THE BOOK--Playing The Percentages In Baseball

"He really shouldn’t then be commenting on what has or has not been studied, since most would assume that he keeps up with the field.  He’s got my email.  I’m sure he’s got Phil Birnbaum’s email, too.  Even if we’re not necessarily the leaders in research, we’re certainly ones who keep on top of things.”

Man, you really come across as arrogant as hell...without obvious justification. James needs to look to YOU to figure out how to think about these questions? Wow.

Uhh, no.  He needs to not make public comments about what has and has not been researched without actually finding out what has and has not been researched. 

Tango’s simply offering himself and Birnbaum up as examples of people who would be able to tell him, with very little effort, what the current status of research is on almost any baseball topic. 

This has nothing to do with James figuring out how to think about different questions, and everything to do with James speaking from a position of authority on a subject of which he’s outright stated he’s ignorant about.

John/1: first I want to thank you for posting your email address.  I always appreciate feedback where the commenter is willing to stand behind what he says by allowing me to be able to reach him (if so desired).

And David/2 read my comment in exactly the spirit I intended.  It wasn’t one of arrogance.  This may simply be the case of the reader relying on the writer’s background in appreciating the basis of the comment.  David read me the way I wanted to be read, and you didn’t.  That’s my fault, but this is a blog, so I don’t have to worry to cater to too many people.  I’ll have to live with misreads from time-to-time.

Now, John, if you are a regular reader of this blog, that’s another story! 

In any case, I really have no idea who Bill James’ circle is.  I do know that he had a very long back-and-forth with Phil Birnbaum (editor of a SABR research journal) once.  And Phil’s blog is at the forefront in keeping up with sports research.  I try to do my part, and I have emailed with Bill James on a couple of occasions.  If James wanted to know the state of some research, he could contact us, and we’d be able to at the least pass him to another group of researchers.  I think that after 4 or 5 emails, James could have had the answer he wanted.

On the other hand, if James DID do this, if he did set out to find out the state of research on the subject, and came up with his conclusion, then you are right: I may be pretty arrogant to think that Phil or I wouldn’t have come up at some point.  But, I don’t see how at least Phil’s name wouldn’t have come up.  And Phil and I have had many online exchanges on the subject, so I’d figure that Phil would then bring me up.

So you’re saying that paragraph DOESN’T come across as arrogant? Read it again.

To each his own, I suppose, but this, as the cliche’ goes, ain’t rocket science. The point James reiterates, whoever has researched it, is simple in the extreme.

But I suppose if you don’t visit this site, you just don’t know what to think about baseball. How did Branch Rickey survive?

My comment (#4) was addressed to (#2), but I stand by it, in light of (#3).

Why are we assuming there is something wrong with the fact that the better teams in the NBA win more often?

The only thing I can think of is that worse teams will really suffer in terms of attendance, moreso than the worse teams in baseball… since attendance is so heavily tied to team performance.  Lengthening the shot clock may allow the worse teams to win more often, but in all other respects I think it would be a travesty.  I’m not a fan of basketball (though as a Boston resident, I must say I’m quite excited about this season), but I feel like the league is sort of thriving, isn’t it?

John/4: no one is suggesting that you need to visit this site to know what to think about baseball.  Please don’t ascribe a position to me that wasn’t even implicitly mentioned.  Try to stick to what I did say and what I intended.

Your righteous indignation is badly misplaced in this conversation, John.

John:
This isn’t that complicated: 
1) James is claiming that a variety of topics hasn’t been studied;
2) James, BY HIS OWN ADMISSION, generally doesn’t keep up with research other than his own.
So, he really isn’t qualified to make claim #1.

In fact, these issues are being studied.  For example, in this journal http://ideas.repec.org/s/jsf/intjsf.html you’ll find an article concluding that attendance is maximized when the home team has about a 66% chance of winning.  Perhaps James is familiar with all this work and finds it wanting.  But he doesn’t claim that, he just says this is “virgin terrority.” And his past admission that he doesn’t follow others’ research makes it reasonable to assume that’s still the case.

RE: #7-We’re talking past each other.

Did you actually re-read the ‘graph, BTW?

Mike/6: the assumption is that if you want a quality of competition equal to that of MLB and NHL, then the NBA season or game time needs to be shortened.

On the other hand, if the competition level of the NBA is at the right point, then for the NHL and MLB competition level to reach to that level (i.e., less randomness, more talent leading to wins), then those leagues need to change something (say, 4 periods for hockey, 4 out innings in baseball).

It’s a question of deciding how much randomness and how much talent “should” lead to wins or league championships, in terms of making things exciting for the fan.

Dubious assumption: “quality of competition"=parity.

Why is the possibilty of a decidedly mediocre team like Arizona winning the whole shebang a good thing?

John/10: actually, I’m not talking past you at all, but we’ve probably both made our positions quite clear by now.

***

Guy: “is maximized when the home team has about a 66% chance of winning. “

When I used to be a in hockey season ticket group, and we had to select tickets, everyone would always fight for the top opponent ticket (which we all agree with).  But when it came time for the 4th or 5th round, I’d start snapping up with worse opponents.

I reasoned that what I wanted was the Canadiens to play a high quality, high tempo game against the Bruins and Maple Leafs, with the great out-of-town crowd that would be there.  Or, I wanted the Canadiens to have a great chance to win the game, and send the fans in a tizzy.

As it turned out, you don’t want the opponent to be too bad, as the Lightning games (at the time) were not pretty to watch.

It’s tough to decide on what my expectation of a “good” game is.  I guess the home team having a 55-65% chance of winning might be around the right level.

I don’t know if 66% is right; was just pointing out that these issues ARE being studied. 

We need to separate two related issues here: 
1) how likely is it that a more talented team will win, given the structure of the game;
2) how much variance in talent is there among teams? 
So another way to improve competitiveness is to reduce team talent differences, such as through a salary cap, changing draft rules etc.  Here again, you can have too much of a good thing:  I think fans like to see excellence, including great teams, so 30 equal true-talent teams is not really our ideal (at least, not mine).

Still another dimension is parity over time.  I want some teams to be better than others, but not the same teams year after year.  So you want some structural forces that bring dynasties to an end and allow new teams to become great.

Lots of variables to consider.....

Reflecting some more:  economists have usually asssumed that fans want as much competitive balance as possible, since uncertainty of outcome makes a game more interesting.  But that’s not quite right, even leaving aside our parochial desire to see our home team win.  Think about a league in which all teams were truly equal in talent:  then we would know, with absolute certainty, that the champion was just lucky.  No team would deserve their position in the standings, good or bad.  The point of competition is to see who’s the BEST, and with perfect parity there would be no best. 

I suppose some would initially reject the luck explanation, arguing that each year’s winner had better character or was more “clutch” or whatever.  But as “clutch” teams routinely went first-to-last, that myth would eventually be punctured.  And all we’d have left is wondering “who will get lucky this year?”

Since 1995 (13 seasons, or 52 teams, an average of almost 4 per team), this is the number of seasons each team has made the playoffs in the AL:

teamID n
NYA 13
CLE 7
BOS 7
OAK 5
SEA 4
MIN 4
ANA 4
TEX 3
CHA 2
BAL 2
DET 1
TAM 0
TOR 0
KC 0

I think you’d want a structure where most teams are at 3-6 times, maybe 2-7 times.  I don’t see where the Tigers, Rays, Jays, and Royals combined make the playoffs once, and the Yanks make it all 13 seasons is good.

In the NL (average of a bit over 3 per team):
teamID n
ATL 11
SLN 7
HOU 6
SFN 4
SDN 4
LAN 4
ARI 4
NYN 3
CHN 3
FLO 2
COL 2
CIN 1
PHI 1
MON 0
WAS 0
PIT 0
MIL 0

(Brewers had 3 seasons in the AL.)

You have a similar situation in the AL and NL.  The difference is the cyclical nature.  The AL is much more inert.  Since 2002, FOUR AL teams have made the playoffs at least 4 times (Yanks, Sox, Angels, Twins).  In the NL, it’s only 2 (Braves, Cards).  If I make it at least 3 times, the A’s also make the cut.

I don’t know what’s better, overall.

I really happen to enjoy the NFL’s ratio of skill/chance in a game and season setting, and also the importance of each individual game.

For comparison to Tango’s list above (since 1994-1995, 13 seasons 156 playoff teams, approx 31 teams in the league, or about 5 appearances on average per team)

AFC
teamID n
IND 9
NE 9
PIT 8
MIA 7
DEN 7
KC 5
JAX 5
NYJ 5
BAL 5 (1 as original CLE)
BUF 4
SD 4
TEN 4
OAK 3
CIN 1
CLE 1 (joined as expansion 1999)
HOU 0 (joined as expansion 2002)

NFC
GB 9
PHI 8
DAL 7
SF 7
MIN 7
TB 6
NYG 5
STL 5
SEA 5 (1 in AFC)
CHI 4
DET 4
ATL 4
CAR 3
NO 2
WAS 2
ARI 1

I also think James’s article involves the cart coming before the horse in a couple of ways.

For one thing, he’s proceeding from the notion that baseball has its competitive balance dynamics right, and that the other sports ought to take measures to catch up.  But this is merely an assertion, and one that doesn’t necessarily fit all the evidence. When parity reigned supreme in the playoffs last year and the Cardinals won the World Series, a lot of people complained that the championship felt a little tarnished coming from a 83-78 team, and the ratings for the WS were not good. 

There is also not any a priori reason to believe that the balance point is the same in every sport.  One of the wonderful things about baseball is that any team can beat any other team on any given night.  On the other hand, one of the wonderful things about college football is the upsets—there is no comparable situation in baseball to where an Appalachian State can beat a Michigan.  The parity in baseball and the upset potential in college football and college basketball (and to a lesser extent the NFL and NBA) are part of the respective cultures of the sport, and what’s good for one might not be good for the other.

Thirdly, James seems to be ignoring another non-zero sum component of league health, which is the quality of play.  Why not cancel the NBA game entirely and just have a 3-point shootout?  This would surely be fairly optimal from the standpoint of allowing any team to win, but it wouldn’t make for good basketball.

Finally, it seems untrue that this kind of thing is going to be the central focal point of future research.  There just aren’t all that many questions to be resolved.  Perhaps we can find an optimal ‘parity point’ for each sport; perhaps we can settle on which potential rule changes could enhance the quality of play.  There’s a little bit of work to do with respect to optimal distributions of franchises between different geographic locations.  And there’s always work to be done with respect to the optimal amount of revenue sharing and things of that nature.  But there already has been significant work done in some of these areas, and it’s not like they’re something that benefit greatly from the new sorts of data that we’re beginning to have access to.  Moreover, these studies are less likely to produce real-world changes.  Statheads have enough trouble getting the 120 or so franchises in the four major sports to listen to them, but for this kind of thing, we’re getting down to just the four leagues, one of which (the NFL) is already *very* cognizant and aggressive about potential rules changes and their effect on TV revenues.  That’s just not a very large matrix of decision points times decision makers.

Having read the article myself now, I agree with Nate that James at least appears to be “proceeding from the notion that baseball has its competitive balance dynamics right”.  Most people I speak to seem to agree that there is not enough parity in the regular season but too much in the playoffs.

Also, the notion that the NBA is too predetermined is anything but a new one.  Even the Nets’ Jayson Smith famously said something to that effect 10 years or so ago (paraphrasing): God doesn’t even care about the NBA until the playoffs.

Actually, it was Jayson Williams in the quote above, who was famous for his wit before he was famous for his legal problems.

bea/17: good list.  That seems like a more reasonable list.  I’m not sure why James thinks that might be getting too close to parity.

nate/18: I didn’t read it as if James thinks that baseball has the notion correctly, but in re-reading it with your thought in mind, I did see it.  By not explicitly mentioning MLB, and deriding the NBA, and a bit on the NFL, his lack of mentioning MLB stands out more.

Interesting!

I have to confess that I don’t see it quite the same way as Tango.  Specifically,

>What is the “perfect balance” point, at which leagues tend most to thrive? I don’t know, because it hasn’t been studied.

Has it been studied?  I’ve seen studies that talked about competitive balance, and studies that compared attendance to league parity (didn’t The Wages of Wins do that?).  But did any reliable conclusions come out of those studies? 

Would the NFL “thrive” better if they changed the rules to create more dynasties?  Is there any research that gave a legitimate answer to that question?

Perhaps James should have written that “studies on related questions so far haven’t cast much light on the subject” instead of “it hasn’t been studied.”

As for Tango’s first set of quotes, on the sabermetrics of how and when the better team wins ... well, I didn’t see James saying *that* aspect hadn’t been studied.  If he had said that, he would have been wrong.

Does Bill James know about Tango’s work on this subject?  Perhaps not.  Tango, do you want to write something for me to post to SABR’s Statistical Analysis list on your behalf?  James definitely reads (and replies to) those postings.

On second thought, I will change my mind.  I think Tango and Guy are right, at least on the non-sabermetric aspects.  There are economists doing this kind of thing, looking at attendance and competitive balance—not exactly those specific (economic, not sabermetric) questions Bill is asking, but along the same lines.  I think if David Berri were to say, “hey, wait a minute, that’s partly what my book is about,” he’d certainly be entitled.

Phil, I wasn’t looking for any extra attention.

When I read articles in newspapers every day, there’s one thing they all have in common: they bring up related research, conclusions, or findings, as the author drives homes his point.  Evem if James’ questions aren’t exactly the ones being asked and answered, the “field” is being looked at.  There’s enough for him to have mentioned *anyone* else’s study (not necessarily mine).  His blanket statemetn about non-research that I quoted was just so out of place from his thesis.

As I read it anyway.

Dave Berri has a good post on this here:  http://dberri.wordpress.com/2007/10/08/bill-james-asks-old-questions-in-sports-economics/.  He lists a lot of academic work that has been done in these areas.  I can’t speak to the quality of all this work, but his point that the level of competitive balance seems to have only a weak relationship to fan interest sure sounds right.

Guy/25, a good post by Berri indeed.  My argument in post 22 does seem to be refuted.  Tango/24 is right too.

I don’t usually respond to people who say that I am arrogant, snark, condescending, funny, sad, etc.  If someone wants to comment on anything substantive I say, I am more than happy to respond.  Whether I, or any other researcher, is arrogant, snarky, etc., is irrelevant to their substantive arguments.  People who talk about those things other than in passing, tend to be trolls, just looking for attention or what have you.  And some people (again, mostly trolls) misread aplomb and certitiude for arrogance.  IMO, Tango tells it like it is (albeit sometimes in a language with which I am not familiar ), is not afraid to admit when he makes a mistake, yet is one of the least arrogant persons I know, which should be pretty obvious if you read his work.  What he said about James e-mailing him or Phil was not even close to being arrogant.  But then again, as I said, most people who harp on a good researcher’s perceived arrogance is usually someone who, for whatever personal reasons, feels threatened by their knowledge and intelligence.  But I am no psychologist so I’ll leave that one alone.

Nuff said about that.