East Bound and Backward

October 7, 2013

I.

                I had a thought about the Texas/Tampa Bay playoff/play-in/play-to-survive game, the 163rd game of the season that preceded the one-game playoff with Cleveland.   Texas and Tampa Bay each entered the 2013 season, April 1st, with a 1-in-30 chance to win the World Series, a 1-in-15 chance to win their league, based on the fact that there are 30 teams, 15 in each league.   But each team entered that game with a 1-in-32 chance to win the World Series, a 1-in-16 chance to win their league.     Each team had a 1-in-2 chance of winning that game, a 1-in-4 chance of winning that game AND beating Cleveland, and each team would then have to win two more series to advance to the World Series, three more series to win the World Series.

                Thus, those teams, despite winning 91 games each, had actually gone backward since opening day.   Each team had less chance of winning, entering this "playoff" game, than they did entering the season.

                By the way, I think that’s what it should be called, a playoff game.  That’s what those games were called before the playoff system we have now, playoff games.   That seems like the natural thing to call them.   I don’t know that the leagues have the right to seize control of the word "playoff" and tell us that a playoff game shouldn’t be called a playoff game, and, if they’re going to do that, the least they can do is give us some other word for it.  

                Anyway, my point was that this is the first time in history that this has happened, and I would think that it might be the first time it has ever happened in any sport—that a team entered a playoff game with LESS chance to win the championship than they had when the season started.    Going back to the last playoff game before the division split, the 1962 San Francisco/Los Angeles games, each team had a 10% chance of being league champion before the season, but a 50% chance before the playoff games.   Thus, each team had improved their position by 400% by reaching the playoff.   But in 2013, reaching the playoff game represented a step backward.  

 

II.

                My two writing interests, as you know, are baseball and crime.    HBO has restarted the series Eastbound & Down, about the life of the washed-up baseball pitcher Kenny Powers.   It strikes me that Kenny Powers is

                a)  not a realistic representation of an athlete, but

                b)  actually a brilliant representation of a criminal, of how a criminal thinks.

                Kenny Powers has a world-class entitlement mentality, very low ethics, and a remarkable ability to rationalize whatever he wants to do, no matter how childish or irresponsible.    I will tell you honestly that I have never known any athlete, and certainly not any major league athlete, who was even remotely like Kenny Powers.   Athletes do have an entitlement mentality, but it’s not an entitlement mentality anything like that of Kenny Powers.    In some way that is hard to explain, the entitlement mentality of athletes is the exact opposite of Kenny Powers.   Athletes uniquely understand—much more than non-athletes, as a rule—that success comes from hard work, success involves suffering and not whining about it, and that the payoff comes through following the rules and respecting the team concept.    Very, very few athletes—basically none—don’t get that message.   Kenny Powers doesn’t have any concept of any of that.   

                The entitlement mentality of athletes has to do with the rewards of athletic success, with what benefits should flow to the person who is able to cut it in the extremely challenging world in which they live.  

                Kenny Powers is not an athlete; he’s a criminal.    What the show illustrates very well—ironically, much better than crime shows do—is how the criminal thinks, how the criminal convinces himself that he is justified in doing whatever he wants to do.   Give Kenny Powers any rule, and he will figure out—and he will explain to you—why that rule should not apply to him.    Show Kenny Powers any possession, and he will explain to you why he should have that possession.    Kenny Powers can express, in words that make perfect sense to him, why he is better than others, why he owes no respect to anyone, why he should be able to treat people as objects.

                In shows like The Godfather and The Sopranos—and obviously, The Sopranos was a vastly better show than Eastbound & Down—but in shows like The Sopranos, the criminals are generally presented as the winners of life’s competitions.   They have troubles, yes, and those troubles often end in death, but while they are alive they get whatever they want and don’t worry about explaining why or how.    Kenny Powers is a loser.   This is what makes the show a comedy, rather than a drama; it is not about a fatally flawed winner, but about a petty, envious, whining loser, endlessly rationalizing his failures as a sort of cosmic conspiracy to deny him what is rightfully his.

                Reading that over, I wonder whether I should have said "baseball player" rather than "athlete"?    The NFL, after all, seems to have a more significant number of criminals in their population.

                But even in the NFL, it seems to me, the criminals are not really very much like Kenny Powers.    The NFL criminals are, for the most part, thugs, a "thug" being a criminal who takes what he wants by the use of force.    Kenny Powers, for all of his failings, is not a thug, either.   He is sneaky, devious and cowardly, but he doesn’t overpower people physically or pull guns on them.    So that leads us back where we were.    Kenny Powers is just really not very much like a real athlete.  

 

III.

                Football’s equivalent of the Balk rule is the False Start rule.   The Balk rule in baseball is a silly rule.   There is no reason to have such a rule, to begin with, and, if we’re determined to have such a rule, we’re going about it in the wrong way.

                Rules designed to prevent one team from deceiving the other, in general, are foolish and unworkable.    Deceiving the other team is integral to sports.    We don’t have a rule saying the base runner can’t fake a start like he is going to steal a base, do we?   We don’t have a rule saying that a player can’t fake a bunt, do we?    The Balk rule is like having a rule in basketball that a player can’t fake a shot and then pass.    Let him fake.   If one side can fake, the other side can fake.   

                Look, I’m not saying you can’t put in any limitations on what the pitcher can do to control the running game.   What I’m saying is:

                a)  Nobody actually knows what is a balk and what isn’t, including the umpires, and

                b)  This is an inevitable consequence of a poorly-thought-through rule.  

                A much better rule would be, the pitcher can only throw to first base once during any plate appearance.   One throw to first (or one fake throw to first); you can do anything you want on that throw (or fake).   After that throw, if you throw to first again and don’t get the runner, the runner goes to second.  

                The parallel rule in football is the rule penalizing the offense five yards if an offensive lineman flinches after the line comes set.   Look, I’m not saying you can’t have any rule requiring an offense to come to a set position.   I’m saying a rule that penalizes the offense five yards for an almost-imperceptible shift of position is silly.   The game isn’t supposed about the officials; it is supposed to be about the players.   Any time the official throws a flag when nothing really has happened, that’s a bad thing.

                I think rules designed to prevent fakes are in general counter-productive, but there is a more specific flaw in the way these two rules are written.    The way the balk rule is written is, in essence, that the pitcher has to make it clear that he is NOT deceiving the runner.    He has to come to a full stop.   That puts the onus on the pitcher NOT to violate the rule.  The same thing with the false start rule:  the onus is on the lineman to be absolutely still after he comes set.    The rule, if you insist on having it, should be written the other way:  the pitcher is only guilty of verboten deception if he affirmatively does something deceptive.    The same in football:  the lineman is only guilty of the forbidden deception if he actually does something deceptive.

 

IV. On the Practical Advantages of BS

                Did you ever think about this:   that Bullshit has a huge advantage over actual knowledge, in that bullshit can be created and deployed wherever it is needed, whereas actual knowledge has very limited ability to travel?

                At this time of year I get a lot of questions from people who want to know who I think is going to win the World Series, and whether it is better to win your division and have a 7-day cooling off period before you go into the playoffs or whether it is better to have to compete right up to the end of the season, and whether it is meaningful that a team ends the season playing well as opposed to limping into the playoffs after a two-week slump, and whether a team whose best pitcher has two different-colored eyes has an advantage in post-season play.

                The only honest answer to these kind of questions is "I have no idea."   This is not a satisfying answer to those who ask the questions, nor is it particularly gratifying to me to have to answer so many questions by saying "I’m sorry; I really have no idea."    Some of these questions I can’t answer because there is no answer; the only way to answer the question would be just to make something up and go with it.   Others I could answer if I had done the right research, but the right research very often would be a two-week project, and I can’t really find space in my life to undertake an awful lot of two-week projects, and anyway, if I could, it would still be two weeks before I could answer the question.

                From my perspective, then, it often seems that what people want from me is that I should be a bullshitter like the guys on TV.    "The guys on ESPN,". . ..this is what I hear, when people ask me questions about which there is no relevant research. . .."The guys on ESPN are able to give expert opinions on every question under the sun.   Why can’t you?   Why can’t you just make something up and go with it, like everybody else does?"

                Generally, we decry and denounce bullshit.   My point here is that actually there is a huge advantage to bullshit.   It fits everywhere.   You never run out of it.    It’s always in stock.   Once you decide that you are happy answering questions with bullshit. . .you’re set.   The cupboard is always full.

                I was watching a football game today.   Kansas University got ahead of a much better team 10-0, then settled in to lose the game 54-16.   The announcer had a ready explanation for the surprising start (paraphrasing).    "Texas Tech had to travel in here yesterday.    They spent most of the day getting to the airport, taking the flight, getting on the bus to Lawrence, getting off the bus, getting into the hotel, sleeping in a strange room.   Sometimes, particularly to a young team, it takes you a few minutes to get your feet on the ground and to get your head back into the game."

                Bullshit?   Of course its bullshit—but he got paid to say that.   That’s the wonderful thing about bullshit:  you can never run out of it.   No matter what happens, you can make up an explanation for it on the spot.   Every effect can always be traced back to some readily apparent cause.   Whenever you have a condition (a) and an outcome (b), you can always assert that a caused b.  Since conditions and outcomes exist in almost limitless supply, any number of supposed causal links can be established.  

                That sounds pejorative, but I don’t mean it that way.  We all do that.  His assignment was to say something that might be true, and. . .that might be true.   Nobody can prove that it isn’t.   We all use bullshit, including myself, because we are all "asked questions" or confronted by issues for which there is no obvious answer except bullshit.  

                But research isn’t like that; research runs out.   Research is always in limited supply.   Research is never around when you need it most.    That’s why knowledge can never take the place of good, old-fashioned, time-tested bullshit.

 

V.

                So I did some research, and the previous bit is relevant to what I have to say here, because the research falls so far short of what we would like to know right now.   We would like to know who the odds favor in the championship series now unfolding.    What I have for you only goes about 1% of the way toward answering that question.   The rest of it. . .you’ll just have to trust the experts on TV.

                Almost 40 years ago, I developed a way to predict who would win a post-season series based on their superiority in each area of play.   Hitting home runs was a huge positive, for example (I think it was +16) because teams which had hit home runs had (at the time I did the research) won many more post-season series than they had lost.    Hitting for a high average, on the other hand, was negative 7, because teams which hit for the higher average had, at that time, most often LOST the post-season series.   I developed this system in late September, 1975, and published it in Inside Sports in 1983. 

                Whether that system has worked since 1983, I have no idea.    60% of all post-season series ever played have been played since 1983.    That research has been buried by an avalanche of history.   I could have kept that research up to date if

                1)  I was really well organized, or

                2)  I was a sophisticated computer programmer who could pull data out of all kind of sources, or

                3)  I had a couple of days to research the subject old style.

                O-for-3.    And also, in my own defense, I just don’t operate that way, and I don’t believe in it.   I do something, I give it to the world, I move on.   Some of my ideas stick; other people pick them up, do something with them.   Most of them don’t.   Either way, I move on to something else.   I believe that in the battle of ideas, your ideas, if they are good ones, will defend themselves.  

                Anyway, I did a little bit of research here, a one-day project; wish I had more time to follow up.    Suppose that Team A has a won-lost record of 94-68, and Team B has a won-lost record of 75-87.    When two teams meet, the chance that Team A will win is:

94 * 87

--------------------------------------------

(94 * 87) +  (68 * 75)

                Which is 61.6%.   When a 94-68 team plays a 75-87 team, the 94-68 team will win 61.6% of the time.

                Well, if we know how often each team should win in one game, we can calculate by ordinary methods of statistical analysis how often each team should win the series.    In 1927, for example, the New York Yankees (110-44) faced the Pittsburgh Pirates (94-60).    A 110-44 team will beat a 94-60 team in a single game 61.5% of the time.    If you have a 61.5% chance of winning one game and you are playing a 7-game series, you have a 73.8% chance of winning the series.    So the chance that the 1927 Yankees would beat the Pirates in the World Series appears to be 73.8%. 

                This calculation, of course, relies on a series of assumptions, many of which ain’t necessarily so.   The calculation ignores the home field advantage.     It assumes that the quality of play in the two leagues (the 1927 American League and the 1927 National League, in this case) is the same.   It assumes that the won-lost record is a true indicator of the quality of each team.    All of these assumptions can be improved.  

                The question, then, is "does the method work, despite these limitations?"

                Well. .yes, up to a point.   In baseball history there have been 18 post-season series in which one team had a 70% chance to win the series, or better.   Of those 18 series, 15 have in fact been won by the team which would have been expected to win.    This chart summarizes those 18 series: 

Series1

 

                You get a 70% chance of a win by one team, basically, when you have a 15-win difference between the teams in a 7-game series, or when you have a difference of 20 wins between the teams in a 5-game series.   All but three of these series, historically, have been won by the team which should have won.

                There are another 33 series in history in which one team had at least a 65% chance to win.    Well, heck, I’ll summarize those 33 series as well: 

Series2

 

                There are 33 series in this range, of which 23 were won by the team which "should" have won.   Above 65%, then, the teams with the better regular-season record have won not merely as often as could have been expected, but more often than could have been expected.    In these 51 series, we would have expected the teams with the better won-lost records to go 36-15.   They actually went 38-13.   

                Below 65%, though, the method doesn’t seem to work. . .the method can’t prove that it works.    There have been 43 series in history (not counting 2013) in which one team had a 60 to 64.9% chance of winning.   The record of the better teams in those series is just 20-23.   There have been another 49 series in which the win expectation of the team with the better record was 57 to 59.9%.    The record of the better teams in those series is just 25-24.   

                There have been 51 series in which the win expectation of the better team is 54 to 56.9%.   The record of the better teams in those 51 series is just 25-26.

                There have been 71 series in which one team had a win expectation greater than 50%, but less than 54%.   The won-lost record of the "better" teams in those series is just 34-37. 

                Altogether, that accounts for 265 series.    There have been 272 post-season series entering this season.   In the other seven, the two teams had the same regular-season won-lost record; thus, each team had a 50/50 shot to win the series.

                Anyway, taken as a whole. . .we would have expected the teams with the better records to win 156 of the 265 series, 156-109.   They actually went just 142-123.    The method performs well above 65%, but has historically been non-instructive under 65%.

                I started this research, about ten days ago, with the intention of "throwing out" the imbalanced series, focusing on those series which appear to be most even, and studying those to determine what factors seem to be predictive in those series.    How often does the team with the better #1 starter win?   How often does the team with the home field advantage win?   How often does the team which ended the season playing well win?   Etc.,etc. 

                But I just never had time to do that research.   Which gets us back to Section IV.    Research takes too long.   We’ll have to rely on bullshit. 

 
 

COMMENTS (21 Comments, most recent shown first)

casejud
On point 1, they may have gone backward from the first game of the season but, if Tampa or Texas had won 90 games after 162, their chances would be ZERO.

Thanks for all of these pieces, are excellent reads, even though I love Eastbound and Down. There is something loveable to some about the criminal eh?
12:20 PM Oct 11th
 
therevverend
Another good comp for Powers is the late 80's Padres' bullpen. That's one thing I love about baseball. Relief pitchers can be fat and surly and loved for it. Comparing Powers to real life athletes is like comparing Leslie Nielsen's doctor in Airplane to real doctors.
8:48 AM Oct 11th
 
therevverend
I thought Pete Rose's mind worked in a similar fashion to Kenny Powers.
8:26 AM Oct 11th
 
David Kowalski
Back in the 1970s I must have had a lot of time to spare because I looked at all the pennant winners up to the start of divisional play. One thing that shone out from the data was that the full season records of teams that were markedly different did not matter. The 1959 Braves 'should" have beaten the full-season Dodgers but substituting Maury Wills for Don Zimmer at shortstop meant that the teams were not comparable. (Yes, the change was that ridiculous).

The rest of it was pretty common sense. Teams that outscored their opponents by the largest margin were likely to win. Homers and walks counted more than batting average and stolen bases. Runs scored was a slightly better predictor than ERA and stolen bases were IIRC a negative correlator.

Since then, the Marlins with their rent-a-teams pulled a similar feat.

Looking at runs scored and head-to-head, all I could do was eliminate the Rays against the Red Sox, despite Will Myers) and surmise that the Dodgers would have an edge over the Braves.

For me, the hardest series to predict was the Tigers vs. the A's. Then again, maybe that was all BS.
6:46 PM Oct 9th
 
OldBackstop
Freakonomics had an interesting piece about pundits' comments and predictions (and how they aren't paid if they don't have an opinion) Excellent site btw: freakonomics.com/2011/09/14/new-freakonomics-radio-podcast-the-folly-of-prediction/

The balk rule is ridiculous. Always thought a throw to first should simply count as a ball.

The season records are interesting, but that April team often has little resemblance to the September team. Maybe post-All Star? Run the numbers please, Bill :-)


10:06 PM Oct 8th
 
RangeFactor
An implication of the first section is: If a genie appeared in the Texas Rangers front office just before the tie-breaker game and offered to go back in time to March 31st and replay the entire regular season over again, the Texas Rangers should say "Yes". Their odds would be better of winning W.S.; they'd have a new chance to win their division, or win a Wild Card spot outright, both better paths to the World Series than the Texas/Tampa Bay tie-breaker game.
4:49 PM Oct 8th
 
tangotiger
1/11 actually, once you start game 163 for this season.
1:35 PM Oct 8th
 
tangotiger
bobb: no, you are doing it wrong.

The Rays/Rangers have 1/32 chance once Game 163 starts. The six division winners have 1/8 chance.

A RANDOM team that plays 163 has a 1/10 chance. But, if you are the Rays/Rangers, you are in a worse spot.
1:34 PM Oct 8th
 
bobburpee
in the first paragraph, you wrote: "I had a thought about the Texas/Tampa Bay playoff/play-in/play-to-survive game, the 163rd game of the season that preceded the one-game playoff with Cleveland. Texas and Tampa Bay each entered the 2013 season, April 1st, with a 1-in-30 chance to win the World Series, a 1-in-15 chance to win their league, based on the fact that there are 30 teams, 15 in each league. But each team entered that game with a 1-in-32 chance to win the World Series, a 1-in-16 chance to win their league. Each team had a 1-in-2 chance of winning that game, a 1-in-4 chance of winning that game AND beating Cleveland, and each team would then have to win two more series to advance to the World Series, three more series to win the World Series."

I think you are comparing an apple to an orange by calculating the odds in two different ways. Talent aside, teams have a 1-in-30 chance to win the World Series at the beginning of the year -- because there are 30 teams and one of them will win. Applying that same logic means that the ten playoff teams have a 1-in-10 chance to win the World Series.



10:01 AM Oct 8th
 
KaiserD2
Regarding post-season predictions, Bill, you are referring to the initial system that you published in Inside Sports (with disastrous 1982 results as I remember) and then in a subsequent Baseball Abstract. At some point in the mid-1980s you redid it in a much more sophisticated form in one of your newsletters and got even stronger results. I didn't keep the newsletter but I don't think you actually published the whole system, just a summary of how it differed.

Essentially both systems assumed that post-season games were generally low-scoring (because the pitchers were always good), that long-sequence offenses wouldn't do well while high-homer offenses would, that pitchers who could throw shutouts were a big advantage, and that--one of the biggest single findings--baserunners were precious and any team that threw a few away on the bases was very likely to lose. I have noticed that principle play out in many post-season games since.

DK
7:49 AM Oct 8th
 
hotstatrat
Bravo to your balk rule suggestion, because that would probably speed up the game. Multiple throws to first base are boring.

Counting a fake throw as a throw is the only problem. That almost brings the balk rule back to a judgement about the pitcher's movements. The idea is a step in the right direction (pun accidental), but, of course, needs to be tested to see if it is as great an idea as we think it is.
11:36 PM Oct 7th
 
CharlesSaeger
Bill: Have you read Harry Frankfort's "On Bullshit"? It isn't completely relevant to how you used the word, but seems vaguely connected to part IV.
8:31 PM Oct 7th
 
shthar
I think you've re-discovered 'male answer syndrome'.
4:05 PM Oct 7th
 
tangotiger
In 1942 NHL, and a few years before that, there were 7 teams, with six making the post-season. With one round of byes, that means there were three rounds for a few teams, or a 1 in 8 chance of winning the Cup.

What was interesting is that #1 played #2, and they waited for the winner of the other 4 for the championship round! That's quite inventive, so that there wasn't such a huge advantage to finishing top 2.

What was VERY weird was that #3 played #4, while #5 played #6. That didn't make any sense.
4:03 PM Oct 7th
 
metsfan17
I liked that 1st paragraph. That was interesting. Of course, the teams that didn't make that Play In game had even longer odds to win the World Series. 0 in infinity.
2:42 PM Oct 7th
 
MarisFan61
I think I laughed through the "BS" thing more than I ever have through a sports article.

I got an extra laugh about how we never run out of it, because it echoed what the wonderful crazy guy said in the movie Network right before he started letting it all hang out: "I ran out of bullshit; I just ran out of it." And one of the reasons that was funny was that of course you never really run out of it.
11:40 AM Oct 7th
 
tangotiger
Fascinating tidbit about the odds for the Rays (or Rangers).

In the NHL in 1979, they had 17 teams, of which 12 went to the post-season, with 4 teams getting a bye, and the other 8 having to qualify for the other 4 spots. That means, those teams had 4 rounds to go through (qualifying round, then Elite 8, Final 4, then the Championship). That's 1 in 16, compared to 1 in 17 to start the season. So, pretty close.

The Rays/Rangers had TWO bye rounds to go through, which is why their odds came out as they did.
11:18 AM Oct 7th
 
tangotiger
Regression Toward The Mean (RTTM) would suggest that the expected .589 matchup is more like a .550 matchup. So, unadjusted, the 156-109 (.589) would really mean we'd expect 146-119.

Bill observes 142-123, which is 0.5 standard deviations, and so, nothing at all.

However, the split based on extreme points is very interesting.
11:10 AM Oct 7th
 
tangotiger
This appears on every page of my blog:

imgs.xkcd.com/comics/sports.png
11:03 AM Oct 7th
 
enamee
(that was from Matthew Namee)
10:37 AM Oct 7th
 
enamee
Bill, your discussion of bullshit reminds me of a passage from The Black Swan by Taleb (page 120 in the 2nd edition):

"My biggest problem with the educational system lies precisely in that it forces students to squeeze explanations out of subject matters and shames them for withholding judgment, for uttering the "I don't know." Why did the Cold War end? Why did the Persians lose the battle of Salamis? Why did Hannibal get his behind kicked? Why did Casanova bounce back from hardship? In each of these examples, we are taking a condition, survival, and looking for explanations, instead of flipping the argument on its head and stating that conditional on such survival, one cannot read that much into the process, and should learn instead to invoke some measure of randomness (randomness, in practice, is what we don't know; to invoke randomness is to plead ignorance). ... But have the integrity to deliver your "because" very sparingly; try to limit it to situations where the "because" is derived from experiments, not backward-looking history."
10:37 AM Oct 7th
 
 
©2019 Be Jolly, Inc. All Rights Reserved.|Web site design and development by Americaneagle.com|Terms & Conditions|Privacy Policy