Knicks Morning News (2023.11.20)

Goga Bitadze is the Mitch of the Magic…

Too early to judge teams out west, imo…and in the East as well. Still lots of settling in with roster adjustments. The Knicks had more continuity than virtually anyone and they still seem to be figuring it out. I suspect that once teams like OKC, IND, MIN, ORL, HOU go around the league a couple of times, some adjustments will be made. The Lakers, Grizz, Clips, Suns and Warriors are all dealing with stuff that will probably stabilize and subside as the season goes on. But hey, who knows?

I love Lauri at$17.2M. Alas, he will be getting a massive extension, probably from Ainge.

I’m not sold on OKC or HOU being for real just yet. The Jazz started out 10-3 last year, then the league caught up. OTOH the Thunder were good last year and have added Chet, so this just might be the year they break out into the top-6 in the West. The NBA season is a long, brutal slog. Let’s see how they physically hold up, especially Chet.

“Man, Detroit has to be bummed. Cade is making last season’s RJ look like Kobe in comparison.

They’d be much better off staring Sasser and Ivey, maybe playing Cade as a 3/D-type forward for now.”

It’s certainly in their best interests to tank for another year. If they are actually trying to win, they’ve been hurt by Bojan being out (purposely?) and Duren getting injured.

As I’ve said previously, I’m not anywhere near out on Cade and won’t be for a long time. Let’s see what happens when he has a real team around him. Playing next to a confirmed bust like Ntilikina, er, Hayes hasn’t helped.

The following stats say that thus far Mark Williams is having a better season than Mitchell Robinson:

BPM 2.6 vs. 0.6
WS48: .210 vs. .183
PER: 24.6 vs. 17.1
VORP: 0.4 vs. 0.3
WP48 (phlogiston): .304 vs. .303

Yet no one who watched the two NYK vs. CHO games would conclude that Mark Williams was the better player.

This raises some pretty significant questions about the way advanced stats are currently used to make arguments about players. Not that this anything new. Some of this site’s most dubious arguments from 10-15 years ago (for example, Tyson Chandler vs. Dirk Nowitski) were grounded in the gospel of WP48…while those grounded in PER (a far more accurate statistic, and that is reasonably aligned with BPM) were dismissed out of hand.

And now we have RAPTOR, RAPM, and others….yet these seem to be flawed as well.

It would seem that the most telling stats would be of the situational and Synergy-type genre (which I believe LOVE Mitch)….and yet box-score based and on-off/plus-minus stats are still the basis for most player evaluation arguments…likely out of convenience, but at great cost to accuracy.

“(*) Kind of like D’Antoni with Marbury, only D’Antoni didn’t do it to show who’s boss; he did it because Marbury sucked.”

…and yet he was the best PG by far on that sucky team.

“It is and remains true at all times and in all places that these metrics are downstream from standard basketball data and standard basketball data are downstream from basketball skills and talent.”

My objection is solely that some folks (including me at times) use stats like BPM to trash one player but are quick to point out the stat’s flaws when it doesn’t align with the side they are defending re: another player.

“I will repeat my take on BPM: if it says you stink, you probably stink. If it says you’re elite, you’re probably elite. For just about everyone else, it’s a starting point and nothing more.”

Isn’t this true about PER as well? Or pretty much all of them?

As to EPM, Mitch is way down the list, below fellow Knicks RJ, Brunson and Julius, and below 17 players designated as Centers, including Chet, Sengun, Bitadze, Turner, Williams, Jalen Smith, Poeltl, Nurkic, Allen, and Duren…and the next C below Mitch is Hartenstein.

Mitch is 21st in BPM for players coded as a C and playing significant minutes. He is also 21st in PER.

However, he is 10th in the NBA in good ol’ reliable WP48! (3rd among starting C’s behind only Jokic and Mark Williams) and second in total wins produced due to having more minutes than Williams.

So the question of where Mitch stacks up vs. other C’s is about as much of a challenge to advanced stats as you could have. And as pointed out above, the other C that is ranked far below where one would expect is non other than tonight’s opponent, Rudy Gobert.

And to me, the question is: if you have to read between the lines to ascertain the impact on winning that each of these two similar (but far from identical) players have, why is this not true for other players?

RJ is a recent example. His steals, blocks and deflections numbers are low, yet a highly-regarded defensive-minded coach like Thibs sticks with him while keeping guys like Obi and Fournier on a much shorter leash.

I’m not trying to bash Mitch or praise RJ, but these are two interesting players to discuss in the context of the validity of various advanced stats. Are they outliers? Are they getting differential treatment?

Mitch is 67th in the NBA in DBPM (20th among C’s), while Gobert is 25th (10th among C’s).

However, Mitch is 19th in defensive EPM (tied for 6th with Bam among C’s) while Gobert is 25th (10th among C’s).

So it does seem like Defensive EPM captures at least some of the value that DBPM misses. But iHart ranks above both Mitch and Gobert by each metric.

PER is dumb. But WP yields results that are so dumb that they make the results of the metric I’m calling dumb look smart in comparison. Therefore I will continue to defend the much dumber metric over the one that closely correlates to today’s supposedly less dumb ones.

Okay.

“Missed FGs in PER is like giving bonuses for usage, which goes up with every missed shot (and even worse, for each TO)”

Even so, PER correlates very strongly with oft-cited BPM, which folks use sll the time here, because, well, it’s pretty good! Whereas, WP correlates with WS, which hardly anyone still uses here because, well, it’s pretty bad!

In other words, it’s way easier to correct for any inherent flaws in PER than it is for WP, except for extreme outliers. For the vast, vast majority of players, PER is the more accurate (or less inaccurate) metric. I don’t see how that can possibly be disputed.