Tuesday, October 16, 2007

Computers can't teach me about the vagaries of playoff baseball, Pt. 2

Devoted reader and relative master of all things grilling Big C has re-assumed his role as TGWNA contrarian, posting a couple of challenges to The Baseball Preachings of Diesel. I feel compelled to defend/clarify some of my positions, as it's apparent that not all of my parishioners are as comfortable falling into step as others.

In re: Luck and the Playoffs

When I, or any other gambler, uses the term "roll of the dice" (or "crapshoot") it's not meant to imply a 50/50 proposition. There's a reason the house owns the seven roll in craps once a point has been established; it's the most common outcome when rolling two six-sided die. However, there's a wide gulf between saying a certain outcome is the odds-on favorite and saying that outcome is preordained; if the two terms were the same, craps would be a boring (and much more costly) game. Instead, despite knowing that seven is the most common roll, it can still be profitable (in the short term) to bet that another roll will come before a seven does, which is the essence of a craps "Pass" bet (or a "Come," which is just an offset "Pass" bet ... but I digress).

When I use the term "rolling the dice" w/r/t seven-game serieseses, what I mean is that the difference between the "favorite" (let's say they're the seven) and the "underdog" (a "Pass" bet on six) is not great enough for anyone to say what's going to happen on the next roll with any great amount of confidence. Yes, betting on rolling a six makes you a 6-to-5 underdog against a seven, but all that means is that for every five rolls of six, you should roll a seven six times. That still means you roll the six five times, though. This is the essence of a gamble, and provided you're paid according to the true odds you're facing — that is, on a $5 bet, you're paid $6 for rolling a six — you should break even.

The playoffs, ostensibly, are contended between the four best teams in each league. We know that's not always the case, as not all pennant winners are created equally, but for the sake of argument we'll leave it at that. By that definition, the gap between any two combatants in a playoff series isn't that wide; in this offseason's biggest mismatch so far, the D'Backs and the Rockies, the underdog (Arizona) was no worse than a 6.5-5 underdog in any game.

As someone who has spent way too many hours at the craps table in his life, I can attest to the fact that you can go through 30 minutes worth of rolls without ever seeing a seven-out; crazy shit happens when you roll dice. You can also go 30 minutes without ever seeing a "Pass" bet pay off. Neither happens all that often, but those things do happen.

The sweeps we've seen so far in the playoffs are indicative of nothing more magical than the fact that, in short series, it's entirely possible for a team to get on a hot roll and defy odds. And while I think all the series so far have held to form — all the "favorites" have won — the fact that the Sox, D'Backs and Rockies (twice) swept is not an accurate representation of the qualities of the teams involved. If the Rockies and the D'Backs played 100 times, it's probably safe to say the Rockies would have won somewhere between 55 and 60 of those games. However, within those 55 to 60 wins, there would likely be multiple streaks of losses that, if broken down into seven-game segments, would lead one to conclude the D'Backs were the better team if that's the only data allowed through the filter.

Furthermore, "momentum" is great until it's gone. I'm not stubborn enough to think that hot streaks can't build up an individual's confidence, but not enough for it to be a proper predictive force. The Rockies' run in the playoffs is a testament to the fact that crazy shit happens every so often, nothing more. It's also a testament to the fact that baseball officiating has absolutely tanked this season, but that's a rant better delivered by others.

One final point: The Cardinals were, without question, the worst team in the eight-team playoff field last season. They limped into the playoffs, almost blowing a prohibitive division lead down the stretch. Everything that you've mentioned as factors in favor of the Rockies was working against the Birds. And they went ahead and won the World Series anyway. Crazy shit just happens in short serieseses, man.

In re: "Clutch"

About the only argument more exasperating to me than the "clutch hitting" one is the existence of god, and for the same reason. Thirty different people will give you 30 different definitions of god, based on the differences found in various religions and personal credos. This makes it difficult to be an efficient atheist; you often feel like you're shadowboxing, since you're never quite sure what kind of being it is you're arguing against. To top it off, in the midst of an argument, the believer can change his/her belief structure to conveniently invalidate whatever it is you're arguing. This can be enraging.

It's the same with clutch hitting: Ask 30 people what defines a clutch situation, and you're likely to get 30 different sets of conditions. It creates a situation in which a non-believer (me) has difficulty disproving anything, ultimately leading to me resorting to snarky comments about unicorns. It's not really good for anyone.

About the only area in which I'll budge is the idea of a "choke," because it makes enough sense; some players might see a consistent drop in performance in high-pressure situations due to various reasons. I will also say that I think this is a much rarer situation that most make it out to be, if it exists at all. I really can't imagine that scores of major league baseball players have gotten this far without a supreme amount of confidence in their ability, if not an out-and-out desire to take at-bats in those pressure situations. There are so many prospective major league baseball players out there that I can't imagine that personality trait isn't almost completely selected out of the pool. And, I fucking promise, A-Rod is not a choke. He is the best player in baseball who has happened to go through less than a handful of short-term slumps when there was 100+ credentialed media members in attendance.

A strong belief in "clutch hitting," however, sounds a lot to me like believing in an intercessionary deity. People will always remember the dude who prayed for five days straight and woke up cured of cancer, but conveniently forget about the dozens of New Orleans residents who drowned in their attics, praying their doomed asses off. All evidence of clutch hitting is anecdotal, and never put into context. Yes, David Ortiz has hit many game-winning hits, but he's also struck out in those situations plenty of times. In fact, he probably does both at almost exactly the same rates as he does in "non-clutch" situations, whatever the fuck those might be. Isn't it enough to say David Ortiz is awesome all of the time? Do you really need the "clutch" qualifier to buttress his greatness? I think not.

8 comments:

larry b said...

Ummmmm... hello... no such thing as clutch? Maybe you've never heard of a dreamy hunk of Yankee shortstop most people like to call "Mr. November." I mean, fucking look at his nickname! Mr. November, for crying out loud! It's not important that his career regular season splits are virtually identical (actually a little better) than his career postseason splits. This one time I saw him hit a single in the 9th inning of a tie game, I think that tells you pretty much everything you need to know.

If Jeter and Ortiz had a clutch-off, I'll tell you who would win: baseball.

Anonymous said...

I think that the craps analogy that you have posited is much more accurate and telling than the coin-flip scenario that I posted earlier. I guess that it was a little silly of me to suggest that the odds may necessarily be even in any luck-based situation. That's not to say that there aren't quite a few gaping holes in the craps analogy (the fact that you can't actually control the outcome of a roll of dice, for example), but that's for another time. Critique of this analogy would be a little hollow right now from the guy who likened the spot to a coin-flip. Reductionism at its worst.

Let me put this playoff-results-are-not-just-chance theory a different way: You've said before, and I agree, that if a team can't gain a yard or two on 4th down, said team doesn't deserve to win. Same goes for the baseball postseason. If you come with your best shit and get stomped, you didn't deserve to win in the first place. I, personally, don't feel that there is anything to gain by trivializing a team's performance by statements such as '..in 100 games, they win 55 or 60, tops.'. They aren't playing 100 games, man, they are playing 7. Both teams know it, and both teams are giving everything they have in every game. To say that the results of any given series are the product of some lucky streak just really sucks the fun out of the whole experience. The D-Bax couldn't muster a win when they needed it the most and got booted from the playoffs as a result, end of story. It's like me saying that the Broncos were fortunate against the Pats in the '06 playoff divisional game by saying that, 70 times out of 100, the Pats win that game. 'Lucky' would be if an acute, localized micro-avalanche buried Tom Brady and Cory Dillon at halftime. The Broncos outplayed the Pats and were the better team that day, as manifested by the result (horrendous officiating aside). As much as it pains me to say this, Denver earned that win.

We'll have to agree to disagree about clutch. You make a very good point, though, that 'clutch' is all about how one defines it. You're spot-on about this; everyone and their mother has a different take on clutch, and all of them think that their definition is incontrovertible. I'm no different in that I believe that the premise for my assertion is inscrutable: All situations (and therefore, results) are not equal. Is 'clutch' overblown, overemphasized, and overwrought? Sure! Does it exist? Of course it does. I'm sure that any player in the majors is there because of his ability to adequately deal with pressure......at the lower levels. I'm certain that the pressure derived from, say, playing at Fenway in Boston is a leeeeetle more intense than quadruple A Palookaville, where even your mom isn't watching. Oh, well, when in Rome.

Unknown said...

"To say that the results of any given series are the product of some lucky streak just really sucks the fun out of the whole experience."

Sometimes the truth just sucks. That's life.

As far as "clutch", you need to read "The Book" Hitters who are "clutch" over a small sample size seem to eventually lose their "clutchiness"

It's human nature to look for patterns in things that sometimes just don't have them.

Anonymous said...

@gregp: I think that you are right about human nature. It is a basic search for meaning that drives any sort of analysis. If there are no patterns in baseball, then what is sabermetrics all about? Baseball is not a random system, it is a chaotic system. The fact that the sport is not utterly random is the only thing that makes any sort of predictive metric (like a teams Pythagorean for projected wins)useful at all.

I'm now interested in "The Book" (actual title?), because it may either bolster or erode my point. You have to have 'clutch' to lose it; if 'clutchy' players lose clutch with age, that could be an indication that a downgrade in skill level correlates to a downgrade in 'clutch'. On the other hand, if players in general seem to come randomly in and out of 'clutchiness', this makes 'clutch' seem imagined in the first place. Perhaps I'll come out of this with a more informed position, after all.

Lastly, if my statement that you quoted in your post is, in fact, the truth about baseball; you are, by extension, saying that baseball sucks. It would be like watching craps on TV and rooting for six. Exciting. But this, I believe, is not the case. The result of a baseball game is the sum of the outcomes of various instances (i.e every pitch has a limited number of results). The outcomes of these incidences are controlled by human beings who are subject to random as well as systematic fluctuations in performance. These individual performances affect, to some degree, the outcome of a game. Baseball is a cerebral game, and it doesn't take much (read: pressure) to affect the psyche, and therefore performance, of a given player. What you are rooting for in every game (especially in the playoffs) is that your team's players are going to perform at a higher level than the other team's players, even if the other team's players normally perform at a higher rate. That's why baseball can be fun, in my mind, because although the reasonable expectation of a team may be a loss, that can be made up for with a proper amount of determination.

Anonymous said...

can't wait for baseball to be over so you guys will start talking about football. You know, like everyone else in this country not associated with AARP.

hey pepe, which hurts more: the fact that the eagles lost AGAIN, or the possibility of the Rams, the team you picked to win the west, might very well go 0-16?

Anonymous said...

Ditto; but it is hard for a Pats fan to talk about football right now without coming off a little smug. Ah, fuck it; YAHDOOD, GO PATS! 16-0 FOR SHO-AH!

Unknown said...

http://www.insidethebook.com/

Unknown said...

I think the main problem with a short series in baseball is that individual player performance in a single game has a huge standard deviation.

A 0.300/0.360/0.420 hitter over a sample of 500 AB can do just about anything over a sample of 20 AB. If the majority of hitters on a team just happen to perform at the lower end of their possible outcomes at the same time (eg Cubs/2007) you end up losing.

As an example, if you look at ARod's career record against pitchers against whom he has more than 50 AB's there is no correlation between his OPS against them and their OPS allowed. And this is even for a "reasonable" sample size.

It's quite a bit different in football and even basketball. If Tom Brady completes 60% of his passes throughout the year, his standard deviation for a single game might only be 10%. It is very very unlikely that he will complete only 30% of his passes in a playoff game. That is way football playoffs are more reasonable, even in a single elimination format.

Now granted, if you could clone Josh Beckett, the chances of winning a best of 7 would be very high, but the majority of baseball playoffs are just glorified coin flips. I totally agree with what Diesel wrote on this subject.