NBA Heuristics

In my conversation on ESPN.com with Bill Simmons, the Sports Guy, I argued that many NBA General Managers—especially Isiah Thomas of the Knicks—could do a better job if they abandoned any pretense at all about exercising their own judgement. It would make more sense, I wrote, simply to draft or trade for players who attended either Duke or the University of Connecticut.

Here's part of the argument:

Let's say I'm so dumb about basketball that all I know is that the best college programs in the country are Duke and UConn, and so as a GM my rule is only draft and/or trade for the first and second best players, in any given year, from those two schools. So I fire all my scouts. I disband my front office, and basically say that I cede my basketball judgment to Jim Calhoun and Mike K. What's my team? It's some combination of Elton Brand, Emeka Okafor, Ben Gordon, Luol Deng, Shane Battier, Mike Dunleavy, Rip Hamilton, Corey Maggette, Jay Williams, Caron Butler, Donyell Marshall and Grant Hill—which is a really wonderful team. Now, of course, in the real world I couldn't get all those people, because lots of them were really high draft picks. But let's say I got Brand in a trade, after Chicago soured on him, and I was lucky enough to be in the lottery for Okafor. Maggette was a 13; Hamilton and Deng were 7s; and Butler was a 10—so at least some of them are doable, particularly since in off-years for Duke and UConn I can trade down and stockpile picks. Battier I wine and dine in the free agent market, because who wants to be stuck in Memphis? Ditto for Gordon, who, it seems, Chicago is thinking of moving anyway. Is that the best team in the league? No. It is better than the Knicks? Absolutely. The point is that clinging to a very simple rule of thumb here—that doesn't require knowing much about basketball—can leave you looking pretty smart.

A reader points out that I neglected to mention one of the very best UConn players—Ray Allen.  And given that, I think I was mistaken to say that an all-Duke/UConn team wouldn't be the best in the league. Think of it (assuming you could put together all those players). Okafor at center. Brand, Maggette and Battier at forward. Some combination of Grant Hill and Rip Hamilton and Ben Gordon and Ray Allen in the backcourt. Is there a better team in the league than that?

In defending the knowing-less-is-more position, I cited research by the psychologist Dan Goldstein. The relevant paper is here. It's definitely worth a read.

Goldstein's blog is here.

If you're interested in this line of research, Goldstein is part of a really fascinating band of psychologists interested in heuristics—that is, mental shortcuts that have the effect of helping us better navigate the world. Take a look here, for more.

Comments

[MR]

I presume that your strategy of picking only Duke and UConn players really must be a strategy of picking only the good players from those schools. Otherwise, you'd end up with a team full of Bobby Hurley, Alaa Abdelnaby, Khalid El-Amin, Brian Davis, etc.

So your strategy is to pick the best players from top schools? Groundbreaking! Not to mention the implausibility of getting 6 lottery picks on one roster. What makes you think you could do that when few if any other teams are able to?

Also, do you really think you'd have been able to get Elton Brand from the Bulls? You expressed hesitance in getting Eddy Curry from the Bulls because of your suspicion that anybody would give away such a talent unless there was truly a problem. I suspect you would have had the same suspicion about Brand.
You're right.

You don't really know what it takes to be a good GM in the NBA. Contrary to your instincts though, that's not an asset.

[Phil Gerbyshak]

I would agree, it seems as though picking off the best players from the best conferences is a great idea. If that were possible, this would indeed be a great strategy, barring injury.

And it couldn't be any worse than trying to assemble a fantasy roster of players that have HUGE upside, because they also have huge DOWNSIDE. For every Dwight Howard and Kevin Garnett, there are hundreds of Leon Smiths and Chris Washburns.

While admittedly it is harder to run an NBA team than this in practice, your theory is a good one.

[sean coon]

i like the concept, but in reality, the team you chose has real issues (unless you're talking about injury not being a factor).

you'd be thin at the 1, as jay williams is a medical wreck and ben gordon is the only pg remaining... which isn't a good thing.

grant hill is a stretch at the 2. he could play it, but his versatility at the 3 is where his value lies. but he's a medical wreck as well.

so that leaves you with two all-star scorers at the 2 and a one dimentional maggette at the 3 backing up an injury-prone hill. battier is nice all around, but he'll get no time at the 2 or 3.

okafur is nice and brand is a special player, but they're left with no one behind them. look at what happened to okafur this year (his injury destroyed my fantasy squad)

the funny thing is that i'd argue that this injured, out of position, team that is sorely lacking depth would *still* beat the knicks 4 out of 7. ;)

[Greg Newburn]

After reading "The Tipping Point" and "Blink," I kind of worship that you're a sports fan.

[Robby]

I don't see what the novel idea is here. Gladwell has not made the point that knowing less is more, but instead made the point that Coach Calhoun and Coach K know more basketball than Isiah Thomas and many other GMs do. The fact that some GMs appear to not be proper experts, does not mean that the use of expertise should be devalued.
Also, if knowing less is more, then why not cede you your judgment to the high school basketball scouts ? Why not say I will only take players who come straight from high school, b/c they must be the best if they don't need college? In that case you might end up with a team which consists of Kobe, Garnett, Jermaine O'Neal, Lebron, McGrady. Or you could end up with a team that consisted of Eddy Curry, Kwame Brown, Leon Smith, etc. So I don't see your point. Yes you could have ended up with a better team by simplifying the drafting process, but you also could have ended up with a much worse team.

[Mark Schraad]

There has been a notable shift (at least my me) shift in college coaching. From the old school version: That being that the coach is the great mind and primary architect, has a specific strategic methodology to building, coaching and competing, to a new approach. That being, that I (the coach) will accumulate the best possible athletes recruitable and assemble an appropriate strategy that best optimizes the competitive potential. I think this extends to the corporate, academic, and social world as well.

FYI - I am a great fan of your ability to communicate thoughts on complex issues to an audience that lacks your background research and detailed insight. It is a goal of mine to move in that direction. Thanks for sharing your leadership, skills and inspiration.

[Run Run Shaw]

Instead of drafting from the best schools, a GM with limited knowledge of the game could draft/trade/sign free agents based on basketball archetypes in what I call the 5 Deadly Venoms approach…for no other reason than it sounds cool.

If a GM just had a checklist for his starting 5 that read, The Scorer, The Passer, The Stopper, The Rebounder, and The Intimidator (aka Charles Oakley), I think he’d do OK.

When most teams with a shot to win the championship trot out on the floor (Spurs, Pistons) they have all 5 of those archetypes covered. Sometimes one player will cover more than one role (Duncan), but the team as a whole has all five represented.

This is the problem with the Knicks. They have an entire team made of only 1 Venom – The Scorer – and it’s streaky at best. They’re missing the other archetypes – the Stopper, the Passer, the Rebounder, and the Intimidator.

Now, if only they could merge with the Nets…they’d still stink.

[Jeff Marquis]

I agree with comments made above regarding the unlikely reality of (a) actually gathering these players all on the same team, (b) their skills translating into success on an NBA team (versus the college systems that they excelled in), and (c) their health remaining intact such that their expected success was realized in any sort of a consistent manner.

I think that picking the Knicks, led from a General Mgr perspective by Isiah Thomas, makes this theory sound more plausible than it probably is in reality.

Very simply, Thomas, based on his accumulated record as a GM, is horrible. He was a great player, and leader on the court, but he has proven definitively that he can't do this GM thing.

Malcolm, I think that you could pick better players than Mr. Thomas, no matter what criteria you use (alma mater, player's favorite color, etc.). Did you ever consider a career in... nevermind.

Glad you started the blog. Love the discourse. Cheers.

[Stefani at Lawlady.com]

I'm going to apply this principle to 1) shopping where my friend Linda shops (she's maniacal about spending), 2) eating where my friend Peter eats (dining out could be his second job), and 3) encouraging my clients to shoot for settlement offers that mirror what other clients worked hard to achieve. Why try so hard to figure out a good settlement when you can piggy back on the exhaustive efforts of other clients similar to you?

This is a great principle. Just hearing it reduces my stress. I see that there is hope for my fried brain syndrome-- that biproduct of living in the information age.

I read that scholarly article Malcolm sites. The language was convoluted and nearly impossible to understand. It reminds me of the source psychology texts I read in college, and only later understood once the self-help community dumbed them down to bite-sized concepts.

[Terry]

Interesting discussion. I'm not sure about Malcolm's GM theory but you gotta think just about anyone could do better than Isiah, sad as that is to say for an old Bad Boys fan.

Here's a question for you basketball geeks, somewhat off topic: a player picks up two fouls in the first quarter, or a third foul in the second and without fail he's sent to the bench. The theory I guess is that the coach wants all his options available at the end of the game when it most matters. Meanwhile the player sits. It doesn't seem to matter if he's the seventh man or if he's an all star.

Does this make sense? Wouldn't it be better to have the players you want on the floor to be playing as long as they can instead of sitting, avoiding potential future fouls? If it's true that he's most valuable at the end of the game, wouldn't it make more sense to sit him down only when he's got five?

[Rick]

Call me a homer, but here in Az we like to think Lute Olsen knows a thing or two about b-ball...

Along with your Blue Devils and Huskies maybe a few wildcats like Mike Bibby, Richard Jefferson, Gilbert Arenas, Andre Iguodala, Channing Frye, Jason Terry, or Luke Walton.

/go suns.

[Jay]

You would do better to pick some different schools. Duke's NBA output lags its college success. UNC Chapel Hill would be a better pick. Wake Forest would give you a current team of Tim Duncan, Chris Paul, Josh Howard, etc. Even Georgia Tech would be worthy of consideration. Plus a stealthier school that has exported more pro's would be easier to draft from in the middle of the draft.

[Jack Coleman]

"The Tipping Point" and "Blink" were among the most interesting books I've read in years (or heard on disc, to be more precise). It's great that you've started a blog, Malcolm, and I will be among your many regular visitors - Jack Coleman, Plymouth, Mass.

[Jim Caserta]

I checked out Goldstein's blog, and his example of which city is bigger - San Diego or San Antonio was troubling. Checking census records for the city limits - San Antonio is 1.236 million, while San Diego is 1.263. No dyslexia, just a 2% difference.
www.census.gov - city populations.

There is a definite advantage of having less information - it's easier to organize and you don't get swamped when it comes time to make a decision. However, it becomes hard (impossible) to justify decisions to bosses/fans when you have nothing but a simple rationale behind them.

[Andrew Kantor]

I had once thought of a variation on this -- perhaps a movie plot. Man becomes coach of NFL team, but is clueless. So he coaches based on what the columnists and commentators say. (Better yet, instant fan voting: "Press one to punt, press two to go fot it.")

It's really no different than any chess-playing computer. Put in enough data and you can pretty much play the statistics. In your example, you're using the success stats of the Duke and UConn coaches, which seems pretty smart.

[Gary]

I have to also wonder if getting Duke players would be a good idea from a Karma aspect, as they are well known to be the evil empire down here in the south. Just look at what has happened to Hurley, Hill, Jayson Williams, etc. There are no examples of major stars from UNC going to the pros and then having serious injury problems that hamper their future in the NBA like the Duke line up.

Plus Coach K is widely believed to be Satan. (Tongue firmly in cheek)

[Paul Musgrave]

Robby has it right: in this case, Gladwell isn't devaluing expertise; instead, he's properly valuing the expertise of others in relation to his own (assuming, of course, that Coach K hasn't had a remarkable run of good luck). In other words, this isn't the "Moneyball" method of picking good teams; it's more like the way that girls almost invariably won NCAA men's basketball pools in my high school--by simply choosing the top-ranked teams (or, in close calls, the team with the better record) to win in each instance, whilst the boys would rely on their false expertise (and I would always put Indiana in the Final Four, out of misplaced loyalty).

Andrew Kantor's comment would be more interesting, and a better test of the value of 'expertise'; maybe James Surowiecki could be persuaded to fund the experiment out of his royalties.

[FriendsterForum.com]

Nice Blog :)

www.FriendsterForum.com

[koreyel]

"What's my team? It's some combination of Elton Brand, Emeka Okafor, Ben Gordon, Luol Deng, Shane Battier, Mike Dunleavy, Rip Hamilton, Corey Maggette, Jay Williams, Caron Butler, Donyell Marshall and Grant Hill -- which is a really wonderful team."

Sorry... this team probably wouldn't even make the playoffs.

And if they did... Detroit or San Antonio would crush them in the first round.

Your heuristic is sort of like what Pitino tried when he became GM of the Celtics.

He ended up killing Boston for a generation.

No Malcolm... building a NBA championship caliber team is still black magic in a black box...

It doesn't succumb to little rules of thumb.

[kappy]

Malcom, you would be a better GM than Isiah, but so would half the country. I did like the Frye draft pick and the trade for Nate Robinson, however.

[Arnie McKinnis]

I agree with your position, but one thing that is taken into account is the "economics" of information and decision making. I truly believe that "less is more" and there is quick route to making decision as long as we are willing to be "wrong". It would be an interesting "study" to rollup all the $$ spent on research and information gathering (people, processes, technology, etc.) to justify decisions that could be made (like the one cited above) - what would economic impact really be? And are there "cultures" within the world that encourage this type of decision making? (I'm making an assumption that our western culture does not). Intereting to think about.

[Robert Schwartz]

No Ohioan could go without mentioning Danny Ferry. 10 year contract and all.

[Jim Jones]

Why is everyone getting hung up the example?

The point was that following a couple semmingly simplistic rules, if grounded properly, can yield successful, complex results.

Although Gladwell's example seems flawed (he seems to be deferring the complexity of choosing the right rules to the coaches of UCONN and Duke), the point is taken.

The idea of a few selective rules yielding complex results is what is called "emergence" in chaos theory.

For example, in the book "Complexity", Waldrop talks about a single water molecule. Water is H2O, two hydrogen atoms, one oxygen. Looking at that water molecule as a single entity, would anyone have predicted that if we assemble billions of these together, we would gain the property of "wetness"? Probably not.

So, following the guidance of chaos theory, maybe the NBA coach example could better be exemplified by a few selective statistical components?

If a coach always went with players who are between 6'7" and 6'11", with a 40 yard dash time of sub 4.6, historically how would they have faired?

Taking all of the players that met this criteria over the course of NBA history, we could look at their individual win/loss ratios and begin to see how effective these simply metrics would have played out.

Jim

[Julie]

Could it be that, statistically speaking, some team has to be at the bottom of the heap?

And all speculating to the contrary is yet another example of too much knowledge at work?

[Goldberg]

interesting, but, Malcolm, I think you need to mention that heuristics, while, yes, they can be a useful mental shortcut, they also can be detrimental. Think, for example, of the availability heuristic and risk assessment--think, esp., the day's after 9/11--i'm sure most thought it was fairly likely to die in a terrorist act or plane crash, when, in fact, either was only a very remote possiblity.

anyway, the point is that you clearly are aware of all this and you probably should make sure your readers are aware, too. Keep up the blogging! a very nice addition to my daily reads!