matter, he can relax and get out? And this is somehow something that A-rod hasn't yet learned? When to be clutch and when not to be? A-rod is a career .307 hitter. I'm very sorry, but he does not get to choose in which at-bats those hits come. He can have good at-bats--those he can control--but no hitter can get a base hit in every situation the team needs him to, regardless of how much heart he has, how much desire he has, how much he wants to, or how hard he tries!

It's very easy to label someone clutch. Just keep changing the definition, and eventually your player will fit the bill. What is clutch? Getting a hit with men on base? With two outs? In the ninth inning? In the seventh game of the ALCS? In September? October? With a man on third? When it's a one-run game and not a blowout?

One of the reasons I like the website Fire Joe Morgan so much, is because they get it. They understand that clutch hitting is, for all intents and purposes, a self-perpetuating myth. Example. As we all know, every hit Jeter gets that drives a man in, and every hit for him in October is called `clutch'. I'll borrow an idea from my baseball mentor, who couldn't wait to talk to me after this year's ALDS. Remember the solo homerun that Jeter hit in game five, Yanks down 3 runs? After he hit it, the announcers went CRAZY saying how 'clutch' Jeter was and what a super October player he is, of course. As you do. Ironclad guarantee: If A-rod hits that home run, it's meaningless because there was no one on base, the Yankees were losing by three and he's `unclutch' for hitting the homerun at the wrong time. Of course, any baseball fan with access to a computer can pull up some stats regarding Jeter's `clutch' playoff performances, not how people view them, but how they actually are, versus his `unclutch' teammate, A-rod.  

Just to illustrate how much people don't want to believe any of this, let's take some real quotes from baseball fans about Derek Jeter and his `clutch postseason hitting':

Statistically speaking, the shortstops you mentioned have had better years than Jeter did this year. What you failed to bring into the equation are the intangibles; overall baseball knowledge, desire to win and being able to get the big hit when the team really needs it.

You do not grasp the one personality trait that causes a person to finish ahead of a seemingly more capable and talented opponent. Desire.

Championships dictate a winner!!!!!!!!!!!!!!!
In a pressure spot,at crunch time, up to this point .....I'll take Jeter over anyone.
AROD might be the best player in the game......................but until he proves he can "win" Jeter is on a level of which alex has yet to attain

And the New York papers...

From the writers at Fire Joe Morgan (taking on Skip Bayless):

In pinstripes, A-Rod turns into C-minus-Rod in October.

(FJM) In last year's ALDS against Minnesota, A-Rod OPSed 1.213. In pinstripes. In October. Similar sample size to what he did this year in the ALDS. Who was writing "A-Rod = Clutch-Rod" articles after that series?

If you could pour whatever is inside Derek Jeter into A-Rod, you would have the greatest baseball player ever. But something has always been missing in A-Rod's makeup: mental toughness, guts, whatever it is that allows Barry Bonds, Albert Pujols and Manny Ramirez to make entire teams better. A-Rod doesn't seize the biggest moments. They seize him, often by the throat.

(FJM) Barry Bonds, up until 2002, was criticized as being one of the worst postseason chokers of all time. Here are his batting averages from three NLDSes and two NLCSes from the years 1990-2000:
.167    .148    .261    .250    .176
For ten years, Barry Bonds performed poorly in October. Then he did something -- regressed to the mean, perhaps, or perhaps he took some magical intangible mental toughness pills that gave him intestinal fortitude -- and he absolutely destroyed the teams he faced in the 2002 playoffs, hitting eight home runs and OPSing three million (I'm estimating).
Then, in 2003, he was terrible in the playoffs again.
Will A-Rod break out in a big way in some future playoff series? If I had to bet, I would bet the house on that happening.

This who eventually could wind up with more career homers than Bonds or Hank Aaron. Yet this is a guy you wouldn't want in your October foxhole. C-minus-Rod finally made Yankees fans long for the days of a far less talented -- but far more clutch -- Scott Brosius.

(FJM) Finally. Finally you write what idiots all across Yankeeland are thinking.
Scott Brosius postseason BA /OBP / SLG: .245 / .278 / .418
Scott Brosius ALDS BA / OBP / SLG: .167 / .196 / .259
That's right. In 54 ALDS at bats, Scott Brosius recorded two extra-base hits. Two.

Still more, this time from Tim Keown:

The only thing harder to figure than Joe West's strike zone in the deciding game of the AL Division Series between the Yankees and Angels on Monday night was Alex Rodriguez. The curious case of the world's most talented head case continued through another shortened postseason for the Yankees. While 22-year-old Ervin Santana was creating a little slice of legend by burying 94-mph fastballs under the hands of some of baseball's best hitters, A-Rod was carving out a chapter of zero-RBI underachievement.

(FJM) Oh boy. Okay, so you're coming out firing. A-Rod is "the world's most talented head case"? Your first reason: he had zero RBI. You classify this as an "underachievement." Perhaps. Let's take a look at the achievements of one Mr. Reginald Martinez Jackson, a man many consider to be the greatest postseason performer of all time.

1973 ALCS, 5 games: 0 RBI
1974 ALCS, 4 games: 1 RBI
1974 WS, 5 games: 1 RBI
1977 ALCS, 5 games: 1 RBI
1980 ALCS, 3 games: 0 RBI

Am I cherry-picking Reggie's worst RBI playoff series? Yes, that is exactly what I'm doing. But Tim Keown is cherry-picking the worst playoff series in A-Rod's career to nail him to the cross. In his last three series, Rodriguez has totaled 5, 3, and 5 RBI. That underachiever.

In case you missed them, here are the numbers:

Career Numbers - Regular Season:
A-Rod - OBP/SLG .385, .577
Derek Jeter - OBP/SLG .386, .461

Career Numbers - Playoffs:
A-Rod - OBP/SLG, 1995-2004 (103 AB): .395 / .583
Derek Jeter (Anti A-Rod) - OBP/SLG, 1996-2004 (441 AB): .380 / .456

Despite what the New York press and Derek Jeter fans want to sell you, Derek Jeter's playoff numbers line up pretty closely with his regular season numbers, something which makes sense. What also makes sense is that Alex Rodriguez is the better overall hitter, and yes, even in the playoffs. Yet how many fans would choose Jeter over A-Rod to make a plate appearance with the game on the line simply because of his `clutch' label?
Let's take a look at Pujols' "clutch" homerun last week. Pujols is a .330 career hitter with power. During about a third of his ABs, he's going to get a hit, and a certain percentage of those hits are going to be a homeruns. Lidge hung a pitch and Pujols didn't miss it. But there is no statistical difference between him getting that hit in the 9th and him getting that hit in the 1st. Pujols was 0-4 leading up to that AB. Did he suck in every other at-bat, but somehow knew how to hit when it was important? No, of course not! That's crazy!

No one in the history of the sport of baseball has ever mastered the art of getting a base hit more than an average of forty percent of their at-bats. It's irresponsible to claim that some hitters hit below their career numbers when there's nothing at stake, yet somehow hit over their career numbers when it "matters", and it all evens out that way.  If that is truly the case, and it is something that players control, then why do they make outs at all? Claiming that `clutch hitting' is real is the equivalent of saying that a baseball player can choose when his outs happen and when his hits happen. There is a certain element of randomness there; there just is! Baseball is not all skill!

Albert Pujols is one of the best hitters in baseball and has tremendous power. One of his homeruns came at the most awesome, dramatic, awe-inspiring moment of the playoffs for the Cardinals. That's what makes baseball so amazing; ANYTHING can happen, and when it does, it leaves a mark. If Pujols grounds out instead, it's just another 4-2 playoff loss that no one remembers when discussing `clutch' players, unless he plays in New York, in which case, he's a choker.

When an established .280 hitter has a .250 season, we need not give up on him, as some teams would. Nor, when he has a .310 season, need we believe he has suddenly discovered the secret of life, and will hit over .300 for the rest of his career. To believe such things is to be in a class with the superstitious dice shooter who believes that the dice themselves are `hot' or `cold' for him.

Adrian Beltre. He had an unreal year last year. He is a career .270 hitter who hit well above his career norms in 2004. Who couldn't logically predict what his numbers would look like overall? (Psssst...the answer is: Seattle) Don't tell me that De Podesta didn't know this. You don't magically figure out how to hit during your seventh year in the league and then forget everything you learned in your eighth. Law of averages says that Beltre will hit .270, no matter how he gets there. Anyone want to talk about Chavez, a career .275 hitter? In 2005, he hit six points off his average. If you want to gripe, talk about his lower .OBP and SLG, talk about his terrible ABs, but there is no use in talking about how many `clutch hits' he didn't rack up.

I believe that a player can have a good at bat, giving himself every opportunity to succeed, which is important, but a hit is a hit is a hit. And even with the best intentions, the best pitch, the best situation, the best baseball heart in the world, he still will fail to get a hit during about seventy percent of his at-bats. But it's our faulty memories that will label a player as clutch or non-clutch based on what happened during the situation in question.

Okay, let's hear it.

Next week: The myth of how important defense is to a team; how a single good offensive player can help a team win more games over a season than an entire defensive team upgrade