Knicks Morning News (2026.04.17)

The type of conversation here probably reflects that. We used to use numbers a lot more, and look down at the eyetest, and it changed a lot recently.

I liked the article a lot, although at the end, it kind of boomeranged back on itself and devolved into “Better numbers are being built,” which sort of defeats the premise.

The critical sentence: “The moment you pick a number to measure success, behavior quietly reorganizes around that number.” When that number doesn’t really even measure anything important — e.g., Mike Brown’s “spray passes” ( insert clown emoji) — the problem worsens.

But bottom line: If you can’t eyetest — LeBron at Rucker Park with no stats kept — you can’t do this right.

(I’ll refine it a bit. If you watched a game in the 50s or 60s or 70s or 80s, you had the eyetest plus you had the boxscore. There is a wealth and treasure chest of information there — it wasn’t remotely difficult to know that Larry Bird was a better player than Danny Ainge or that Bernard King was a better player than Pat Cummings. So what we call “analytics” doesn’t build on nothing; it builds on a treasure trove. Which means the proper delta to look at isn’t “improvement over nothing,” it’s “improvement over treasure trove.”)

I paused on this too, because I’d like to think something else is needed, not just numbers. But numbers are all(?) we’ve got right now, so I guess the safe road for analysts would be to look for numbers that are the product of complexity, or even simpler come from earlier in the play, i.e. “screen assists”, or “the pass that leads to the pass”. I hope something richer that that emerges. Like the science of chemistry between players.

Maybe it’s out there already, I don’t know, but I’d like to see something like a “Burden” calculation of the cost of players who can’t (or don’t or won’t) generate shots and the cost to the offense of that. Like maybe start with eyeballing the number of times a player passes up a shot with 8 seconds or less left on the shot clock that the median player would take. Then measure the PPP the offense scores on shots after that pass-up, and dock the player for the delta between that and the team’s PPP as a whole.

That’s entirely back of the envelope and on the fly, but a player who just picks the best parts of the smorgasbord for himself and leaves the rest to others should pay a BB-ref-page price for that. “Quentin Grimes had a TS% of 592, but that was offset by 6.1 Burdens” would be the general idea.

Knicks wearing white, Hawks black for Saturday’s game. Excellent uniform matchup.

It’s hard to think of a player who is clearly good, but not ranked as such by either the box score AIOs or the impact-adjusted metrics. If you fail both tests you’re almost certainly just bad.

If you went back to 1972-73, let’s say, and just looked at the various players’ box scores, FG%, PPG, assists, RPG, AstPG, FT%, and team record and just ranked them all one to whatever based on that, you’d be something like 95% “right.” Maybe even more. You’d want to try to find a way to adjust for minutes played, obviously, and maybe try to go through the box scores and sample who the player likely guarded for 10-20 games and see how that player performed but other than that, you’re pretty much all the way to where you need to go.

The notion of players who are “good” but who don’t otherwise show it through this data list is not entirely specious, but it’s pretty specious nonetheless. Good basketball players are good at making baskets and assisting teammates who make baskets and getting rebounds and at not getting torched by their man on defense, and not clearly missing assignments on defense off the ball. This isn’t a big mystery. (Obviously there are players who take a lot of shots to make their baskets, but FG% weeds those out and helps us distinguish who’s actually good at making baskets.)

Jalen Brunson makes a lot of baskets and can create the opportunities for himself and his teammates to make baskets. Jalen Brunson is, pretty much ergo, an excellent basketball player. Eye test, standard box score.

Klay’s a 44.8% FG shooter on his career, peaking at .488, playing in a system tailor-made for his skills. That ain’t top-10 material. I’m not even sure where the “underrating” is. He was a key cog on a bunch of extremely good teams and was given a lot of credit for them.

(Just as a personal aside, looking at his BB-ref page, I’m shocked that he never had a single season with a TS% of .600 and his career number is only .571. I’d have guessed those numbers were quite a bit better.)

As a result of these kinds of empiric advancements, there’s absolutely a class of players who have become very wealthy that would’ve been ignored in the ’70s, and conversely a different class of players now playing overseas that would’ve become very wealthy in the ’70s.

Still a bunch of misses, start with Jokic and Giannis and on the flip side, Anthony Bennett. Then all the low second-rounders doing great work in the league. Then move to G-League guys who come into the association and play really well. Daniss Jenkins would be one. Plenty of others.

Not really sure there are any fewer “misses” today than at any other time. It’s an inexact art-science. Always has been, always will be. That makes it fun.

Were the all-NBA voters BITD confused by pointzzzz? Not really, although they did elect two players to the second team who shot under league average TS%.(*) First team average TS+ better then, TS+ for first two teams slightly better now.

BITD:

First team:

Abdul-Jabbar — 117
Archibald — 111
Havlicek — 101
Haywood — 107
West — 107

FIRST TEAM AVERAGE: 108.6

Second Team:

Barry — 102
Cowens — 96
The Incomparable Walt Frazier — 107
Hayes — 96
Maravich — 101

OVERALL AVERAGE: 104.5

Modern:

Giannis — 108
SGA — 111
Joker — 115
Spida — 100 (**)
Tatum — 101

FIRST TEAM AVERAGE: 107

Second team:

Brunson — 105
Curry — 107
Ant — 103
LBJ — 105
Mobley — 110

OVERALL AVERAGE: 106.5

It was before my real time, but Cowens played for a 67-win Boston team and had a rep as a hustlebunny, and didn’t really score much, so it couldn’t have been pointzzz. Kind of the same with Hayes, who only averaged 21 a game but 14.5 rebounds. That said, if someone wants to make the case that their efficiency wasn’t looked at right, there’s at least a superficial one to be made. Flip side, both were second-teamers and the first team that year was more efficient than the 2025 first team.

BIg picture, no real fooling, nothing really missed, and the voters got to “efficiency” without it being overtly measured but merely conceptualized and inferred.

(**) Isn’t Spida being on there at least kinda, sorta a bit of a pointzzz illusion?

(*) TS% wasn’t a number in and of itself then, of course, but instead a concept gotten to through other numbers. No voter had the ability to just look at a guy’s TS% as they do now.