Nylon Notebook: Found at the Finals

Jun 4, 2015; Oakland, CA, USA; Cleveland Cavaliers forward LeBron James (23) reacts after a play during the third quarter against the Golden State Warriors in game one of the NBA Finals at Oracle Arena. Mandatory Credit: Kyle Terada-USA TODAY Sports
Jun 4, 2015; Oakland, CA, USA; Cleveland Cavaliers forward LeBron James (23) reacts after a play during the third quarter against the Golden State Warriors in game one of the NBA Finals at Oracle Arena. Mandatory Credit: Kyle Terada-USA TODAY Sports /
facebooktwitterreddit
June 7, 2015; Oakland, CA, USA; Cleveland Cavaliers forward LeBron James (23) reacts after the Cavaliers 95-93 overtime victory against the Golden State Warriors in game two of the NBA Finals at Oracle Arena. Mandatory Credit: Bob Donnan-USA TODAY Sports
June 7, 2015; Oakland, CA, USA; Cleveland Cavaliers forward LeBron James (23) reacts after the Cavaliers 95-93 overtime victory against the Golden State Warriors in game two of the NBA Finals at Oracle Arena. Mandatory Credit: Bob Donnan-USA TODAY Sports /

In an NBA Finals edition of The Nylon Notebook, a few statistical tidbits through three games:

LeBron JAAAAAAAMMMESSSSS

As has been discussed, well, everywhere (but especially here and here) LeBron is doing many many things this Finals.  His 65.2% True Usage over the first three games (nearly 68% in both Game 1 & 2!) would be a close second to Russell Westbrook’s top regular season figure.[2. Though still well short of Westbrook’s absurd January/February run in terms of sheer play-ending involvement.] But he’s doing it in a style dramatically different than RWB’s face-breaking, verging-on-okay-now-completely-ok-I-got-it-back-out-of-control affect. Pick a cliche about imposing his will, controlling the pace, grinding down the clock, they all apply. James’ 12.7 minutes of average possession time is 4.5 minutes higher than John Wall’s season lead. His approximate[1. Due to the way nba.com displays much of the SportVU data as per game averages to only on decimal place, single game statistics are often close estimates rather than exact figures.] 26.8% time of possession while on the floor would also have topped the league. Most remarkably, I think, is that he’s doing this without actually turning the ball over that much. Out of every 100 plays in which he’s figured a major role (shooting, passing, turning it over), only 6 have been turnovers. Among high possession, high usage players, the closest analog in the regular season is Chris Paul. who also sported a True TO% of just over 6%.[3. Ty Lawson and Kemba Walker are two other primary ball-handler types who were extremely low turnover on this measure, but safe to say neither approaches James or Paul’s level of impact measured against those turnovers.]


Warriors’ Offensive Woes For Real?

Using the Expected Effective Field Goal Percentage method, Golden State is getting a similar mix of shots than they did during the regular season:

How much of the diminution is just “Make or Miss League” variance, how much is players such as Draymond Green and Harrison Barnes performing at a lower level for either psychological or physical reasons and how much is Cleveland’s defense is a question somewhat beyond our ability to totally answer. But 538’s Neil Paine and our own Justin Willard make some headway towards unpacking things. An additional point is that for whatever, a mix of Cleveland’s defense and Golden State’s impatience with the slow pace perhaps, Golden State is passing less, and taking a far higher percentage of their shots in self-created situations in the finals:

The Warriors have an eFG% of 43% on self-created shots, against 49.4% on other shots in the Finals, so the 8 or so extra self-created shots per game aren’t helping, especially with all the final scores being so close.


“NBA Open”

With all the talk about “contested” or “wide open” threes, I made this chart as a guide to what’s actually going on:

open3s graphic
open3s graphic /

Especially important is the far right, because “NBA wide open” and “normal people basketball wide open” are not at all the same thing – to be an NBA level shooter, not being bothered by a guy sort of close to you but nowhere near close enough to actually block the shot is part of the job description. To put it another way, this (from Game 2) is an NBA open-to-wide-open look for Steph Curry:

Stephopen
Stephopen /

He missed the shot, so Dellavedova is getting credit for shutting him down. But in the best of times, even a shooter as great as Curry missed half or more of his wide open threes. This is a big part of the reason “crediting” three point defense is so tricky – if he’s going to miss a lot anyway, how do we know if the defender affected a single shot or small set of shots?


LeBron vs. the Bigs

In talking about what the Warriors can do lineup wise, someone on Twitter asked friend of the program Kevin Pelton an interesting question:

I looked into this a little, and while deriving overall offensive rating wasn’t quite possible to do quickly with available data, I could look into field goal shooting. Using a combination of detailed shot logs and lineup data[4. Thanks as always to TNC’s data sourcing wizard, Darryl Blackport.], I was able to isolate shots taken by every player in the league with and without a “traditional rim protector” on the court defensively. I defined a rim protector as a guy who defended at least 8 field goals per 36 at the rim per SportVU’s “Defensive Impact” stats[5. I’ve previously found a league average big defends right around 8.7 attempts per 36] at a percentage of 51% or lower. I admit to cheating on the 51%, I was going to use 50%, but opponents scored at a 50.9% rate on Tyson Chandler at the rim and if Chandler doesn’t count as a rim protector you’re probably doing it wrong.

In any event, on aggregate about 63% of all shots in the NBA during the 2014/15 regular season were taken against a team with a “rim protector” on the floor. Effective Field Goal Percentages were 2% lower with a rim protector on the floor than without one (48.9% o 50.9%), FG% at the rim was 3.8% lower (60.2% vs. 64%), while a slightly higher percentage of shots were taken within 5 feet of the rim with no paint patroller in the game – 26.4% of all FGA with and 27.5% without.

While it’s not totally and apples-to-apples comparison because of Cleveland’s different personnel and resulting LeBron dominant style of play, during the regular season he wasn’t notably more effective than the average player against non-rim protecting lineups. He shot 2.3% better overall, 4.6% better at the rim while his percentage of attempts at the rim rose from 23.9% to 25.9%. This possibly suggests that LeBron’s effectiveness shouldn’t on its own prevent Golden State going small, sans Bogut or Ezeli for the rest of the series. It still might be a bad idea with Cleveland’s offensive rebounding, Green’s general lack of production and possible back issues as well as the usefulness of Bogut[5. Or Ezeli perhaps.] as a screener however.

Interestingly, this analysis does reveal a possible cause of Atlanta’s struggles in the Eastern Conference Finals. Even during the regular season when their offense was humming, four of Atlanta’s five starters saw more than twice the average decline in eFG with a rim protector facing them on defense, with only Jeff Teague not being majorly affected. Of course, the Cavs negated Teague through other means, namely daring him to shoot. With Timo Mozgov and an invigorated Tristan Thompson blotting out the sun, maybe Atlanta couldn’t get quite the same inside-out action whic allowed them to feast from the arc during the season?

Looking to how the Warriors fair in this analysis, Barnes had one of the biggest differentials in the league, shooting 9.6% worse with a rim protector on the floor, while Curry was 5.5% less efficient. In a parallel to the Atlanta series, Green was actually quite good against lineups with a paint space eater, largely by getting to and finishing at the rim at a much higher rate. Much like Teague, Cleveland seems to have negated this by simply ignoring Green on the outside and daring him to hit alone-in-the-gym jumpers or finish floaters over Mozgov, neither of which has happened much yet.


Repeat After Me, The Finals Are Not A Referendum On Anything

Finally, wanted to reiterate a point I made on twitter yesterday:

As I’ve talked about before, the study of the numbers can describe the situations likely to lead to good outcomes in terms of the shots a team takes or allows. How to get from a team’s personnel to those desired outcomes is where the “basketball qua basketball” part comes in, and we’re nowhere close to “solving” the game to the point where that kind of expertise isn’t hugely important. Plus, all the strategy and tactics doesn’t much matter if LeBron stands up next to the mountain of statistical accomplishment and chops it down with the edge of his hand.