I can only imagine he's thinking "everything I'm writing here is BS" as he does it, but he does it anyways, for a check. Or is this author just a moron?

I’m not sure – maybe the folks at BP can tell me since everything I know is wrong – how it is that you measure personality and tenacity with a bunch of digits? Don’t get me wrong, I’m as big as the next guy in discovering new ways to look at the game, but if we base it entirely on sabermetrics, it all feels a little hollow, doesn’t it? What the numbers can’t judge is how a player reacts to certain situations, what his make up is.

Yeah man, how could anyone possibly measure it? Not by performance, which shows up in statistics!

As, I believe, mgl has said, anything intangible is worthless.

Notice that Kirby Puckett is mired on the list at No. 9. I sent three copies of the 1991 World Series to the BP offices just to reiterate the quite obvious point.

I bet he didn't really send the copies, either.

As I've said before: Measuring clutch by comparing a player's performance in other situations to his performance in clutch situations is the wrong way to measure clutch. A player who performs no worse in clutch situations than he does normally may still very well be a clutch performer - if the typical performance from a player in a clutch situation is "worse" than the normal.

The right way to do this is to establish a baseline expected performance in the clutch, then compare the players to the expected baseline performance. If the average player sees his OPS decline by 15%, then a player who sees his OPS decline by only 5% is a clutch performer.

-- MWE

A player who performs no worse in clutch situations than he does normally may still very well be a clutch performer

Indeed, and if you slug .600 to begin with, like David Ortiz, you can still come through a lot in the clutch even if you get quite a bit worse. Basically, Ortiz seems to me like a monster at every moment of the ballgame.

All these analyses of clutch hitting depend on arbitrary definitions of "clutch." Which is fine; arbitrary can connote "heuristic," i.e. necessary as a starting point: let's try it this way and see if we get interesting results. But one problem with this arbitrariness in discussions of clutch hitting is that the "clutch" has an almost infinite number of degrees of gradation, where the highest-grade situations are so clutch that they outweigh bunches of less-clutch situations. You can say quite objectively that Francisco Cabrera is the greatest clutch hitter in baseball history. The bases are loaded, two outs, if he makes an out his team will lose the pennant, but he gets a hit and his team wins the pennant. How can a few 7th-inning at-bats in close games in May compare to the pressure that was on him at that moment? As a clutch situation, it approaches the Platonic ideal.

As, I believe, mgl has said, anything intangible is worthless.

On the other hand, just because you can't quantify something doesn't mean it's not important.

As, I believe, mgl has said, anything intangible is worthless.

That sounds like rationalism reduced to absurdity.

"The Book" talks about clutch hitting and if I recall correctly the conclusion they come to is that such a thing as clutch hitting does exist but that its effects tend to be so minor that a great hitter who is a poor clutch hitter still tends to be way better in clutch situations than a bad hitter who performs better in the clutch.

That sounds like rationalism reduced to absurdity.

I don't know about that. But it sure sounds like BS to me.

If i understand it, the situation with clutch hitting is:

1)It's easy to show that clutch hitting can't be a major effect. IOW, you could run a team completely ignorant of any clutch effect and it would have essentially no effect on the outcome

2)But, as a minor effect, it is devilishly hard to prove whether or not clutch hitting exists.

That sounds like rationalism reduced to absurdity.

How so? If you can't measure it, it means that it has no discernable impact on events.

If clutch hitting exists and have value, it has to have an impact on events somehow -- otherwise, what's the point?

I don't know about that. But it sure sounds like BS to me.

Everything that differs from your opinions sounds like BS to you.

Or is this author just a moron?

He is exactly that. He's a blogger who somehow has a job for Boston.com, the online wing of the Boston Globe. He brings nothing to the table---basically a written version of a sports-talk opening monologue.

The issue with saying "if we can't measure it we don't care about it" is that this assumes we have all the proper measurement tools. Someone, at least in theory, could be clutch but it could also be the case that we don't have the proper tool to measure it right now. This doesn't mean we should pay someone based on it now (because we can't measure it sufficiently to support that) but there still may be a real skill there, and at some point we may understand how to measure it and pay for it.

Everything that differs from your opinions sounds like BS to you.

Not true - just most of the things you spout.

As, I believe, mgl has said, anything intangible is worthless.

Something is worth what someone else is willing to pay for it. If someone is willing to pay me for something intangible, you better believe that intangible is worth something. The accumulated goodwill of some of the organizations I work for, for example, is worth hundreds of millions of dollars if not more. I assure you that Derek Jeter doesn't find his Awesomely Clutchity Clutchness to be worthless, since it's worth at least $5 million a year to him.

Now something intangible might be useless, but (1) that's a horse of a different color; and (2) in many cases it's not true. My standing amongst my colleagues is intangible, but hardly useless. For a ballplayer the same thing matters... a player's standing with other players is useful to him as well, and often to his team.

And Matt, get off your high horse. I thought you were *proud* of being a BS artist yourself?

A player who performs no worse in clutch situations than he does normally may still very well be a clutch performer - if the typical performance from a player in a clutch situation is "worse" than the normal.

I thought that hitters as a whole perform better in RISP situations.


2)But, as a minor effect, it is devilishly hard to prove whether or not clutch hitting exists.

The Book proves it pretty conclusively, IMO. Of course, the issue that follows is whether we should care.

Not true - just most of the things you spout.

"I don't need data when I have sense."

Appear!

#11 That's ONE way of looking at it, but not the only way. It could also mean that you have not identified the tools or methods to measure the effect.

I agree with MWE and will also offer another critique of these measurements of clutch. The BPro passage admits things like (with 2 outs, runner on 3rd and tie gane, a single is as good as a homer. Yet, they use very upper level stats like OPS to measure performance in the clutch. You could hit a bunch of singles in the aggregate of the aformentioned situations, and have your OPS go DOWN, and still have performed in the clutch.

You can't measure situational success/failure with blunt measuresments like OPS.

How so? If you can't measure it, it means that it has no discernable impact on events.

Nah, let's say a team wins 95 games, while their runs scored and allowed would project them to 90 wins, does that mean that those other 5 wins are imaginary?

Maybe that's not a good example, mayber something like the 1977 White Sox is better. How did they win 90 games with that roster? Was there some Bob Lemon Effect?

This is the type of stuff that Don Malcolm railed against; a sabermetric view of the game that is just as closeminded as some of the mainstream media.

Something is worth what someone else is willing to pay for it.

All of your examples involve subjective valuations rather than objective. Just because there is a market for something (indicating a belief in its value) has no bearing on whether there is an actual objective value to a baseball team.

In reference to your point, I think the proper statement is "intangible things shouldn't have value" if your goal is to win baseball games.

Craig - I don't ride a horse, but a pony.

BLB - That was in re: to one specific point, where data wasn't needed. Usually you need data. Sometimes that lump above your neck works though, especially when the data you need doesn't exist, to either confirm or deny the point I was making, in re: to that specific issue you were referencing.

It's funny that Lar, a Yankee fan who has watched Clutch God Papi shatter his dreams repeatedly would say clutch doesn't exist.

Then again I guess the existence of the greatest hero in American history could just be a coincidence.

GGC gets it. Don awlays did too, of course sometimes it was tough when sorting through the Qmax to the max.

How so? If you can't measure it, it means that it has no discernable impact on events.

Wrong. Fielding was for practical purposes impossible to measure for most of baseballs history.

Take the hypothesis: Having happy ballplayers leads to a positive impact on performance.

Presently, you have no means of measuring happiness, that doesn't mean the hypothesis is wrong or unimportant.

Some people are uncomfortable without data, some are very comfortable with data. Without this, we wouldn't have these article posting/thread discussion merry-go-rounds...

I don't ride a horse, but a pony.

I hate anyone who ever had a pony.

Just because there is a market for something (indicating a belief in its value) has no bearing on whether there is an actual objective value to a baseball team.

If that's what's he meant, then he should say it that way, instead of using a term like "worthless" which means something totally different. At any rate, there is no such thing as "actual objective value" and I object to the use of the term. All values are subjective.

I don't believe the assertion as restated, either, though I understand where it's coming from. What I suspect you mean, is that nothing intangible contributes to winning ballgames.

Hey guys,

Have you ever been on a sports team and was the best player during practice, but never the best player on the team?

Have you ever sang a song a million times in your car, but forget the words in Karoake or a performance...or just not sing it well?

Why is this? Because, retards, there is a such things a "Clutchness." I used to believe there was no such thing as this either, but then I asked myself these questions, and found out I was stupid.

If you think that batting in the bottom of the ninth inning with two on, and two out, down by three is the same as batting in the top of the fifth with two on, and two out down by three, you've assured to me you have no idea about life.

Oh yeah, how about this, when you're playing a baseball video game and the situation is critical, don't you find yourself a little more nervous than when the situatiion isn't critical.

Guys, life is life and the factors that make it interesting such as getting nervous in clutch situations exist in baseball as well.

Oh, and though nobody here will admit it, I'm sure there has been times when you've not been able to perform with a women because you were nervous.

Yeah, clutchness exists. Is it measurable? No. Does it exist? Yes. Is David Ortiz more calm when the situation matters most? Yes.

Stop being idiots, and remember baseball isn't different from real life.

These folks have obviously lost their marbles in their constant attempt to prove to the rest of the baseball world how much smarter they are.

1) When you title your stuff the way BP does ("everything you know is wrong!") you're inviting Wilbur's brand of thoughtless sneering. Moral to all sabermetricians: try to be a little humble, and more people might listen.

2) Emeigh is quite right when he says Measuring clutch by comparing a player's performance in other situations to his performance in clutch situations is the wrong way to measure clutch. His suggestion is one possible improvement to BP's approach. But the big problem here (and one sabermetrics exists in part to remedy) is the chasm between our perceptions and reality. Better hitters get more clutch hits (and clutch opps) than the "average hitter," so we perceive them in a certain way. Some sabermetric studies use over-achievement relative to the player's baseline performance as the criterion for determining clutchness, so OF COURSE the casual observer like Wilbur is going to see the results as non-intuitive. If he took the time to understand this point, he might actually find some value in the rankings, but BP botched Step #1 of any discussion: get the other party to listen and think.

3) If MGL said what Larry M. quoted, that's another problem with sabermetricians. That which is intangible is often merely difficult to measure with current technology. Far, far too often we statheads interpret "no evidence of a statistically significant effect" as "proof of non-existence." That's wrong, statistically; failing to find something doesn't mean it's not there--and even potentially large. It's what Bill James was talking about in his "Fog" article. Again, a bit more modesty is called for in some of our pronouncements.

It's funny that Lar, a Yankee fan who has watched Clutch God Papi shatter his dreams repeatedly would say clutch doesn't exist.

Um, what? When did one series negate everything else?

He "shattered my dreams" a total of two times, in Games 4 and 5 -- but they only shattered one dream.

Notice that Kirby Puckett is mired on the list at No. 9. I sent three copies of the 1991 World Series to the BP offices just to reiterate the quite obvious point.

This is the stupidest thing ever. There have been thousands of players since 1972, and Kirby Puckett is ranked #9 in clutch wins. Yet somehow this is proof that BP doesn't understand that he's a clutch hitter.

In other news,

Eddie Yost had no plate discipline
Max Carey was slow
Despite using steroids, Raffy Palmeiro was a singles hitter
Todd Helton can't get on base
Babe Ruth couldn't hit for average
Walter Johnson was a soft tosser
Steve Carlton wasn't very durable
and Mike Mussina has mediocre control

If that's what's he meant, then he should say it that way, instead of using a term like "worthless" which means something totally different.

He's making the bizarre assumption that we're valuing things on how they generate wins in baseball games.

What I suspect you mean, is that nothing intangible contributes to winning ballgames.

Not at all. If the uncertainty approaches infinity then the value we place on it should approach zero. This doesn't mean that these things don't affect outcomes (they most certainly do) but we are unable to control or predict their effects in any reasonable manner. Clutchness, BTW, does not fall into this category. It exists; it's been measured.

Let me make it clear, I'm not sure it was MGL, not even kinda sure. I know someone whose opinion I respected said it, and I vaguely recall it being MGL.

Further, don't misinterpret that statement: it's not saying that which hasn't been measured is worthless, it's that which can't be measured in any way at any time is worthless. If David Ortiz is clutch, but it never actually helps his team win a game, ever, then it's worthless to be clutch.

As, I believe, mgl has said, anything intangible is worthless.

he didn't really say this. did he?

If David Ortiz is clutch, but it never actually helps his team win a game

wha?!?!

This is the stupidest thing ever.

Particularly considering that Puckett didn't really hit that much better in the World Series than he did the rest of the season.

If David Ortiz is clutch, but it never actually helps his team win a game, ever, then it's worthless to be clutch.

Yes - Papi's worthless in terms of helping his team win his games. Thats exactly why there us a huge difference between his static run estimate values, and his game stat run estimate values last year.

wha?!?!

A) You're misinterpreting a hypothetical example, or intentionally taking that statement out of context to make me seem ignorant

B) David Ortiz's home runs last year in clutch situations weren't necessarily the result of clutch ability, but of random variation. In fact, most of them were probably random variation.

Yes - Papi's worthless in terms of helping his team win his games. Thats exactly why there us a huge difference between his static run estimate values, and his game stat run estimate values last year.

That's not what I said.

Also, Game States are worthless.

But when a game counts, really counts, do I really want Mark Grace up to bat in lieu of David Ortiz?

I guess that clutch hit in the bottom of the 9th in Game 7 of the 2001 World Series doesn't count.

WRT to Ortiz, back on Thanksgiving I hypothesized that he matched up better with the typical closer than he does with a typical starter; nothing that I can prove, but it sounded like an interesting idea. Let me see if I can dredge that thread up.

life is life and the factors that make it interesting such as getting nervous in clutch situations exist in baseball as well

Undoubtedly. But as I've noted in several of these threads, anybody who makes the major leagues has been through innumerable clutch situations. In Little League, in high school, in college or the minors, in Spring Training with everyone watching him trying to win a job, &c. Everyone who has ever played in the majors has gotten used to performing in front of thousands of people, several of whom are quite vocal about his potential to suck.

Do psychology and emotion still come into play? Sure. Might Joe Rookie be insecure about his ability to hit Roger Clemens in a playoff game? Sure. But he might feel better in the seventh inning with two men on after seeing Roger a couple of times than he did in the top of the first with nobody on. Who knows? Might David Ortiz be really, really confident about his ability to hit any pitcher? Obviously. But Royce Clayton is brimming with self-confidence, too. None of that has anything to do with "clutch" moments in baseball games as opposed to non-clutch, per se; it just has to do with the psychology of high-level competition.

Has anyone ever choked in a high-pressure situation in a major-league game because the awfulness of their immediate predicament dawned on them? I'm sure it's happened once in a while. If it didn't happen to the Cub Alex Gonzalez in the 8th inning of Game Six of the 2003 NLCS, he gave a really good impression of it happening.

Have there been major-league talents who could not deal with the stress of playing at such a high level? I'm sure there have been. They either wash out, or, in an occasional situation like Jim Eisenreich's, they overcome a purely medical condition to do just fine.

I am simply unconvinced that there are some major-league players who do fine in front of 15,000 fans, or in the fourth inning, or in June, but significantly and consistently worse (or better) in front of 50,000, or in the ninth inning, or in October.

intentionally taking that statement out of context to make me seem ignorant

You don't have to take you out of context to make you seem that way.

Have you ever been on a sports team and was the best player during practice, but never the best player on the team?
Have you ever sang a song a million times in your car, but forget the words in Karoake or a performance...or just not sing it well?

Well, yes. But I am not a professional sportsman or entertainer. If Barbra Streisand weren't a clutch singer, she'd never have made it to Vegas.

You're comparing skills that we use occasionally in social situations to the skills of professionals who've reached the top of their chosen profession. The one thing they have in common is the ability to continue to play well under pressure as they've climbed the ladder. If there's a talented ballplayer who folds under pressure, do you really suppose he'd ever make it out of the minors?

So, yes, for the reasons you give, there are definitely players who perform well under pressure and those who don't. Who are these clutch hitters?

Pretty much everyone who's ever made the major leagues.

Yeah, clutchness exists. Is it measurable? No. Does it exist? Yes. Is David Ortiz more calm when the situation matters most? Yes.


How do you know?
They've been saying the same thing about Jeter for years- but as far as I can tell since the World Series when he picked up that rep he's been a dog in clutch situations (which has had no effect on the media's opinion of him of course)

Ortiz has been clutch in a very small sample size
it could mean that "David Ortiz more calm when the situation matters most"
but doesn't quite prove it.

You (or I) as a fan/observer can look at Ortiz in a late August game and marvel at how he seems to approach the at bats that matter the most. It doesn't mean that the next time he won't get tied up by thr pitcher, or line into a hard out and start pressing the next time after that.

Todays dog may turn around and produce when it matters later. Geaorge Steinbrenner once called Dave Winfield Mr. May. Year later with the Blue Jays he drove in teh winning run of a World Series.

I have no doubt "clutch" exists
I have serious doubts that it cabn bve accurately measured, and have no doubt whatsover that it can't be accurately predicted.

Larry, how do you know that Ortiz's clutch homers are the result of random variation?' Can you prove it?

My guess is you "believe" that they are, without any data. No different that those who "believe" they are/were clutch.

Larry: Since we can't remember who said that, or what exactly was meant, maybe we should take the debate away from it, then. It's quite likely that there's some confusion here about "intangible," "measurable," "worth," etc. And I'm definitely confused about how clutch ability can never actually help a team win, so I'm not going to go there.

The point I'm making is simple but important: We often use data to test theories about the way the world works, and often the data lead us to (in stat jargon) "fail to reject the null hypothesis" that X has no effect on Y. This does not prove forever after that X has no effect on Y. The finding is specific to a set of data, a modeling method, and lots else, and in any case we have to restrict ourselves to saying we've failed to find evidence of an effect rather than proved a negative--and then seek lots of replication, await more better data, subject our model to robustness tests, etc.

This is especially so with things that are dicey to measure. Like fielding (as Swedish Chef pointed out), or intangibles or happiness or managerial inputs or whatever. Statistical researchers have often gone along for years believing that X doesn't seem to cause Y, and then better data or a better specified model or including a previously-omitted variable challenges that belief. Empirical work is hard, and findings usually tentative. Again, that's why a bit more humbleness would sabermetricians well.

You don't have to take you out of context to make you seem that way.

Did it ever occur to you to try and not be a total dick?

Larry, how do you know that Ortiz's clutch homers are the result of random variation?' Can you prove it?

Because random variation is the main cause of when almost everything happens in a baseball game.

A) You're misinterpreting a hypothetical example, or intentionally taking that statement out of context to make me seem ignorant

no i'm highlighting something you said that makes zero sense. zero. larry, that's what clutch IS. it's not clutch if it doesn't help the team win.

Because random variation is the main cause of when almost everything happens in a baseball game.

That doesn't answer the question asked, however.

Did it ever occur to you to try and not be a total dick?

It's occured to me. You'be proven over time to not deserve that curtousy. Others have shown they do deserve it, and do get it.

Because random variation is the main cause of when almost everything happens in a baseball game.

Then stop watching games and start flipping coins.

My guess is you "believe" that they are, without any data.

There are mounds of data which tell us that we should regress drastically to the mean when attempting to predict performance in clutch situations. Attributing Ortiz' homers to "clutchness" or "variation" at the exclusion of the other is silly and pointless. It's like saying a strikeout is due to the pitcher or is due to the hitter.

Thus far in his career, Ortiz has had a relatively small number of PAs in clutch situations. Therefore, when predicting his future performance in like situations (guessing at his "clutchness"), we will basically regress his performance to his general baseline. When looking at his performance retrospectively we are inclined to reduce the "clutch" factor involved because we as yet have insufficient data to conclude anything decisive about his ability in this area. Without a sufficient sample size, we should assume that it adheres to historical norms.

This does not prove forever after that X has no effect on Y.

This is absolutely correct. What I am saying (and I think Larry as well) is that when studies have attempted to find this effect and have failed to see causation time after time and year after year then we should grow more confident that the effect is - if existant - relatively small in comparison to the other factors which are obfuscating a final judgment as to correlation.

no i'm highlighting something you said that makes zero sense. zero. larry, that's what clutch IS. it's not clutch if it doesn't help the team win.

It made perfect sense.

Nobody has ever questioned that clutch hits exist, but the question has ALWAYS been "is it a skill?"

Just because someone got a hit in a clutch situation doesn't mean it was because of any especial skill beyond the ability to hit well. That doesn't mean it doesn't exist, but the mere fact that the hit occured doesn't mean the skill does exist.

And I'm not saying that clutch hitting isn't a skill, either, I'm saying that anything that cant' be measured under any circumstances has no value. Clutch hitting, as a skill, has apparently been measured in small amounts, so I'm willing to accept that it is a small skill, though seemingly one of minor importance.

Nah, let's say a team wins 95 games, while their runs scored and allowed would project them to 90 wins, does that mean that those other 5 wins are imaginary?

Of course the wins exist and (on some level) they very likely indicate some level of clutchness -- scoring runs at more important times, maximizing chances of those individual runs leading to wins.

The super-major problem that these discussions have is in the participants conflation of "clutch performance by a hitter" with "clutch ability by a hitter" into a catch-all "clutch hitters." People argue over whether or not "clutch hitters" exist and the conflation gets everyone confused and upset.

(1) Clutch performance exists. There are some times with more "pressure" in them and some players naturally (either by luck of distribution or through ability) will hit better in those times. There will always be "clutch performances by players."

(2) Clutch ability may well exist but is damned difficult to identify. People don't agree on what a clutch situation is. They don't agree on what a clutch performance is. (Is a home run in the 9th inning of June for a contending team "more clutch" than a single in the first inning in September for a player fighting not to lose his job to a rookie? Is Alex Rodriguez less clutch for hitting home runs to put his team in front in the fourth inning rather than waiting until the 7th inning to hit home runs? And so forth). And even if they agree on both of those things, it's hard to tell whether a player was "clutch" or merely getting one of his randomly distributed hits. It's apparently proven difficult to predict ahead of time which players will perform better in pre-defined "clutch situations" in a given year -- is that just the result of small sample problems or because they weren't "clutch" to begin with?

Bleah.

Clutch performance exists, but what are clutch situations and who will be the players who hit better in them, and how will we identify what "better" is and whether or not such performance (in what is likely a small sample) is anything more significant than a random distribution of their hits? These questions need to be answered in order to identify "clutch hitters" numerically, but since they're hard questions to answer, we end up with subjective, "Well, Ortiz is a clutch hitter, I can just tell," which is all well and good except for the bit where "I can just tell" is way too often a reason given for things which turn out not to be true once we find a way to measure them. "Camden Yards is a good hitters' park, I can just tell" is one such instance. "Stealing bases disrupts the pitcher, I can just tell" is another.

But the first huge problem is with people who argue whether "clutch hitters" do or don't exist because some people will think it's talking about "performance" and some people will think it's talking about "ability." The performances exist. The ability may well exist, but it's damned hard to identify.

It's occured to me. You'be proven over time to not deserve that curtousy. Others have shown they do deserve it, and do get it.

Not a Christian, huh?

Then stop watching games and start flipping coins.

Well, that's a well-reasoned response to a fairly well established fact.

Pops,

And as many others have described already, these past studies aren't necessairly up to the task at hand. Defining clutch, defining the appropraiet baseline, and defining what outcome measure to use have not been done to satisfaction.

The skill in baseball for a .300 hitter is getting 3 hits out of 10.

Random variation determines which of those 3 ABs he'll get the hits in.

And as many others have described already, these past studies aren't necessairly up to the task at hand.

No, they weren't but if the effect were something HUGE they would have shown something.

The Tango, MGL, Dolphin study is far more thorough and comprehensive. I see it as pretty much the last word on the matter barring some basic shift in what we view as clutch. If you don't have The Book, you should pick it up. It's fantastic.

49. Larry Mahnken Posted: April 07, 2006 at 10:12 AM (#1945347)
Larry, how do you know that Ortiz's clutch homers are the result of random variation?' Can you prove it?

Because random variation is the main cause of when almost everything happens in a baseball game.

But taking this vast generalization and applying it without test to the specific situation of David Ortiz is no different than people applying clutchness to David Ortiz because in general they believe "Players have the ability to step it up in general".

It's occured to me. You'be proven over time to not deserve that curtousy. Others have shown they do deserve it, and do get it.

I find this to be a really nasty mode of human interaction. Other than the pleasure from inflicting pain, what does anyone gain from treating another poster like crap?

I think that MHS, MWE, OleP and quite a few others have made points I agree with.

One other way of looking at this, though, might be useful. Player just don't get that many clutch PAs in a season. Statistics are a field perpetually bounded by sample size - in other words, how much you can say and how confidently you can say it is determined greatly not by the effect observed, but by how many times that effect is observed. And baseball necessarily limits our sample.

It seems perfectly reasonable that a player might be very, very clutch for a short period and then somewhat clutch and then not so. And there would be basically no way of knowing that statistically, and it would appear as a marginal tendency to clutchness, or a statistically insignificant one. That doesn't mean that skill would be less valuable or or unpredictable - just unpredictable using statistical tools.

That's how I've taken the "lefty masher" debate. What MGL has shown is not that there are no lefty mashers, but that the sample size isn't large enough for us to statistically identify lefty mashers. (Or at least to statistically identify based on basic "split" data.)

no i'm highlighting something you said that makes zero sense. zero. larry, that's what clutch IS. it's not clutch if it doesn't help the team win.

I think this is wrong. Last night, Ryan Doumit of the Pirates hit a two-run home run in the 8th to tie the game. I would say that was a clutch hit, even though the Pirates lost the game when Dunn hit home run in the bottom of the inning.

Yeah, clutchness exists. Is it measurable? No. Does it exist? Yes. Is David Ortiz more calm when the situation matters most? Yes.

Talk about things that don't make sense. And things that truly can't be measured. Where's the data for Ortiz' state of mind in game-wining (or game-losing) situations? How do you know that his performance in these situations doesn't result from the way his body responds to being less calm?

Stop being idiots, and remember baseball isn't different from real life.

Stop being silly and remember that it is. The consequences of failing to perform in key situations in a baseball game, even a major league baseball game, rarely approach the consequences of failing to perform in the real-life situations (war, most often) that we like to hyperbolically compare sports to.

And I'm glad that someone pointed out that being 9th on a 30+ year list of MLB perfromances is good, not bad.

Because random variation is the main cause of when almost everything happens in a baseball game.

I'm not sure I understand this Larry. Let's take a simple example.

Bottom of the 11th inning at Fenway, 2 on, 2 out, Blue Jays lead the Sox 5-4 with Papi at the plate facing Justin Speier. Papi takes two balls just off the outside corner, takes a high strike, then Speier throws a fastball down but over the plate. Ortiz gets his arms extended and puts solid wood on it, an excellent swing and drives the ball deep to straightaway centerfield. Vernon Wells sees the ball right off the bat and turns quickly; the wind (which was blowing in slightly) holds the ball up a touch and Wells catches it on the full run just before the little nook area in straightaway center. Ballgame.

I see the variations or counterfactuals, dozens of them. Maybe Papi might not have played that day. Maybe Speier might have pitched the last three days and Downs has to pitch the 11th. Maybe one of those close pitches gets called a strike. Maybe Ortiz is late on the pitch or doesn't get his full extension and he swings through it. Maybe the wind switches direction and blows it out. Maybe V-Dub can't pick up the ball right away, and the runners score.

But "random"? I don't see the randomness. Maybe the umpire's call, sure. But everything there happens for a reason, doesn't it?

...when studies have attempted to find this effect and have failed to see causation time after time and year after year then we should grow more confident that the effect is - if existant - relatively small...

It's noteworthy, however, that the oldest studies (and therefore the ones limited by poor data) failed to reject the null of no effect. And many sabermetricians therefore came to be branded (perhaps unfairly) as believing clutch ability did not exist. More granular (play-by-play) data has led to the conclusion that there might be a statistically significant though quantitatively small effect for some. IOW, the headline summary of the intellectual history here might be "Statheads Backtracking."

The next step might be to massage our models a bit (perhaps along the lines Emeigh suggested; perhaps in yet-unimagined ways) and the quantification of this effect might become more precise. It still might turn out to be small, Pops, as you say. But we've progressed down a path from "proven zero effect," and people like Wilbur are partly slamming us 'cause they remember those early erroneous pronouncements.

But taking this vast generalization and applying it without test to the specific situation of David Ortiz is no different than people applying clutchness to David Ortiz because in general they believe "Players have the ability to step it up in general".

I didn't say that David Ortiz wasn't clutch. I didn't say he was. I said that not every hit he got in a clutch situation was because he had a special clutch skill, but because he had a special hitting skill. If he has a clutch skill, I'm sure at least a few, maybe more, of those hits were made because of it, and wouldn't have been made without that skill. But to imply that every clutch hit Ortiz got was because he was a clutch hitter -- that's silly.

It's noteworthy, however, that the oldest studies (and therefore the ones limited by poor data) failed to reject the null of no effect.

Yes.

It still might turn out to be small, Pops, as you say. But we've progressed down a path from "proven zero effect," and people like Wilbur are partly slamming us 'cause they remember those early erroneous pronouncements.

I think we finally have a conclusion on this. I missed the major "clutch" arguments on this site so I can happily plead neutrality. If I ever said anything like that; then I was plainly wrong.

I think this is wrong. Last night, Ryan Doumit of the Pirates hit a two-run home run in the 8th to tie the game. I would say that was a clutch hit, even though the Pirates lost the game when Dunn hit home run in the bottom of the inning.

hmmm. maybe i should have phrased that sentance differently.

Other than the pleasure from inflicting pain, what does anyone gain from treating another poster like crap?

Um... the first one.

Craig B: Like I said a little further down, the randomness is the when, not the what. If you throw the same exact pitch to Ortiz in the same exact place at the same exact speed at the same exact time a hundred times, he's not going to do the same thing with it every time. Which result happens that particular time is the random part.

The skill is how often and well he can do it, not the when.

I'm saying that anything that cant' be measured under any circumstances has no value

I think you might want to consider rephrasing that slightly. Things that cannot be measured cannot have any valid predictive value, but they can still have actual value in determining the outcome of games (at least hypothetically). Or if that's not what you meant, maybe you should try reversing the order: if something has actual value, how do we explain our inability to measure it?

we should regress drastically to the mean when attempting to predict performance <strike>in clutch </strike>situations.

Trying to predict anything in baseball is pretty much a fool's errand anyway. Projection models for specific players in all situations aren't generally worth very much, why would a model that predicted clutch be any better.

I don't know if Papi or anyone is clutch because it is a skill, or if it was random variation or whatever term we are throwing around to discount a players performance now days. In fact I don't really care what it is. I know that it happened, and it happened with an awful lot of regularity, and I know it helped the RedSox win more games than one would expect from a player.

Papi in 04 & in 05, was like Yaz in 67. When things were on the line, he came through far more often than you would expect. That is what his legacy will be. Now if you want to say well who cares it wasn't a "skill" - I don't know if your right or wrong - I do know you missing a great game though.

Thanksgiving thread

I've started to see what RossCW was on about, though he over-emphasized it. A lot of statheads see the newest research as always being definitive. When I first came here I was skeptical of the research that clutch hitting didn't exist (ask Dan Werr, I was a Tinophile), but everyone here was almost religiously of the opinion that it didn't. I was converted, though not to the extreme. Then I saw Andy Dolphin's research a couple of years back and started to believe that clutch hitting may exist as a minor skill, though his research could, in fact, be measuring something else. There are other who have taken that research and gone to the other extreme from the old argument: clutch hitting totally exists, and you're an idiot to question it.

I believe there's some clutch skill, but I think there does need to be more research to come up with a definitive answer.

Projection models for specific players in all situations aren't generally worth very much

Yep. I never once saw Willie Mays get 302/1000 of a hit in an at bat.

I do know you missing a great game though

Nobody's missing anything. They're just explaining what they saw differently. Larry's right about one thing: you really are an ass.

Trying to predict anything in baseball is pretty much a fool's errand anyway. Projection models for specific players in all situations aren't generally worth very much, why would a model that predicted clutch be any better.

Shoot, if I'd known he wuz gonna get a hit, I'da just plunked him.

But taking this vast generalization and applying it without test to the specific situation of David Ortiz is no different than people applying clutchness to David Ortiz because in general they believe "Players have the ability to step it up in general".

Exactly----it's an obfuscation not an answer.

<iNobody's missing anything. They're just explaining what they saw differently.</i>

Exactly. On one extreme, there's the guys who look at the Yankees' last two games and see doom, then there's the other extreme: me, who says that I saw nothing happen that I didn't think could happen, and my expectations haven't changed because of two freaking games.

I've started to see what RossCW was on about, though he over-emphasized it. A lot of statheads see the newest research as always being definitive. When I first came here I was skeptical of the research that clutch hitting didn't exist (ask Dan Werr, I was a Tinophile), but everyone here was almost religiously of the opinion that it didn't. I was converted, though not to the extreme. Then I saw Andy Dolphin's research a couple of years back and started to believe that clutch hitting may exist as a minor skill, though his research could, in fact, be measuring something else. There are other who have taken that research and gone to the other extreme from the old argument: clutch hitting totally exists, and you're an idiot to question it.

I believe there's some clutch skill, but I think there does need to be more research to come up with a definitive answer.

Thanks, Larry. I appreciate that post, I think that I have mistakenly pegged you for going from a Tinophile to the exact opposite and I apologize.

Personally, I think that the most interesting study of clutch performance was mentioned in the

article. Maybe the answer lies in psychology and not stats.

Larry, can't argue much with #79. Can you see why this type of post gets different reactions than your first in this thread?

Primer discussion of that article linked in #80.

My guess is you "believe" that they are, without any data. No different that those who "believe" they are/were clutch.

The difference is that the default hypothesis is that a phenomenon/skill does not exist. The null hypothesis. That only changes when sufficient evidence is brought in on the other side. So to believe, in the absence of data, that something does not exist is epistomologically superior to believing it does exist.

Random variation determines which of those 3 ABs he'll get the hits in

Usually, facing the Texas Rangers determines that he'll get the hits then.

Usually, facing the Texas Rangers determines that he'll get the hits then.

LOL.

Anyway, back to the last two Yankees games. Obviously, I was pissed as hell that they lost, and pissed that Jeter bunted and that Proctor pitched on Tuesday, but I've come around to the belief that every particular game they win is because they generally got lucky, or played a weaker team where neither team got lucky, and every game they lose is because they got unlucky, or played a stronger team where neither team got unlucky.

Like every fan, I LIVE for the luck. If it weren't for random variation, sports would be VERY boring.

"If you think that batting in the bottom of the ninth inning with two on, and two out, down by three is the same as batting in the top of the fifth with two on, and two out down by three, you've assured to me you have no idea about life.

Oh yeah, how about this, when you're playing a baseball video game and the situation is critical, don't you find yourself a little more nervous than when the situatiion isn't critical.

Guys, life is life and the factors that make it interesting such as getting nervous in clutch situations exist in baseball as well.

Oh, and though nobody here will admit it, I'm sure there has been times when you've not been able to perform with a women because you were nervous.

Yeah, clutchness exists. Is it measurable? No. Does it exist? Yes. Is David Ortiz more calm when the situation matters most? Yes."

If "clutchness" exists in baseball, but isn't measurable, then it has no effect on winning, so it's really of no interest. What good is it that a player is calm in a "clutch" situation if he doesn't produce more than a player who isn't calm?
Cheers,
Alan Shank

The difference is that the default hypothesis is that a phenomenon/skill does not exist. The null hypothesis. That only changes when sufficient evidence is brought in on the other side. So to believe, in the absence of data, that something does not exist is epistomologically superior to believing it does exist.

This is a game people play but it doesn't actually work.

We can phrase the question differently and, if you are correct, change the default assumption:

"all players hit differently based on different contexts"

Playing with phrasing shouldn't control the default condition. The only issue is the relative strenght of the proof offered on each side of the debate.

"The skill in baseball for a .300 hitter is getting 3 hits out of 10.

Random variation determines which of those 3 ABs he'll get the hits in. "

Random variation does a heeeeeeeeeeeeeeeeel of a lot more than that. Try this. Write a program that simulates "seasons" of, say, 500 at-bats, where the probability of getting a hit is .3. Use a good pseudo-random-number generator. Have it run 1000 such seasons and analyze the results. You are going to find "seasons" in which the "batter" hits .340 (about two standard deviations above the "true skill") and .260 (two below), as well as lots more between .320 and .280 (roughly 68%).

The issue here is trying to find the "population percentage" by looking at the "sample percentage."
Cheers,
Alan Shank

We can phrase the question differently and, if you are correct, change the default assumption:

"all players hit differently based on different contexts"

Changing the phrasing doesn't change the default assumption.

The only issue is the relative strenght of the proof offered on each side of the debate.

Except that you can't prove a negative.

Random variation does a heeeeeeeeeeeeeeeeel of a lot more than that.

Well, yeah. I was just simplfying it. It's not really 3 of 10, it's 30% of an infinite number of ABs, which ABs the hits come in is the random variation part.

If you throw the same exact pitch to Ortiz in the same exact place at the same exact speed at the same exact time a hundred times, he's not going to do the same thing with it every time.

This is an untestable hypothesis and I don't believe that it's true.

This is an untestable hypothesis and I don't believe that it's true.

Of course it's true, because a batter can't reproduce the exact same swing every single time. If the ball hits the bat just an little bit later, in just a slightly different part of the bat, it's going to do something different.

The skill is in being able to hit it consistently with the good part of the bat, not to a certain place. If it was the latter, we'd see guys hitting a lot better than what they do now.

Of course it's true, because a batter can't reproduce the exact same swing every single time.

But why?

This is a semantic distinction that makes no difference at all, but I happen to believe that everything is purely deterministic. There are reasons why David Ortiz's swing differs "every single time", and if, as in your hypothetical, we repeated the exact same moment in time, with the exact same conditions 1,000 times, I believe that the result would be the exact same 1,000 times.

Or, to put it another way, I would phrase your more general claim as follows: I believe that there is variation in a player's true skill from at-bat to at-bat, for reasons that may well be "random" from the player's perspective.

But, as I said, it's untestable either way.

Random variation does a heeeeeeeeeeeeeeeeel of a lot more than that.

Well, yeah. I was just simplfying it. It's not really 3 of 10, it's 30% of an infinite number of ABs, which ABs the hits come in is the random variation part.

Stuff like this makes me cringe. There is no such thing as a ".300 hitter"... players are not machines and they don't act that way.

Whether a player gets a hit in a certain plate appearance is NOT a random event, it depends on him getting the bat on the ball and hitting it safely between the fielders. That's a skill, and while the skill varies widely depending on the conditions he faces, it doesn't alter the fact that it is human agency that is by far the largest determinant of the outcome of an at-bat. If I was a manager and a player reacted to striking out by saying "it was just a random variation in my performance" I would be seriously tempted to punch him in his goddamned self-satisfied face. Thankfully, I guess, no one who plays baseball thinks like that.

If you throw the same exact pitch to Ortiz in the same exact place at the same exact speed at the same exact time a hundred times, he's not going to do the same thing with it every time.

That's the thing. Are batters seeing different pitches in different situations? There was a recent thread on Kirk Reuter's retirement and the Giant fans were talking about how he changed his approach depending on the sitution (Becoming more of a nibbler with men on, IIRC). Is there anything in that? How many pitchers change their approach like that? Is it enough to make a difference?

There are reasons why David Ortiz's swing differs "every single time", and if, as in your hypothetical, we repeated the exact same moment in time, with the exact same conditions 1,000 times, I believe that the result would be the exact same 1,000 times.

Sure, all hail determinism, that's fine. The point I think Larry was trying to make is that if Ortiz gets the same exact pitch twice in a row (not twice in the same slice of time) he won't reproduce the exact same result. You can test this by buying a David Ortiz and setting him up in front of a pitching machine.

Ah, heck, I don't know what I think anymore.

If I was a manager and a player reacted to striking out by saying "it was just a random variation in my performance" I would be seriously tempted to punch him...

Of course, this is highly unrealistic, as I realized after writing it. The player would undoubtedly say "it was just a random variation in my performance, skip."

Equally amusing is the thought just expressed by Kiko, which is the player coming up and saying "sorry, skip" after striking out with the bases loaded, and the manager turning to him with a beaming, beatific smile and saying "that's all right, son... you could not have done otherwise."

I just thought that I'd mention that these types of threads are my favorites. I appreciate having something like this on a relatively slow workday for me.

So, yes, for the reasons you give, there are definitely players who perform well under pressure and those who don't. Who are these clutch hitters?

Pretty much everyone who's ever made the major leagues.

I agree with this. And back to the original poster who posited the opposite of this by analyzing his own life and recalled areas where he felt the pressure to perform, he stopped one step short. Whether one feels the pressure to perform is directly related to the confiedence one has in one's abilities -- clutchness doesn't only vary from person to person, it varies within a person from task to task.

As a teenager, I played in national/provincial playoffs/championships in two activities -- tennis and Space Invaders -- in tennis, I succumbed to the pressure (my arm actually felt heavier) and I performed below my capabilities in the clutch because I was insecure about how good I was. In Space Invaders, I believed (I think accurately) that there was no-one in the world who was better at it than I was and I felt no pressure as I competed for hours on end in the finals of a Canadian championship.

I'd guess 95% of MLB ballplayers have that latter confidence feeling when they are performing and that clutch success is more an effect of randomization than of succumbing to pressure less than the other competitors at that high a level of performance and confidence.

Hey guys,

Have you ever been on a sports team and was the best player during practice, but never the best player on the team?

Oh, and though nobody here will admit it, I'm sure there has been times when you've not been able to perform with a women because you were nervous.

Others have touched on this post; I'm going to try and expand on their responses, and throw in a few real-life examples. It's easy to think of the power of clutch when looking at these awesome athletes, and then remembering the things that post #28 mentions. So of course it seems intuitive to some people to see substantial differences in clutchness. Because in our lives, we've all seen people perform better or worse in pressure situations.

But the problem is, our lives are very very average (sorry!) in comparison to the average pro athlete's. So while the same things that effect us also effect pro athletes, I don't believe they do in the same proportion. As others have posted, you don't get to the major leagues without already having gotten through thousands of clutch situations.

But what others haven't really mentioned is that since all MLB players have gotten through all of that, and all of them have had to perform, thousands of times, there's really a baseline minimum of clutchness that is substantially larger in the majors than Joe Beer experiences on his weekend softball team.

Think of another profession, like investment banking. There's the cream of the crop, firms like Goldman Sachs, and then there's your small town Midwestern bank. Goldman would be the equivalent of MLB, and the Midwestern bank would be like A-ball. What if you compared the clutch performance BETWEEN the leagues? No doubt Goldman's boys bring home the deal substantially more than the Midwestern bank. They're all basically clutch. But the clutch differences between each investment banker at Goldman Sachs would be relatively small - there's a (substantial) baseline level needed to even get in the door there. But comparing those clutch performances to the average investment banker would show substantial differences.

An example from post 28 - singing in a car versus singing in front of people. The baseline for this isn't whether I can sing in the car on the way home from work vs. sing at Yankee Stadium; it's whether someone who has sung their whole lives, who does this for a living, who has already done this thousands of times, can go sing at Yankee Stadium.

Another example from 28. Of course many of us, myself included will have had anxiety issues in bed. But this is again a poor example. If you wanted a good comparison, think of male porn stars. They are pre-selected for their ability to perform when the pressure is highest.

You have to keep in mind, pro athletes have been pre-selected because of their ability to perform under pressure. So you simply can't compare your daily life to the ability of an athlete to perform in the clutch.

Are there clutch abilities at the pro level? I believe there are, but because the minimum baseline is so much higher than we understand, the differences are much smaller than we may recognize.

Pro sports isn't like life at all. It's nothing like life. We need to stop kidding ourselves that it is.