Very surprising image to be sure.
Quite a few surprises here. For one, SpursTalk whipping boy Jakob Poeltl is the Spurs' best player, according to the metric, largely on the strength of his strong defensive rating. Others like Rudy Gay, Patty Mills, and Devin Vassell also perform exceptionally well.
Meanwhile, SpursTalk favorites such as Keldon Johnson have a negative rating, along with other players from the youth movement like Dejounte Murray and Lonnie Walker, who is the team's worst player. Aldridge is a major liability as well.
Obviously, all the negative rated players (minus DeRozan) are starters, while all the positive rated players are bench players. So it's consistent with the idea that the Spurs' starters regularly get outscored, and then the bench comes in to make up the difference.
Last edited by Uriel; 01-13-2021 at 11:41 AM.
Very surprising image to be sure.
I knew I was right about Patty Mills' and LMA's defense. According to the metric in this image, it is non-existent. That's how bad their defense is.
Last edited by Dejounte; 01-13-2021 at 11:26 AM.
LA def raptor is -0.7 but it is the second worst among full time starting center, easily drags the starting lineups defense. The center affects the defense more than any other position, he is in charge of communicating switches/rotations, deters drives and erases mistakes.
I've never found 538's basketball stats to be useful at all. They certainly aren't so strong to go against intuition.
Sample size still way too small to make much out of plus/minus stats this season. I mean, if we're going to put any stock in plus/minus right now, Vassell looks like the second-coming of Jordan (+24 net rating) and lock for DPOY (92 defensive rating compared to a league average of ~108). In the box score RAPTOR, Poeltl's getting a boost right now because of the low FG% of the man he's defending. That's good but, again, tiny sample size and part of Poeltl's issues right now is he just isn't challenging enough shots in general (as evidenced by his low block totals).
With the small sample size, it makes sense that all the bench players are clumped together and all the starters are clumped together. Revisit these stats in another 20-30 games and then they will begin to be somewhat meaningful. But, even then, part of Pop's coaching strategy is prioritize the strength of his bench, which is why Spurs bench players have had outstanding plus/minus number for the last 20+ years no matter who is coming off the bench, tbh.
I had issues with their older CARMELO rating system, but the newer RAPTOR one that they've used since last season is probably the best free, publicly available NBA advanced statistic. The methodology is at https://fivethirtyeight.com/features...-metric-works/ and goes into the various counting stats, on-off, play-by-play, and player tracking that go into the rating. timvp's point about sample size is valid though. However, I do think it can be used to show high level trends on the team, but it's not necessarily useful for comparing players to each other at this point in time.
Even that they get wrong.
Then it's time to double-check that metric.
"Surprise" is not the word for that one. ROFL is the word for it. With 5 Poeltls on the floor, the Spurs would lose games by scores like 300 to 2. Or maybe 300 to 0.
Yeah. Any metric that has that soft pos as our best player needs to be recalibrated
Will always LOL at "Advanced" stats/metrics. Just fodder numbers trying to make sports seem more complicated as it is and make monkey boys feel knowledgeable.
538 is absolutely terrible
Yeah, I like RAPTOR. At the very least, it's unique and gives a different perspective than other advanced stats due to the use of player tracking and niche stats that aren't factored in elsewhere. It's only glaring limitation is sample size -- but that's true of everything else other than RPM (but RPM cheats a bit by using rigid baselines, tbh).
I'd say that RAPTOR actually correlates with reality well. For example, last year's box score RAPTOR (the half of the equation that uses player tracking, etc.) found that the three worst defenders on the team were Belinelli, Forbes and DeRozan and the two best were White and Poeltl. That's pretty accurate. Compare that to the unsanitized plus/minus that found that Belinelli and White were equal defensively and Mills was a top two defender on the team -- and box score RAPTOR looks great by comparison.
Rewind to the 2017-18 season and all advanced stats at the time pointed to Dejounte Murray being the reason why the team's defense was so good -- so much so that those advanced stats resulted in Murray being named to an All-Defense team. Now we know in hindsight that while Murray can be a good defender, that team wasn't great on defense because of him. Using box score RAPTOR, it credits Danny Green and Kyle Anderson as being the two main reasons why that 2017-18 team was so good on defense ... and that, again, is closer to reality.
Which tells you that this stat is not as accurate as you think it is
Not that Poeltl hasn't been playing like , but you can't really win any games with 5 centers on the floor for one team, tbh. There's better ways to say he sucks
I agree with Timvp. RAPTOR is gold standard. It combines both box scores metrics with plus minus stats and ridge regressions.. so it's fairly robust. And Neil Paine himself suggests that it is too early to take note of these numbers as the sample size is small. Add to the fact that the Spurs play bunched line-ups a lot (starters in full and bench players in full) ..the low sample size can show up some individuals poorly if they are bunched with other poor players much more. Ergo Johnson and Walker have got rhe rough end of the stick. I expect their numbers to pick up over time, especially Walker's as he gets involved in more plays and scores more.
I was not writing about winning. I was writing about an annihilation. Because Poeltl doesn't score. Do you understand?
If you want to know why Lonnie's defensive score is low, watch him switch vs watching Vassell switch. Vassell is so quick & decisive about when to switch & when not to & essentially makes decisions for the defender he is switching with. Lonnie may improve in this area, but it seems as though quick decisions are instinctive more than anything.
None of these numbers are surprising. The bench has had better advanced stats in most games so far this year, and Demar's effect on winning has been pretty much the same as it's been his entire career-- relatively minimal in relation to his minutes & max salary. Superstar first options like LBJ, Jokic, Doncic, Durant, etc are consistently the most productive players on their teams, and their teams pull away in terms of scoring when they're on the floor... DeRozan has never been that guy.
No surprise on Poeltl. Just watch his defense, watch ballhandlers shy away from attacking him and watch him point out assignments and switches.
Yeah, and I was saying it's hyperbolic, and that a team of 5 centers (who don't even shoot) is unlikely to score many points at all, not just a team of Poeltl's, making it a bad way to say that he sucks. Do you understand?
Guy, that's what I wrote, about Poeltl.
Are you not swift? Do you just not understand things?
538 is a liberal hack
Surprising how defensive minded Rudy Gay is according to his 538 stats. Better play him more in the 4th quarter when we need stops, eh?
There are currently 1 users browsing this thread. (0 members and 1 guests)