Limitations of Moreyball: Analysis of the Rockets Clutch Performances

I do not agree that a supposed inability to find "clutch" performers is a "limitation" of "Moreyball." It appears that the OP misunderstands where the Rockets stand on this topic.

The most important part of the Morey clip quoted by the OP is this: Morey didn't say "clutchness" was not important or doesn't exist. He said that he wouldn't pay extra money for a player who put up "clutch" stats because "it has not been shown that this can be replicated" from year to year. Cuban responded by saying they do pay for "clutch" performers because they've found certain guys like Jason Kidd who did play differently in "the clutch"-- whether it's defined by the last 3 minutes of a game or by when a team is desperately trying to make the playoffs.

I don't think there is a fundamental disagreement between the two:

What Morey said is true in general-- if you look at the entire NBA population, a player's shooting at a high efficiency in the last few minutes of a tight game for one season doesn't predict that he will do so the next season. So, it's mostly a foolish to pay a guy $2M/year extra because he had, say, a 60% TS% in the clutch last year. The issue here is that the sample size for most guys is just too small. Take a look a Chandler Parsons' and Jeremy Lin's shot charts in the OP, for example, we are not talking about many shots from each of these spots so much of the shooting %s are a matter of luck.

However, there are a few players who obviously do perform at a higher level in "the clutch" and have proven so during the course of his career. Jason Kidd is one of them and Chris Paul is another. It is notable that Cuban's example, Kidd, was an older guy-- we are talking years and years of data, so the sample size is large enough for the numbers to not just be a matter of luck.

Morey doesn't ignore "clutch" performance. His methods are also not unable to be applied to "clutch" stats. In fact, he had already analyzes these numbers in order to determine whether past stats in the clutch predicts the future for a player. In the case of guys who had played a ton of "clutch" minutes, he most likely can reach similar conclusions to Cuban and understand that, say, Chris Paul tends to take over during the last few minutes if the game is close based on years of data. The issue with most guys, especially young guys like Parsons, Lin, Harden and others on the Rockets recent rosters, is that they have played relatively too few "clutch time" minutes for that sample size to be meaningful predictors of the future.

Also, it's not like the Rockets have gone out of its way to avoid taking on stars who have demonstrated "clutchness"-- they did chase Chris Paul and Melo is another guy who had put up great "clutch" stats that Morey chased. When it comes to elite players, you just go with whoever luck allows you to have.

 

Great post. Thanks for the effort you've put in!

Which part of Aus are you from?

 

"Clutch play," in reality is just the ability to execute your offense and get the best shot possible on every possession when the defense is ramped up to the max. If a team isn't good in the clutch, it simply means their playmakers don't make very good decisions or run the offense well. That's it. No magic to it.

So guess who the best closer in the league is, son? It's Chris Paul. And it ain't no coincidence that he just happens to be the best court general in the game either. His clutchness and his ability to run an offense go hand in hand. Whiiiiich, brings me to the point that in crunch time the Rockets offense is limited because basic play execution isn't really an aspect of the team philosophy. They're just making it up on the fly out there while the other team has a sound defensive strategy, so that's knock number 1.

Knock number 2 is that the Rockets make Harden the primary playmaker in the 4th when Lin is the best court general on the team. Just because Harden is the best on the team doesn't mean he should be handling playmaking duties, especially when he's not the best playmaker on the team and especially when the defense is going to be looking to pressure him with multiple defenders.

The numbers in OP to my lame ass suggest what I've always seen from watching close Rockets games from the past year, which is to say that their clutch time offense suffers from a lack of actual offensive strategy and that the go-to-Harden-iso in the 4th quarter isn't a good strategy. Sure, Parsons seems to be cold blooded from 3 but he has some ball handling but there's a reason that he doesn't have any clutch-drives-to-the-hole and it's because both Lin and Harden are better playmakers and drivers than him, so Parsons really should only be shooting J's in the clutch. No problem with him there. Lin is a gangsta in the clutch as well, and Harden has been clutch in the past when playing a secondary/third offensive type role in OKC but has not been great as the first option. Why? Because the offensive sets in clutch time when Harden got the rock in OKC revolved around Harden either running the PNR or him spotting up for a 3 pointer/J after Westbrook/Durant made a play. In other words, the offensive set during the clutch in OKC was actually BALANCED and Harden got catch-and-shoot options (something he excels at) while in Houston Harden doesn't receive any catch and shoot opportunities in the clutch because he's always asked to pound the rock. He then shoots a contested 3, or drives hoping to draw the foul. Neither of these options are even close to being better than a catch and shoot shot by Harden and the options are made worse if they're shut down because Harden doesn't have a reliable mid-range game.

Whiiiiiiiiiiiich, brings me to the point once again that Lin should be the primary playmaker in the clutch, as he is composed enough to run the offense well, and instead of completely taking away Harden's clutch ability to hit catch-and-shoot J's, you give that back to him while giving the team a more balanced offense.

 

Not really. Luck is just a culmination of all things that cannot be repeated with success. Harden being a 25ppg scorer is not luck. He can put up numbers in that range repeatedly game after game. Harden making an off balanced jumper between Leonard and Duncan to make a game winner is luck. Have him try that over and over, and he'll maybe make 2 out of 10 times.

 

"Moreyball" doesn't come with any accompanying defensive philosophy like Popovich's '48 minutes of hell' or the 'white line defense'. Players are only supposed to defend on a 1-on-1 basis, and get lost in defensive rotations, over help, or don't help enough.

 

Here's how the term "clutch play" can become subjective.

You can create a case that Clutch play is comparable to something as blind as luck as Meh states below-

Then you can take the term Clutch and apply it to analyze player performance when they are faced up against the better team defenses & matched up with the top individual defenders in the league to get a good indication of how well they perform when the defense (individual and team) is really good.... which doesn't tell you everything, but IMO it tells you a lot.

You can also analyze stats for players with experience based on how a players produces in the playoffs vs. the regular season & comparable players in the playoffs at the same time.

As Morey states in the previously referenced video from the 2010 Sloan Conference, they absolutely were evaluating players on this notion, and in 3 years of the league continuing to advance, I'm sure there is even more data out there to analyze on this topic.

So while Cuban say's absolutely we measure "clutch-ness" and Morey say's they throw away the notion of "clutch-stats" because of their sample size, they are still both doing the exact same thing, just calling it different things.

 

Harden didnt shoot the ball well last year, thats what the numbers say and the eye test. In prior years, he was a more accurate shooter. I think we can expect a better and far more consistent year out of all the Rockets main players (Lin, Harden, Parson, Jones, Dmo)

 

Oh man... now the fun begins... the good ole Lin vs. Harden at the end of games argument.

All I will say is this- The Rockets and generally the rest of the NBA's most effective play has been of recent memory the "Pick and Roll"

So who would we say is the more effective pick and roll player???... both are pretty good, but its pretty obvious to me that the best pick and roll guard in any situation is James Harden. Add to the fact his ability to get to the free throw line, and finish with contact, and you have yourself your pick and roll guard of choice.

The whole run the offense argument is irrelevant if a big man is coming up to give whichever guard a pick.... if that's the "play" the Rockets are running... Harden is the guy I want over Lin.

 

I think most credible ppl agree that Harden was the best player for the Rockets last year but he also has many flaws. He turned the ball, alot. He did not shoot a high percentage from the field and he was the worse defender on the team by far. Now I can say these things because they are honest. Lin was the better passer and floor general but that was his job. I still dont understand ppl's point in fighting over who is better. They are both talented but still flawed and inconsistent. But making exaggerated statements that Harden is amazing and Lin sucks or Harden is a ball hog and Lin is a victim are just ridiculous. For the old heads this arguement reminds me of the Kenny Smith vs Sam Cassell debate which lasted until Sam and Kenny were both gone from the team. Instead of just enjoying 1 of the top 5 guard combos in the league, we get stupid arguements about who is better or calls to replace Lin with backup pt guards or rookies who havent even played a damn game in the league. Just enjoy the team cause this is better than watching Moochie Norris and Kelvin Cato play.

-Drop the mic

 

I had forgotten how white our team was that year.

 

Oh yes I remember this well.

 

43% FG for a shooting guard is not that good, no matter the excuse. He will improve because he's only 24 and most NBA guards improve their shot over time. He will take smarter shots and will become more consistent which is really the mark of greatness.

 

Do not make the mistake of looking at FG%. You need to look at TS% instead.

 

I think you misunderstood me. I'm not being subjective at all here, because I based my post upon the OP's definition of "clutch". Which has the following characteristics.

1. It's based on a 1-yr sample of 2012-2013
2. It's based on the idea that clutch is defined by shooting percentages(i.e. not a culmination of team offense and defense)
3. It's based on <5pt differential and <5minutes 4th quarter/OT

I used the above definition of clutch when I formed my response. I'm not using Morey's internal definition nor Cuban's internal definition, which I'm sure is quite different and much more detailed.

Do I believe the Rockets have detailed situational stats? Of course I do. But I bet that is much more than just shooting. It probably involves offense and defense, shooting, passing, FT% of players, how teams are more likely to play during end-of-game situations and the best counters, etc. Do I believe the Rockets look at Chandler Parsons' shooting during clutch situations and make any conclusions based on that? No.