“Art Howe seems to get short-shrift in the film. I understand that he’s upset about his portrayal—specifically that the film portrays him as insubordinate. Frankly, I felt more sympathetic towards him. I mean, if you’re paying the guy to manage the team, let him manage the team.”
I do think Howe is portrayed somewhat sympathetically, but his problem is a classic middle management problem. I think what you’re really paying a manager for is to get a maximum number of wins from the talent on a team’s roster.
“My sense is that statistics don’t capture this psychological component of handling pressure very well”
Well, the problem with basing evaluation on quantifiables is that there are always going to be features that elude easy quantification, yes, but baseball statisticians are fairly adavanced in this regard—you can easily find stats for how, for example, a pitcher does against certain counts, how he does “close and late”, etc. None of these are definitive, but . . .
" To be successful post-season team, you have to build a team that has the personnel and approach that will be the most successful in the post-season."
Not to be glib, but one thing you’re leaving out is that, to be successful in the postseason, you have to make the postseason by winning enough games in the regular season. Sabermetrics was never really intended as a magic bullet—you still probably need either a good core of young players via the draft or virtually unlimited payroll (ala the Yankees) . . . ideally both. But eventually, if you’re a small market team, those young players’ first contract is eventually going to run out and you’re likely going to have to replace them.
The Red Sox success following the years the film is set were a product of a big payroll + the application of some of the principles of sabermetrics.
I think what you’re really paying a manager for is to get a maximum number of wins from the talent on a team’s roster.
I agree it’s a classic middle-management problem—specifically micro-management or interference from superiors. If the manager has to get maximum number of wins and you’re telling him who he must start and altering his batting order, that’s really not a fair proposition, imo. (But life’s not fair, I know.)
Well, the problem with basing evaluation on quantifiables is that there are always going to be features that elude easy quantification, yes, but baseball statisticians are fairly adavanced in this regard—…
And yet there are still significant limitations, imo—limitations that are just as bad as the type of blind-spots that stem from conventional thinking.
Btw, I mentioned psychological effects because statistics sometimes fail to capture this. (The theories on always going for it on fourth down are another example of this, imo
Not to be glib, but one thing you’re leaving out is that, to be successful in the postseason, you have to make the postseason by winning enough games in the regular season.
Right, but incorporating tactics or players that have great value in the post-season, but less value during the regular season, doesn’t significantly jeopardize making it to the post-season, imo. Beane (or Lewis) fails to acknowledge the way certain tactics and personnel can have great value in the post-season, and therefore should be incorporated in the regular season, even though this may not be the most statistically sound decisions. (The thing is, Lewis scoffs at people who value things like base-stealing and closers—as if the people who value these things are blind by tradition—i.e., that there isn’t a sound reason for adopting these tactics or getting these players. But I think he has a blindness like the people he criticizes.)
Before I forget, I also wanted to ask if anyone has analyzed the Twins’ success and compared this to the A’s approach? My sense is that they were successful without using or heavily relying on sabermetrics, yet they are a pretty small to mid-sized market team.
One thing about statistics: Statistics are never intended to be infallible magic numbers. Any self-respecting statistician will make sure that his results make sense, and not use the results of a statistical analysis as an absolute certainty. So a statistician would probably agree with your argument that both statistics and judgment/perception… only he would incorporate judgment/perception into his analysis in the first place.
Just thought I should chime in, since I recently finished my Statistics degree… I don’t really know anything about sports though.
“Before I forget, I also wanted to ask if anyone has analyzed the Twins’ success and compared this to the A’s approach? My sense is that they were successful without using or heavily relying on sabermetrics, yet they are a pretty small to mid-sized market team.”
Yeah, their payroll has snuck up into the upper tier over the years, but generally speaking, they were pretty successful with a low payroll. There’s certainly no one right way to build a team.
@Matt
But do you know if anyone has analyzed the reason for their success?
@Drunken
So a statistician would probably agree with your argument that both statistics and judgment/perception… only he would incorporate judgment/perception into his analysis in the first place.
Here’s the reason(s) people over-rely on statistics: a) sometimes statistics can reveal silly biases in conventional thinking and so there’s a temptation to dismiss judgment; b) some people love statistics, while having a more negative opinion towards judgment (which is more subjective and hard, if not impossible, to measure).
“But do you know if anyone has analyzed the reason for their success?”
You mean in terms of how they approach player decisions?
Player decisions as well as style of play (as you know Beane never wanted the team to attempt to steal bases because of statistics).
They’ve not very orientated toward statistical analysis (and have in fact been very vocal about this over the years). Style of play, they’ve generally been closer to a traditional NL style of play that what you expect from an AL team.
That’s interesting. I’d be interested in reading a book about them as a comparison to Moneyball.
Not sure that there’s much of an angle there, except to show that occasionally a more traditional approach can still sort of pay off (the lesson there is probably that a combination of execution and luck mattter as much as theoretical orientation).
What potentially interesting, though, is understanding how and why a team like smaller market team like Minnesota (or the Rays) could be so successful—if they were not using a more statistically based approach. What made their management, coaching and scouting successful? Were there other factors, unrelated to management, coaching, etc. that made these teams competitive? That’s where I’m coming from.
“What made their management, coaching and scouting successful? Were there other factors, unrelated to management, coaching, etc. that made these teams competitive?”
But this would really be getting back into the sort of elementary, conventional speculation about the game that has dominated for years and years, wouldn’t it? I suspect it would just come down to the Twins doing things the conventional way better than most of the mid-level and larger payroll AL teams . . . plus luck.
Tampa Bay actually does incorporate sabermetrics into what they do. In fact, I haven’t read it, but there’s a book called The Extra 2%: How Wall Street Strategies Took a Major League Baseball Team from Worst to First (obviously indebted to Moneyball), that details how they got where they are.
The most important thing for a small-to-medium-sized market team, though, is probably to develop players internally, so that’s where I’d start if I were writing such a study. I can speak about the Rays with a little more authority than I can about the Twins. The difference between their philosophy and Beane’s, for example, is more in the on-the-field approach. They stole 155 bases this season, but they also got caught stealing a league-leading 62 times. Did + 155 bases plus -62 baserunners create runs? Generally, I think statistical analysis would show that in order to generate more runs by successfully stealing than you lose by getting caught stealing, you have to be successful stealing bases about 69-70% percent of the time (the majority of MLB teams had a % between 68-80% this season.
But this would really be getting back into the sort of elementary, conventional speculation about the game that has dominated for years and years, wouldn’t it?
What’s wrong with that—unless you mean that “conventional speculation” doesn’t include a well-argued and well-researched opinion. Of course, any response would technically be “speculative”, but making a reasonable, well-argued explanation is possible; and I’d find that interesting.
The most important thing for a small-to-medium-sized market team, though, is probably to develop players internally, so that’s where I’d start if I were writing such a study.
That would be one place I’d look, too. If this seemed promising, I’d also compare the way this team developed players to other teams. Jumping to football for a moment, I think you could identify qualities that made coaches like Bill Walsh and Bill Parcells so successful at developing QBs. Part of their success had to do with their philosophy towards the game and part of it had to do with their ability to assess QBs and develop them. It might have been possible to identify specific ways these coaches were so successful at developing QBs.
Generally, I think statistical analysis would show that in order to generate more runs by successfully stealing than you lose by getting caught stealing, you have to be successful stealing bases about 69-70% percent of the time (the majority of MLB teams had a % between 68-80% this season.
As I mentioned earlier, I think there might be another way to assess the value of base stealing, though—and statistics may not necessarily capture this value. I’m thinking of the effects of base stealing and overall speed on the base paths during the post-season.
“I’m thinking of the effects of base stealing and overall speed on the base paths during the post-season.”
But if base stealing doesn’t lead to runs, what’s the value in it?
Here’s the thing about base stealing.
If you have a good base stealer, stealing a base increases your odds of scoring one run.
But, reduces the average amounts of runs you will score.
So, the only time you should be stealing bases is if it’s late in the game and you are behind one run.
This is not ignored by saber metrics, you just have to take the game situation into account. In real life they do this, they just didn’t take it into account in the film because it was over the script’s complexity level.
You have to be careful not to mislead yourself with the mythology of the game, and the personality it takes on based on the way it’s marketed (Individual personalities). They say somebody is ‘hot’ or ‘cold’, but most of the time it’s just drawing a conclusion based on the size of the sample group. There is certainly more pressure in the postseason, but you can’t look at somebody who did badly over 20 at bats and say they can’t handle postseason play. So long as they can handle pressure, which pretty much anyone who can play in the majors can, there shouldn’t be that big a difference.
Of course, there is a difference between playing only great teams and playing a mix of good and bad teams, and maybe the formula should take that into account as well.
Also, saber metrics don’t apply the same way to another game like basketball or football like they apply to baseball. Baseball results are mostly based on the attrition of individual accomplishments, and there’s generally only one dominant offensive and defensive strategy to play. Statistics apply a lot more to baseball than other sports, whereas in basketball you can play with dramatically different offensive and defensive strategies, and any individual play depends on the entire team working together.
@Matt
Ultimately, I’m saying that it can lead to runs—but not in terms of increasing the average over many games. In other words, in can lead to runs in one or a few games. Here’s what I said:
Moreover, I think base-stealing players/approach may possibly make teams more resilient to pressure and put more pressure on opponents. Here’s my thinking: a) if you have a speedy and effective base stealer, batters don’t feel like they need big hits to drive in runs—especially if the base runner can steal to second; b) at the same time, this puts more pressure on the pitcher and defense, because he knows that the base-runner is fast and that if he steals second, he’ll be in significant threat. He has to worry about the guy stealing and he might be troubled that a single can drive in a run. Now suppose the team isn’t built for speed and eschews base-running. This takes away quite a bit of pressure on the defense.
So I’m hypothesizing that the strategy could lead to more runs because of the way it affects pressure—mitigating its effect on your team while exacerbating it for your opponents. This can increase the performance of your team while decreasing the performance of your opponents—i.e., getting runs.
@Jirin
This is not ignored by saber metrics, you just have to take the game situation into account.
Based on the book, the As almost never attempt to steal bases. It’s a strategic decision. Moreover, when you hear Lewis talk about base-stealing, he sounds like stealing is absolutely foolish—in any situation.
Statistics may apply to baseball more than football or basketball. I honestly don’t know enough about managing baseball to say one way or the other. But there are situations in basketball and baseball where a statistical approach becomes relevant. For example, there have been papers examining punting on fourth down. Some have argued that teams should go for it on fourth down more often—and some go so far as saying teams should never punt. (There’s a high school team somewhere that does this.)
In basketball, I ran across an article discussing end of the game situations—specifically when you should and shouldn’t intentionally miss a free throw or put the opponent on the line.
“I’m hypothesizing that the strategy could lead to more runs”
It could in certain situations, yes, depending on who’s pitching, who’s catching, who’s batting, who’s on base, etc.
“the As almost never attempt to steal bases”
Well, remember that was that specific team, which stole a league low 46 bases all season, and were also caught stealing only 20 times (second fewest in the league that year), so it was a low-risk, low reward strategy. The Marlins that year stole 177 bases and were caught stealing 73 times. Oakland had a dead-slow team that year, so it makes sense for them not focus more on moving base runners other ways.
It could in certain situations, yes, depending on who’s pitching, who’s catching, who’s batting, who’s on base, etc.
Therefore, to dismiss base-stealing in outright—to ridicule a manager or player stealing a base, as Lewis has done—is unwise and, well, off base.
Well, remember that was that specific team, which stole a league low 46 bases all season, and were also caught stealing only 20 times (second fewest in the league that year), so it was a low-risk, low reward strategy. The Marlins that year stole 177 bases and were caught stealing 73 times. Oakland had a dead-slow team that year, so it makes sense for them not focus more on moving base runners other ways.
But they chose players (overlooking speed) and strategy (almost never stealing) based on statistics—while ignoring the possible value of base stealing in the playoffs. What I’m saying is that even if we assume that base-stealing is not a strategy over the course of a long season, it may be good in a short one, namely, the playoffs. And if this is the case, then might want to build a team for speed and integrate base stealing during the regular season. Why? Because you can’t completely switch your identity and style of play when going to the post-season.
“it may be good in a short one”
It may be, but I’m not sure you can rely on it being a good one, because you don’t know who you’re going to face in the postseason.
Also, though, the game has changed significantly from 2002 to now in terms of the norms of base stealing. Back then it was a much more of a station-to-station (and PED) game, so the smaller market teams—Kansas City and Florida and Montreal—tended to be the ones that stole a lot of bases. By 2011, you have even the bigger market teams stealing more bases, and among this year’s playoff teams, the Yankees, Tampa Bay, Texas and Arizona all finished the season among the top ten is team steals. The A’s meanwhile are right in the middle of the pack in total steals, though (and Coco Crisp, 2nd in MLB in steals this year, plays for the A’s). So, yeah, Lewis may be sticking to the “steals are bad” line, but Beane himself these days, not so much. If he’s got someone like Crisp who is stealing at about an 85% success ratio, he’s going to let him steal . . . especially when his team’s on-base percentage is in the bottom third of the league.
It may be, but I’m not sure you can rely on it being a good one, because you don’t know who you’re going to face in the postseason.
The opponents matter and your personnel (including injuries) matter as well. You might factor in conditions like the ballpark, weather and umpires as well. My point is that stealing shouldn’t be dismissed outright; that it can offer advantages through short-series, where every game counts—in a way that statistics don’t really capture.
Yeah, I don’t think you can lay out too many unchanging precepts about the game because it changes too much over the course of a few seasons.
Jazzaloha
Some questions and comments relating to statistics, sports and other aspects of the film:
>Art Howe seems to get short-shrift in the film. I understand that he’s upset about his portrayal—specifically that the film portrays him as insubordinate. Frankly, I felt more sympathetic towards him. I mean, if you’re paying the guy to manage the team, let him manage the team. Maybe Howe, as portrayed in the film, behaved in an insubordinate fashion, but I think he makes valid points.
In any event, my main question revolves around Art Howe’s contribution to the team’s success. Does anyone know if there was anything written about his performance and contributions to the team’s success as a manager. I think another book from the perspective of the way Howe’s expertise and knowledge contributed to the success of the team, if it did at all.
>My feeling about the use statistics in terms of management and coaching is similar to my feeling towards the nature (genes)-nurture (environment) debate—namely, that both are required. I think human judgment/perception and statistics have blind-spots. If a GM or coach/manager uses both wisely they can avoid those blind-spots and make wise decisions for the team.
>Billy Beane (supposedly) said two things that I find ironic: 1) The playoffs are a crap-shoot; 2) Nothing matter if you lose the last game (which is equivalent to saying you’ve got to win the championship, or all your accomplishments in the entire season will be negated). I basically agree with both statements, but I have a very different philosophy from these two points. If the last game(s) is the most important and the last game(s) are a crap-shoot, then building a team and style of play based on statistics is unwise. (If championships depended strictly on the regular season record, generally, an approach that depended more on statistics would be the wiser approach.) Let me briefly lay out some of my ideas about post-season play and how it relates to the regular season and sabermetric approach:
1. The biggest difference between the regular and post-seasons are mental—specifically dealing with pressure (and to a lesser extend fatigue and injuries—sometimes physical injuries are beyond mental toughness). The stakes are higher—and keep getting higher near the end of the game and as the series moves to the final game (the final out of the final game being the peak of pressure, generally). Frequently, the champions handle the pressure the best—more specifically, championship teams can be built in a way and play in style that is not only resilient to pressure, but turns the more pressure on their opponents.
2. My sense is that statistics don’t capture this psychological component of handling pressure very well, so to build a time more on statistics we tactics and styles of play that do make a team resilient to pressure and make the opponents feel pressure at the same time.
Let me try to use one example from the book, namely, base stealing. From what I understand, Beane basically doesn’t want the A’s to ever steal bases because the risks aren’t worth the benefits. These odds may be true over the course of many games. Of course, in the playoffs, the sample size is much smaller. Moreover, I think base-stealing players/approach may possibly make teams more resilient to pressure and put more pressure on opponents. Here’s my thinking: a) if you have a speedy and effective base stealer, batters don’t feel like they need big hits to drive in runs—especially if the base runner can steal to second; b) at the same time, this puts more pressure on the pitcher and defense, because he knows that the base-runner is fast and that if he steals second, he’ll be in significant threat. He has to worry about the guy stealing and he might be troubled that a single can drive in a run. Now suppose the team isn’t built for speed and eschews base-running. This takes away quite a bit of pressure on the defense.
3) To be successful post-season team, you have to build a team that has the personnel and approach that will be the most successful in the post-season. So if the sabermetric approach works in a long season, but doesn’t really work in a short one (i.e., the post-season), then you should base your team on this approach. Instead, you have to make a team that has the best chance of success in the post-season—this is a team that takes makes teams more resilient to pressure and increases pressure on opponents.
If a team could play one style and have one set of personnel in the regular and dramatically switch the style and personnel in the playoffs, relying heavily on statistical approach during the regular season and switching to another approach in the post-season, then that might be the way to go. But that way probably won’t work. So, the wiser move, imo, is to make sure your team can play in a way that will give you the most chance to succeed in the post-season and to play this way in the regular season. (This doesn’t mean you ignore statistical approach, but incorporating base-stealing might not be as foolish as Beane seems to think.)