Olynyk = DPOY?

[clover]

Some of the joke responses here make it seem as if Olynyk is some huge slouch defensively, which i don't believe to be the case at all. He just needs to do much better at contesting without fouling.

He's been doing that terrifically all year. Last night is the first game he's got into significant foul trouble at all.

And KO was among the best in the league in defensive +/- for the second half of last season as well. His opponents have had horrible shooting percentages and he's all over the place providing help defense and zipping between the perimeter and the basket as well.

He's completely underrated here.

[fairweatherfan]

Some of the joke responses here make it seem as if Olynyk is some huge slouch defensively, which i don't believe to be the case at all. He just needs to do much better at contesting without fouling.

He's been doing that terrifically all year. Last night is the first game he's got into significant foul trouble at all.

And KO was among the best in the league in defensive +/- for the second half of last season as well. His opponents have had horrible shooting percentages and he's all over the place providing help defense and zipping between the perimeter and the basket as well.

He's completely underrated here.

Absolutely - stepping out of the weeds of how the stats are calculated, they're just supporting what a careful watch of Olynyk's recent play can tell anyone.  His defense has improved sharply, to the point that he's consistently a plus defender now. 

Best in the league or close to it is of course ridiculous but he isn't hanging around the top of these metrics by accident.  It's also not a coincidence that our DRating with Olynyk off the court is worse than when any other player is off the court.  It just seems like a lot of fans are still hung up on how goofy he looks out there and ignoring what he's actually doing to help the team. 


EDIT: This is a pretty good Reddit thread with some good clips and analysis of Olynyk's recent D:  https://www.reddit.com/r/bostonceltics/comments/3sefy9/how_good_is_kelly_olynyks_defense/

[kozlodoev]

People who think that BR's individual defensive ratings are worth anything clearly have no understanding whatsoever of how they're calculate.

I have a thorough understanding of how they're calculated, and while they're clearly prone to noise, that's not the same as being worthless.  Individual results can and will be fluky, especially after a handful of games, but on aggregate better defenders will find themselves closer to the top of the rankings than worse defenders.  The smaller the sample size and the more dissimilar their team situations are, the more variability there is, but there remains a signal under the noise.  Doesn't have to be all or nothing. 

Most people would consider players like Whiteside, Kawhi, Gobert, Draymond, Duncan, Chandler, LeBron, etc to be among the top individual defenders in the league.  Is your argument that their presence toward the top of DRatings is a total coincidence and completely unrelated to their actual defensive ability?  Because that's what the ratings being worthless would imply.

My argument is that the individual stat correlates so closely with team achievement that it is practically worthless as an indicator of a player's defensive ability.

[The Oracle]

  What Moranis posted from Basketball reference is not true defensive rating.  It is some formula they have come up with to try and evaluate each individuals defense and it is clearly highly dependent on box score events.  Sully's #'s look much better in their formula due to their inclusion of rebounds that really skew his #'s and make it appear as he is much better than he is.  Most of these advanced statistics that can be found on various sites are completely bogus. 

  Defensive rating that can be found on NBA.com is the number of points the team allows per 100 possessions while that player is on the floor.  This is not some concocted formula, it is exactly what it says it is.  For the Celtics Kelly Olynyk has a def. rat. of 82.5, Sully has a def. rat. of 99.3 worst of all regular rotation players on the C's.  So yes K.O. is currently leading the league in def. rat. the team has only allowed 82.5 points per 100 poss. while he is on the floor.

[Celtics18]

Let's not ignore the rate at which Kelly steals the ball.  He's averaging 2.9 steals per 36 minutes, which is good for third in the league among players who have played at least 100 minutes so far (Jae Crowder is first).

On a team that is relying so much on creating turnovers to generate fast break offense, his ability to get his hands on the ball in the lane has been a huge factor for this team.

He's been a legitimate defensive playmaker this year.  It's been good to see. 

[fairweatherfan]

People who think that BR's individual defensive ratings are worth anything clearly have no understanding whatsoever of how they're calculate.

I have a thorough understanding of how they're calculated, and while they're clearly prone to noise, that's not the same as being worthless.  Individual results can and will be fluky, especially after a handful of games, but on aggregate better defenders will find themselves closer to the top of the rankings than worse defenders.  The smaller the sample size and the more dissimilar their team situations are, the more variability there is, but there remains a signal under the noise.  Doesn't have to be all or nothing. 

Most people would consider players like Whiteside, Kawhi, Gobert, Draymond, Duncan, Chandler, LeBron, etc to be among the top individual defenders in the league.  Is your argument that their presence toward the top of DRatings is a total coincidence and completely unrelated to their actual defensive ability?  Because that's what the ratings being worthless would imply.

My argument is that the individual stat correlates so closely with team achievement that it is practically worthless as an indicator of a player's defensive ability.

That logic cuts both ways, of course - good team defenses tend to be made up of good individual defenders.  It's difficult to disentangle individual and team defense from either end, simply because the team defense is just made up of a handful of individuals' efforts.

But again, this "practically worthless" measure correlates fairly well with other individual defensive measures, and with the eye test stuff; the same names tend to keep showing up toward the top.  And of course we can begin to account for teammate effects with on-court/off-court analyses; Olynyk leads the team in those defensive measures as well, the team's DRating is worse with him on the bench than for any other player. 

I guess I have a hard time grasping how one could even believe in the possibility of a non-worthless measure of individual D if the standard is that it can't strongly correlate with team performance.  Isn't the entire reason we emphasize the value of individual defenders because of the assumption they improve the team's overall defense? 

[littleteapot]

I'd say points per possession when you're defending the PNR or when you're defending in the post or when you're the help defender are very valuable.

[fairweatherfan]

I'd say points per possession when you're defending the PNR or when you're defending in the post or when you're the help defender are very valuable.

Sure, those could give you useful info on specific defensive skills, but they're more specialized than overall D, so they're not as informative about overall contributions.

More importantly, though, how would you confirm that these are especially valuable skills, if not by correlating them with the team's overall performance?

[The Oracle]

  With Olynyk on the floor the Celtics allow 82.5 points per 100 poss. which would easily rank them #1 in the league by almost 11 points(Miami currently has the best def. rat. at 93.5).  Without K.O. on the floor the C's allow exactly 100 pts. per 100 poss. which would rank them 14th in the league.

[littleteapot]

Sure, those could give you useful info on specific defensive skills, but they're more specialized than overall D, so they're not as informative about overall contributions.

More importantly, though, how would you confirm that these are especially valuable skills, if not by correlating them with the team's overall performance?

That's true. You have to factor that into the role you'd like for any player, and how the other team could possibly try to expose him. These stats also don't remove who you're playing with from the equation, but they at least show what the other team was attempting to do to score on you and what happened. I just don't see how the team defensive rating when a player is on the floor tells you anything.

By the way, I think the main reason Olynyk looks so good in these on/off stats is that he plays 17 MPG and most teams in the nba just don't have a 4th big. He gets to completely ignore the crappy bigs he's tasked with defending, and if someone does kill him Stevens just benches him and puts one of his other bigs back out there.

[Celtics18]

Based purely on the dreaded "eye test," I contend that Klynyk is not a particularly good one-on-one defender, but he is an exceptionally good help defender.  He reads the pick and roll well and he has quick hands on the interior.

[The Oracle]

Sure, those could give you useful info on specific defensive skills, but they're more specialized than overall D, so they're not as informative about overall contributions.

More importantly, though, how would you confirm that these are especially valuable skills, if not by correlating them with the team's overall performance?

That's true. You have to factor that into the role you'd like for any player, and how the other team could possibly try to expose him. These stats also don't remove who you're playing with from the equation, but they at least show what the other team was attempting to do to score on you and what happened. I just don't see how the team defensive rating when a player is on the floor tells you anything.

Thats the thing though it doesn't matter who K.O. is playing with the team is performing much better defensively with him on the floor no matter who is out there with him.  Never in my wildest dreams would he have ever become a positive impact defensive player but he clearly has.  All defensive metrics point to it and he passes the eye test with flying colors.  His defensive positioning is really good and even guards have extreme difficulty in taking him off the dribble.  His steal rate is extremely high(exceptional for a big man) and he defends a ton of shots spanning the entire floor at a very good percentage.

[littleteapot]

I'll give you the steals. On the other stuff... I don't know I don't think he's consistently contesting shots. It looks to me like the whole defense collapses in anticipation of Kelly trying to cut off a guard and they make most of the plays for him against shots. When someone tries to get into the paint and pass or they hesitate, Olynyk has been a great opportunist. Personally I want a guy in the paint who's scaring the other team into settling for jumpers, not scaring Crowder into a mad dash to the rim to try to make a play.

EDIT: I also don't think he's cutting drives off, pushing people into sidelines or helpers or any of the stuff that made Noah an all star either.

[fairweatherfan]

Sure, those could give you useful info on specific defensive skills, but they're more specialized than overall D, so they're not as informative about overall contributions.

More importantly, though, how would you confirm that these are especially valuable skills, if not by correlating them with the team's overall performance?

That's true. You have to factor that into the role you'd like for any player, and how the other team could possibly try to expose him. These stats also don't remove who you're playing with from the equation, but they at least show what the other team was attempting to do to score on you and what happened. I just don't see how the team defensive rating when a player is on the floor tells you anything.

Well, there are only 4 other players out there with a guy, and they tend to vary a bit, so it can be informative, especially when contrasted with off-court performance.  It's difficult for a truly bad defensive player to consistently put up good DRatings, or vice versa.  It's flawed though, for sure.  But looking at things like opponent FG% and specific situations like you mentioned can help isolate the one player's impact.  It's basically impossible to completely identify one player's impact, though.

There are actually very similar issues for individual offensive performance, but they don't seem to get the same criticisms since we have more counting stats on that end.

By the way, I think the main reason Olynyk looks so good in these on/off stats is that he plays 17 MPG and most teams in the nba just don't have a 4th big. He gets to completely ignore the crappy bigs he's tasked with defending, and if someone does kill him Stevens just benches him and puts one of his other bigs back out there.

This I completely agree with - he's a much better defender than in the past, but the rating #s are being inflated by not matching up as much with starters.  I think he's the only one of our bigs that hasn't started yet.  His numbers would almost certainly drop a bit if he was playing more against first units, but I still think he'd be a plus defender overall, unless a stronger big just kept abusing him in the post.  Though in that case he'd probably get in foul trouble before giving up too many points.

[Celtics4ever]

Still I do not think he wins DPOY, because he is not flashy.   We folks see this stat and think beyond it, I am guessing nope.