NBA / Jun 28, 2011 / 5:30 pm

Dissecting The Truth: The Sabermetrics Effect


Can you cut open a frog to see how it jumps? Jonah Lehrer, a contributing editor at Wired, argues on Grantland.com that you can’t. Hardcore statistical analysis of sports – sabermetrics, a baseball term, is used as a catch-all – is leading to grievous errors on teams’ and fans’ parts by ignoring what can’t be measured by obsessing on what can. Stat-heads, he argues, believe statistical accomplishments trump the rest, turning human competition into a series of fantasy leagues where everything is on paper and nothing is appreciated. There is no real insight, only leader boards. Sabermetrics is missing the point of the team.

Lehrer’s piece, posted this morning, has caused a bit of a stir, and rightly so. Broadly criticizing statistical analysis is a time-honored editorial decision, and as good a foot to start off on for a non-sports writer as any. Lehrer’s piece is succinct but bubbly and cleverly written. Fans can tire of sabermetrics if they’re not following the discussion in real-time. At first take, much of sabermetric discussion is dismissive: “Scoring baskets is overrated, let’s move on to efficiency.” It can suck the fun out of a game. But while Lehrer does well at focusing his attention on sabermetrics, he’s not altogether honest and he might be entirely wrong.

Lehrer is off on many levels – no specific stats or metrics are mentioned, all evidence is anecdotal – but is most transparently incorrect lumping coaches in with statistically-obsessed general managers and fans. It’s simply not the case.

He argues that “coaches and executives” use sabermetrics to neglect important values that can’t be qualified in players, like “not being an asshole, playoff experience … and listening to the coach.” His argument that “coaches and fans use the numbers as an excuse to ignore everything else,” sticks out: Most coaches recuse themselves from outright statistical analysis. “It’s not a fantasy team,” is a common refrain among MLB managers; NBA coaches play the best guys they have as much as they can.

Lehrer doesn’t name a single GM or coach who forms rosters or runs lineups this way. The putative inattentive asshole with no playoff experience isn’t named; neither are the scores of willfully ignorant general managers and coaches.

(It might be true these qualities are important to team success, but the extent to which is unknown. Chris Paul did no favors for his then-coach Byron Scott; Kevin Durant and Dwight Howard may be too friendly to capitalize on their skills. All three are most likely the best players at their position.)

To Lehrer’s credit, some organizations are indeed statistically obsessed. One team hired a man named Roland Beech, formerly the webmaster of the statistics site 82games.com. Beech sat on the bench and helped the head coach set optimum lineups. The team’s assistant coach in charge of defense went on record proclaiming his statistical bent. The owner called out the statistician as a “key part” of how the season finished. If Lehrer had bothered cutting open the frog, he’d have seen it was Dallas.

Follow Sam on Twitter at @ samreiss_ .

Follow Dime on Twitter at @DimeMag.

Become a fan of Dime Magazine on Facebook HERE.

Related Posts with Thumbnails
  • Promoman

    I think Lehrer is right. Just look at the Rasheed Wallace era Trail Blazers, the New York Knicks just 2-3 years ago, and the Washington Redskins. Both teams had rosters full of elite talent on paper but weren’t shit when it came to performance. Like the old saying goes: Sports don’t build character, it reveals it. The same goes for work too. If you’ve ever taken or seen a personality test, there’s guaranteed to be at least one question that that asks about how you behave and/or cope with things that happen in both instances. If you’re an asshole in sports or work, then you’re an assshole in other areas of your life. The only question is what circumstance makes it manifest. You have to consider politics in team sports. In team sports, you have players who trash and even try to sabotage other players because they’re jealous. They catch feelings to point of pulling bitch moves but not enough to try to invest in themselves to try to be better players or to even try to unite to win. As for Chris Paul, he’s been pulling slick shit for years. He’s like Shaq, they’re both people who are really good at presenting themselves as cool but they will fuck you with a sandpaper condom if you’re not hip to their shit. Jason Kidd has gotten coaches fired too and old heads remember the Toni Braxton thing too. Some players set themselves up to get fucked over by tolerating bullshit. Yao Ming was in the same boat as Dwight Howard and Kevin Durant are in now.

  • sam

    Agreed with you on most of that.

    Shaq, Paul and Kidd have pulled shit, but coaches and GMs will live with it because of their game. Yao was soft, sure, but he was also permanently injured — that’s what killed him.

    I can’t think off the top of my head of a player who wouldn’t pass to another. If it’s that bad, it’s either in practice or well-hidden.

    As for those three teams — I don’t know. Knicks had scorers and nothing else. Skins had old marquee names; Blazers came close with Pippen that one time. I can’t see a metric giving those teams preseason title chances, much less any midseason love.

  • http://deleted dagwaller

    Seems to me that the title of this article is misleading. It should have been called, “What I think about Jonah Lehrer’s article: A book report”.

  • Darnell

    That’s great and all, except for the fact that this very good post analyzing a recent article with a lot of buzz, incorporating a relevant topic to the most recent NBA champions, in no way whatsoever resembles a “book report.” I’d have no choice but to assume the whole conversation and concept just went over your head.

  • Darnell

    @Promoman — No love lost, but none of the three teams you cited would have held up under any advanced statistical measure. The Knicks, for example, didn’t exactly have elite talent in a “sabermetric” sense.

    Using Hollinger’s PER as a relatively well-respected baseline analytical tool, the 1998-99 Knicks’ best-ranked player were David Lee (34) and Nate Robinson (36). For their next-best player, you have to go all the way down to Al Harrington at (107). Early on, they traded Zach Randolph (25) and Jamal Crawford (124). They were 17th in offensive efficiency and 24th in defensive efficiency. Conversely, the NBA champion Lakers had Kobe Bryant (6), Pau Gasol (14) and Andrew Bynum (24). They ranked 3rd on offense and 5th on defense.

    Take it back to three years ago, before NYK started trimming the fat in their quest for LeBron: Randolph (48), Lee (52), Crawford (89), Nate (106), Eddy Curry (!) (115). They were 24th in offensive efficiency and dead last in defensive efficiency. The NBA champion Celtics featured one elite player in Kevin Garnett (4), along with Leon Powe (22, outlier skewed by small sample size), Paul Pierce (30) and Ray Allen (83). But though their Hollinger offensive efficiency was 12th, they ranked first in defense, and featured a good blend of relatively elite talent and a solid team concept, which showed up via both the eye test and statistical metrics.

    The 2002-03 Trail Blazers (the last full year for Rasheed, and the first recorded year of Hollinger’s stats on ESPN) tied for ninth on offense and ranked 13th on defense. They had solid individual players, with no singular elite talents — I won’t count Rasheed, who never averaged 20 ppg. They won 50 games, one better than their expected result, and lost in seven in the first round to the Mavs. None of this information is incongruous.

    I’m less a football guy, but when have the Redskins ever had elite talent compared to the rest of the NFL? Start at quarterback; how would they have ever competed? The closest they had was McNabb, and they got him way too old.

    Just saying — like you said, statistical metrics obviously aren’t perfect, there’s a human element you can’t factor out, but their potential usefulness can’t be discounted, which was the writer’s point.

    Sorry for the book report.


    Wrong. Statistics/metrics tell all, they tell us the Rasheed Wallace era Blazers teams were not as good as we thought/assumed. Everything shows up in the statistics… everything.

    Same with players who might be negative in the locker room… it shows up in the stats if it is having any effect… FACT.

    If a player is an asshole and this affects team chemistry then it WILL show up in the statistics, if a player has playoff experience and can guide a young team to success then it WILL show up in the statistics. All these sports writers are dinosaurs, who hate advanced statistics because they can no longer bullshit their readers because fans know more than they do now and can prove it.

    People who want to ignore statistics are the same people who assume a short, white second baseman is a scrappy player and the tall black outfielder is a rangy athlete.

  • sam

    Darnell — thanks for taking the time to address these teams’ shortcomings.

    iCarnacki – you got data on that? Not sure if you’re being sarcastic. The point I was making was that advanced metrics are reasonably more predictable or descriptive than basic stats which everyone already uses to begin with.

    There’s no one who follows a team and watches sports stat-free, excepting maybe a developmental scout.

    Anyways thanks for reading

  • Darnell

    @sam – Sure thing, it’s an interesting discussion.

    @icarnacki – I guess I see what you’re attempting to get across, but you almost certainly won’t find anyone other than legions of straw men arguing that modern statistical understanding can override a guy being an asshole, or a criminal, or whatever — which obviously, can’t be determined by a number. This doesn’t take away from sabermetrics’ usefulness in projection or comprehension when it comes to player or team analysis. Obviously, character is important to keep under consideration, and you’ll see guys drop in drafts because of it. But more often than not, when a team is good, there are underlying statistical factors that explain that, and can help overcome having a supremely talented dick (Kobe) or a total nut job (Artest).

  • http://deleted dagwaller

    @Darnell – No, the point didn’t go over my head. Maybe when you read the original article, you projected the creativity onto this one. Every paragraph by Reiss references the original article directly. In other words, it’s a summary and commentary on the original article.

    Almost like…a book report. I’ll forgive you for not understanding, since it’s probably been a while since you did anything of the like.

    My comment had nothing to do with the original article. For what it’s worth, I agree in general with Lehrer’s piece.

    I was pointing out that the author of THIS article, Sam Reiss, has merely summarized the original. He offers no original thought, other than at the very end, when he mentions the Dallas Mavericks hiring of a stathead (ignoring other teams around the league who have no doubt done the same thing).

    I know that you enjoy cutting and pasting, so it doesn’t surprise me that you would stand up for a piece that has done little more than that.

    Your own book report has received an F. You made two posts outside of your original unprovoked (unless your last name is Reiss) attack on me. The first rambled for a while before finally stumbling to its end:

    “Just saying…there’s a human element you can’t factor out, but their potential usefulness can’t be discounted, which was the writer’s point.”

    Ok, so you agree with Lehrer…right?

    “But more often than not, when a team is good, there are underlying statistical factors that explain that, and can help overcome having a supremely talented dick (Kobe) or a total nut job (Artest).”

    So…you don’t agree with Lehrer?


  • Darnell

    @dagwaller — “Just saying…there’s a human element you can’t factor out, but their potential usefulness can’t be discounted, which was the writer’s point.” — “The writer” referred to Sam. I’d still think that would be an obvious interpolation for most anyone who read the rest of my comments, but I suppose I could have phrased that part a bit better. It happens, life goes on.

    I guess we just come from different perspectives; I saw plenty of original, worthwhile thought in Reiss’ commentary. (Isn’t that what commentary — your word — is?) But I often find it’s easier for some people to attempt to tear other people down than to attempt to raise their own level. “Book report” was an unprovoked comment with a negative connotation; I liked his post, so I responded. You agree in general with Lehrer’s flawed article; you don’t mention why. You say I cut and paste, with no actual basis. You imply Sam and I are the same person — we’re not, though we’d probably get along.

    My point, dumbed down a bit, was that statistical analysis and traditional scouting/personality study go hand in hand. You can’t really discount either one, which was what Lehrer was trying to do with analysis. And I think more often than not, we find that if a team performs well in advanced statistical metrics, that they’re probably good in real life, too.

    That’s it. And it’s all good, my dude, I respect that you’re passionate about whatever it is you believe. One love.

  • http://deleted dagwaller

    No, you’re right. When you post (relatively) anonymously on websites, you open yourself up for people to attack you out of nowhere. I’m used to it, it’s all good!