Game Nineteen - Raptors at Grizzlies - Wed. Nov. 24th 8:00pm on TSN

[planetmars]

Precious is really important for our defense to work. I don't feel like chasing stats, but I suspect he's our best defender.. even better than OG IMO.

[Jclaw]

[DanH]

Precious is really important for our defense to work. I don't feel like chasing stats, but I suspect he's our best defender.. even better than OG IMO.

It's way too early for any impact stats to be reliable as the on/offs are mostly noise at this point, but purely for interest's sake, using DARKO's DPM and dunks-and-threes' EPM (EPM has a bigger spread than DPM this early in the season):

Precious +2.1 D-EPM, -0.6 D-DPM
Birch -0.9 D-EPM, +0.9 D-DPM
FVV +1.7 D-EPM, +0.8 D-DPM
Pascal -0.7 D-EPM, +0.6 D-DPM
Anunoby +0.5 D-EPM, +0.5 D-DPM
GTJ +1.2 D-EPM, -1.0 D-DPM

As you can see, wildly different results across the two models. Has to do with how much each depends on small sample on/offs, individual counting stats as priors, previous season data as priors, etc. This early in the season it's mostly noise.

But I have to admit, those DPM numbers make a lot more sense to my eyes than the EPM numbers this early. I personally have not been impressed at all with Precious' defence in general, though he's had a few useful stretches (was very solid in the second half last night for example). Same goes for GTJ - he's generating a lot of steals individually but gets blown by more than anyone and often misses his closeouts.

One thing that makes me laugh is that both models agree with my eye test that Svi is an absolute disaster defensively. I can't wait until we are fully healthy and Yuta's minutes (behind OG, etc) force Svi and Boucher both either out of the lineup or to spot minutes.

[Jclaw]

I think his by quarter slash of 0/0/17/0 might sum up the Heat fan experience with him last year. But there's something there. Looking forward to it.

[planetmars]

It's way too early for any impact stats to be reliable as the on/offs are mostly noise at this point, but purely for interest's sake, using DARKO's DPM and dunks-and-threes' EPM (EPM has a bigger spread than DPM this early in the season):

Precious +2.1 D-EPM, -0.6 D-DPM
Birch -0.9 D-EPM, +0.9 D-DPM
FVV +1.7 D-EPM, +0.8 D-DPM
Pascal -0.7 D-EPM, +0.6 D-DPM
Anunoby +0.5 D-EPM, +0.5 D-DPM
GTJ +1.2 D-EPM, -1.0 D-DPM

As you can see, wildly different results across the two models. Has to do with how much each depends on small sample on/offs, individual counting stats as priors, previous season data as priors, etc. This early in the season it's mostly noise.

But I have to admit, those DPM numbers make a lot more sense to my eyes than the EPM numbers this early. I personally have not been impressed at all with Precious' defence in general, though he's had a few useful stretches (was very solid in the second half last night for example). Same goes for GTJ - he's generating a lot of steals individually but gets blown by more than anyone and often misses his closeouts.

One thing that makes me laugh is that both models agree with my eye test that Svi is an absolute disaster defensively. I can't wait until we are fully healthy and Yuta's minutes (behind OG, etc) force Svi and Boucher both either out of the lineup or to spot minutes.

It's funny because to my eyes, the EPM stats look better. And it's not just Precious. I feel like both Pascal and Khem have both been negatives on defense so far this season. And yet they are graded as positive in the DPM stats.

DFG% isn't everything, but opponents are scoring at 53.7% against him compared to 43% against Precious.
https://www.nba.com/stats/players/de...mID=1610612761

Precious is a terrible offensive player (not including his 3rd quarter yesterday of course).. but I feel like he's one of our best defenders. Will be watching more carefully as the season goes along.

[DanH]

It's funny because to my eyes, the EPM stats look better. And it's not just Precious. I feel like both Pascal and Khem have both been negatives on defense so far this season. And yet they are graded as positive in the DPM stats.

DFG% isn't everything, but opponents are scoring at 53.7% against him compared to 43% against Precious.
https://www.nba.com/stats/players/de...mID=1610612761

Precious is a terrible offensive player (not including his 3rd quarter yesterday of course).. but I feel like he's one of our best defenders. Will be watching more carefully as the season goes along.

DFG% is less than "not everything", it's nothing. Even at the rim it's mostly noise over a full season except the top and bottom 5% of the league. Nevermind over smaller samples. And away from the rim it is entirely noise; no matter the sample size, it never becomes predictive.

But I agree, will be watching Precious (and everyone) carefully on defence. They have a long way to go to get their defence fixed and will be interesting to see if anyone stands out in terms of improvement from this point on.

[golden]

No centers.. means a lot more Boucher and Pascal. And they have Steven Adams. So yeah this will be great!

Standard RR refrain after most losses: if only we had a REAL center.... like Steven Adams. lol. Did he even play last night?

[slaw]

It's funny because to my eyes, the EPM stats look better. And it's not just Precious. I feel like both Pascal and Khem have both been negatives on defense so far this season. And yet they are graded as positive in the DPM stats.

Part of the issue evaluating defense is that you really need to understand both the defensive schemes and what the individual players are being asked to do in the applicable schemes. You then need to look at each unique game being played and the particular sequences and patterns.

I'm sure the models have some value but there's a reason why teams break down games into very small time periods and try to figure out what is happening in each little segment. It's more telling than aggregating numbers.

[golden]

Part of the issue evaluating defense is that you really need to understand both the defensive schemes and what the individual players are being asked to do in the applicable schemes. You then need to look at each unique game being played and the particular sequences and patterns.

I'm sure the models have some value but there's a reason why teams break down games into very small time periods and try to figure out what is happening in each little segment. It's more telling than aggregating numbers.

Yeah, perfect example is the Raptors selling out as a team to stop Curry at all costs. That makes his primary defender (usually Fred) look like the world's greatest Curry Stopper, and the rest of the team look like chumps who let role players go off.

The inability to capture "intent" is a fundamental flaw of many stats. Same issue with offense. Stan Van Gundy complained about this at a MIT Sloan sports analytics conference. SVG said that stats showed his team being bad in the pick & roll, but SVG said he uses PnR as a starting point to get the team into more complex actions. Whoever is scoring that stat is in no position to say whether that PnR was successful or not, in a vacuum.

[inthepaint]

It's funny because to my eyes, the EPM stats look better. And it's not just Precious. I feel like both Pascal and Khem have both been negatives on defense so far this season. And yet they are graded as positive in the DPM stats.

DFG% isn't everything, but opponents are scoring at 53.7% against him compared to 43% against Precious.
https://www.nba.com/stats/players/de...mID=1610612761

Precious is a terrible offensive player (not including his 3rd quarter yesterday of course).. but I feel like he's one of our best defenders. Will be watching more carefully as the season goes along.

Birch is position-sound and experienced, but he's slow, has a tough time switching onto quick guards, and doesn't have the wingspan/height/bulk to deter rim attacks as your classic rim protector. Precious is also short for a C, but he's very quick, will stay in front of guards with no problem, and can leave the paint to close out on wing shooters well (which Nurse's system calls for)

Birch is better offensively than Precious, but the bar has been low there. Birch has no range, but he's a good pnr target and passer, and can clean up those bunnies around the rim. Precious offence now is mainly fighting for put-backs on the glass and DHO's on the perimeter. Neither is a particularly stellar rebounder. If Precious develop that 3-ball though, that's huge.

[big boi]

Birch is better offensively than Precious, but the bar has been low there. Birch has no range, but he's a good pnr target and passer, and can clean up those bunnies around the rim. Precious offence now is mainly fighting for put-backs on the glass and DHO's on the perimeter. Neither is a particularly stellar rebounder. If Precious develop that 3-ball though, that's huge.

Precious is a good rebounder. If he can do other things well enough to stay on the floor, he'll get a lot of rebounds. As far as I can tell he's averaging over 8 RPG in 24.5 MPG, that's good. And the eye test shows me he has skill as a rebounder.

[DanH]

Precious is a good rebounder. If he can do other things well enough to stay on the floor, he'll get a lot of rebounds. As far as I can tell he's averaging over 8 RPG in 24.5 MPG, that's good. And the eye test shows me he has skill as a rebounder.

Precious has put up dominant defensive rebounding rate the likes of which we haven't seen (anything close to) since JV was here. Nearly 30%! His offensive rebounding is... OK. Under 9%, which is fine, though to describe it as a strength, if you compare again to JV he floated in the 11-13% range most years - but he again is elite. 9% is OK.

Birch has actually been amazing on the offensive glass, grabbing 14% of all misses. Defensively, far less so (13%), actually grabbing less of the other team's misses than our own! Usually players grab about twice as many boards on the defensive glass versus offensive, and that's for bigs where they tend to get disproportionate offensive rebounds (league wide it is more like 3 or 4 defensive boards for every offensive board). Weird rebounding year for him.

Always worth checking on-court team rebounding for players too, as sometimes guys do their job and box out and the team succeeds on the glass in spite of their lack of individual rebounding.

With Birch: 38% OREB%, 71% DREB%
With Precious: 31% OREB%, 75% DREB%

For reference, league wide those team numbers would rank 1st (38% OREB%, by a mile) vs 4th (31% OREB%); and 27th (71% DREB%) vs 4th (75% DREB%). The net differences (offensive boards grabbed versus given up) are very impressive for both players.

[Kagemusha]

Standard RR refrain after most losses: if only we had a REAL center.... like Steven Adams. lol. Did he even play last night?

Steven Adams has been crappy in Memphis so let's not use him as a measuring stick.
We still need a good center.
Let's not brag about one game.
Memphis has been beating good teams and we'll probably not win 2 games out of 3 against them.

[slaw]

Yeah, perfect example is the Raptors selling out as a team to stop Curry at all costs. That makes his primary defender (usually Fred) look like the world's greatest Curry Stopper, and the rest of the team look like chumps who let role players go off.

The inability to capture "intent" is a fundamental flaw of many stats. Same issue with offense. Stan Van Gundy complained about this at a MIT Sloan sports analytics conference. SVG said that stats showed his team being bad in the pick & roll, but SVG said he uses PnR as a starting point to get the team into more complex actions. Whoever is scoring that stat is in no position to say whether that PnR was successful or not, in a vacuum.

One of the more interesting conversation I've had was a couple years ago with a key guy in a NHL analytics dept and how they review games in 6 second increments and give the coaches feedback every intermission on line matchups based on that review. And he was saying how they are focused on individual events cause what the coaches really want to know is who they should put on the ice, when and against whom. He described it as super-advanced scouting as opposed to statistics.

The models can be useful for player comparisons and some of them are more predictive than others but I don't trust them as much as I do baseball stats, for example.

[Primer]

Birch is position-sound and experienced, but he's slow, has a tough time switching onto quick guards, and doesn't have the wingspan/height/bulk to deter rim attacks as your classic rim protector. Precious is also short for a C, but he's very quick, will stay in front of guards with no problem, and can leave the paint to close out on wing shooters well (which Nurse's system calls for)

Birch is better offensively than Precious, but the bar has been low there. Birch has no range, but he's a good pnr target and passer, and can clean up those bunnies around the rim. Precious offence now is mainly fighting for put-backs on the glass and DHO's on the perimeter. Neither is a particularly stellar rebounder. If Precious develop that 3-ball though, that's huge.

Birch is and always has been an elite offensive rebounder.