Limitations of Moreyball: Analysis of the Rockets Clutch Performances

To clarify, I did not think your post was being subjective at all... just that the term "clutch" and how it might be interpreted is.

You can label a certain set of statistics as "clutch stats" like some of the situational stats we previously noted, but painting it with a broad stroke without explaining the situations you are evaluating is subjective.

Player or team performance at the end of a game with the score close is a situational stat that can be evaluated, but much like general +/- and others alike, there are certainly problems with it as comprised, and cannot be the one, single over-riding statistic that interprets success or failure.

 

44% is average, nothing impressive but this is also not accurate since it is also lower because of the 6 threes Harden takes and also does not account for the free throws he makes. Why do you think Durant and Harden are such great scorers, free throws.

And the only reason it is down is because he always has the defense collapse on him cause he is the man in houston. In OKC he was shooting 49% from the field and had a 66%TS (GODLY).

 

even though it was not perfect, such as it does not account for how defenses collapse on players and some other stuff, it was a very good read

 

In all seriousness, I have turds that are bigger then him. He's kinda cute in a feminine kinda way, NO HOMO.

 

still think that the midrange shot is needed at times despite morey only wanting 3s all the time.

 

Yeah I don't think he ONLY wants 3's. I think that's a little overblown. I do think he wants to maximize high percentage shots though. So if you have a player who makes midrange jumpers at a high-rate... it may be a high-percentage shot for that player to take a midrange jumper. Let's put it this way... If LMA comes to town, it's gonna be fire at will, efficiency be damned.

 

The only limitation of Moreyball is that you can only win one champsion a year with it.

 

To be absolutely clear, below is the actual value of basketball shots.

Shot %Made Value Outcome in Points
Foul Shots* 75.4% ** 1.51
< 8 feet 55.3 % 2 1.11
8-16 Feet 39.50% 2 0.79
16-24 Feet 39.50% 2 0.79
3 Point Shots 36.22% 3 1.08


* For the sake of simplicity I am ignoring technical free throws and and-1's.
** I used last seasons team free throw percentage. Otherwise I used NBA averages.

So, why should a team ever shoot a mid-ranged shot unless the player is absolutely wide open?

 

Can i be honest with you. I keep reading your username as Voice of Anus.

 

Nice analysis,

Can do you the same clutch analysis for this guy?

Spoiler

 

WHOA WTF, i thought that his name was voice of anus the whole time..

ahaha

 

Yea see that might be a bigger than anything u find researching the Rockets

 

I think DM touched on he luck factor a bit during last season when he made a comment that he didn't like to play a game close to the buzzer since a lucky shot, tipped ball, bad bounce, bad call etc.. could win the game. i.e. the last second shot in Houston by Monte Ellis last season.

 

when you add the word "City" to that term, it means something more powerful.

 

Thanks for the read. I enjoy all the stats and the graphics. :grin:

However, I believe TS%, FG%, pts, are not good measurements of clutch play. The problem has everything to do with sample size as other posters have point out. The defensive rotation, offensive scheme, shot clock management, shot selection, coaching decision, home vs away, Manu sausage, etc., will all effect the shot. There's not enough shots to balance out the variables.

Couple examples from last NBA final
1. Lebron James defending Tony Parker at end of games.
2. Removal of Duncan that led to offensive reb that turn into Ray Allen 3.

These plays has less to do with the shot but more with match up and offensive/defensive schemes. I understand Ray Allen's shot was CLUTCH, but was Bosh's rebound clutch? Or the lucky bounce off the ring clutch? Does Ray's FG% encapsulate all the other things that happened in that play? More comments were made about Pop's decision over Duncan then Ray's amazing shot.

If you want to look at shots at clutch time, I think looking at FT % may be the way to go. At least the shot itself is consistent, so its less effected by variables. Harden shot about same %, Lin shot 4% better, Parsons shot 6% worse but he only had 12 shots, Asik went from 56% to 100% (okay he only shot 2), and Dwight shot 2% better. Lebron is 1 % better (I got these # from 82games). This made me think that, if clutch free throw % normalize over a large sample to regular free throw %, FG% during clutch should normalize if a large enough sample is available.

I'm not saying clutch can't be analyzed. In fact, I believe "clutchness" can be looked at from different angle.

In my opinion, the last 5 min of close games has more to do with possessions. If you are down, you need more possessions to catch up. And if you are up, more possessions = win secured. Which gives us 3 things too look at: shots made, rebound, turnovers.

OP already gave us a look at the shots made. It will be interesting to look at rebound, and turnovers (i will group ball handling here).

I'm not sure how good "hands rating" is on 82games. But lets look at the two stars last NBA final.

Lebron: 23.7 hands rating normally, and 65.4 hands rating during clutch.
Parker: 26.5 hands rating normally, and 52.6 hands rating during clutch.

Lets look at Harden and Lin:

Harden: 17.2 normally, and 14.0 during clutch.
Lin: 24.2 normally, and 10.0 during clutch.

Now, I believe these number suffers from sample size as well. But, it is clutch just to keep the ball and not turn it over. And this is why I think Lebron and Parker are so good. They are like baseball closer. You just need to stay ahead.... so don't turn the ball over.

Same argument can be made about the flip side. If a player can create turn overs, its very clutch play. And if you look at our team from this angle, it kinda makes sense that coaches played Bev and Tony over Lin at clutch time. Coaches wanted to get more ball possession before worrying about offense.

Sorry about this long post. I'm bored with work . Anyways, it'll be interesting if someone can crunch the numbers and find out who takes of the ball, rebound at a good clip in addition of shooting.

I just think the last 5 min of a close game is more about shots being clutch. Its about that star rebounder, that star ball handle, and of course, the star shooter.

Anyways, thanks OP. Made me used my brain today =)

 

I believe mid-range jumpers can be very much valuable because by being a few steps closer to the basket naturally more opportunities will present themselves to cut towards the hoop for high percentage shots or pump-fake and get to the basket with one or two straight dribbles, which is about as many precise dribbles as most bigs have in them.

Of course it really comes down to specific players and their skill-set. I don't think it would be very wise at all for Tim Duncan, Chris Bosh or Al Horford -- all elite mid-range shooters -- to suddenly pass up on open mid-ranger jumpers and instead swing the ball or start hoisting 3s, even if they were able to hit a higher eFG% from 3 than from mid-range. The (threat of the) mid-range jumper opens their driving game and directly leads to high % scoring opportunities near the basket, which does not get reflected in value of shots based on shot location.

In many other (most) cases, where players are only able to connect on a below-average or even average % from mid-range, I very much would agree that mid-range jumpers are best to avoid, especially if such players can hit 3s on a reasonable %.
Still, I would not dismiss the value of the mid-range jumper completely. eFG% and PPP only tell a part of the story.

 

Since dunks are the most efficient shot, why doesn't the team only dunk the ball? Because when you only do one thing or you only do two things, you become easier to defend and your efficiency on those shots go down. A healthy offense takes mid-range shots to keep the defense guessing.

 

Towards the end of a game, the value of midrange shots goes up relative to the 3 point shot, because you are more likely to hit the midrange than the 3 pointer, and you don't have enough time to shoot enough 3 pointers to make up the difference in odds.

Most obvious example, last shot of the game, down by 1 point, a midrange is better than a 3.