Extensive Clutch Stats

[Doctor MJ]

Great thread.

And yeah, I too am thinking a lot more about clutch stats than I used to. The clear trend that came out when we first got these stats was that clutch performers aren't actually very effective (with Kobe being the face of this). It now seems pretty clear that LeBron at the very least shows glaring clutch trends along the lines of what naive fans assume of other stars. (Dirk's the next guy to talk about, and when you factor in playoffs, the luck/skill of the '11 Mavs' clutch game is completely off the charts).

[SideshowBob]

why not use the NetRtg that NBA.com provides (differential between offensive rating and defensive rating) instead of basic +/-?

That's what I used for the team section. For the players, it would be inconsistent to use net +/- per 100 when everything else that is listed is per 48 minutes. Understand that both are the same stat, one is just per 100 possessions, while the other is per 48 minutes.

[colts18]

That's what I used for the team section. For the players, it would be inconsistent to use net +/- per 100 when everything else that is listed is per 48 minutes. Understand that both are the same stat, one is just per 100 possessions, while the other is per 48 minutes.

Both aren't the same stat. They use different defensive ratings which is why you will see some inconsistencies in the stat. NBA.com confirmed it to me.

[SideshowBob]

That's what I used for the team section. For the players, it would be inconsistent to use net +/- per 100 when everything else that is listed is per 48 minutes. Understand that both are the same stat, one is just per 100 possessions, while the other is per 48 minutes.

Both aren't the same stat. They use different defensive ratings which is why you will see some inconsistencies in the stat. NBA.com confirmed it to me.

Yeah, true, I saw that earlier. The point still stands though, it makes sense for me to list +/- per 48 minutes as it would be inconsistent and misleading to use whatever NBA.com is listing on each player's advanced stat page.

[Gideon]

good work

according to +/-, Lebron and Dirk are the greatest clutch performers in the NBA for their career. do I read that right?

LeBron and Dirk also have the two best career ON/OFFs since that stat has been tracked by basketball-reference. I think there's a strong case to be made that they are the two most impactful stars of their generation (with LeBron first, and Dirk second). It's hard to put Dirk about Duncan, and I still have Duncan ranked higher on my list at the moment, but some of these stats that are emerging are at least making me consider that question.

[GreenHat]

Lebron just doesn't have the clutch gene.

Your numbers don't mean anything, the only ones that do when it comes to clutch are 5>1

[SideshowBob]

Updated first post with career numbers for Anthony, Bryant, Durant, James and Rose

[SideshowBob]

Harden added. I'll eventually add playoff numbers as well.

[qm22]

There seems to be a lot of noise in +/- probably mostly due to change in their teams. E..g, for many of these guys you would probably say in their prime years they did not have a drastic variation in their clutch performance, yet statistically the same player can have +/- vary by 20-30 between 2 prime years.

[SideshowBob]

Its raw +/- adjusted to 48 minutes in very small sample sizes. I wouldn't make anything of the +/- numbers other than noting trends.

[aal04]

Clutch to me is last 24 seconds of the clock.

[SideshowBob]

Clutch to me is last 24 seconds of the clock.

No one has to take this as the definition of "clutch," simply take it at face value, which is player/team performance in close late game situations.

[HilltopperJay]

Clutch to me is last 24 seconds of the clock.

Why? Do you feel that games are only decided in the final 24 seconds? And why the arbitrary selection of 24? Why not 28? 33? 42.5?

[TheOUTLAW]

I will never understand why people think that clutch is only during the last few seconds of a game. You can be just as clutch earlier on when your team needs a basket to end a run by the other team.

[ElGee]

good work

according to +/-, Lebron and Dirk are the greatest clutch performers in the NBA for their career. do I read that right?

LeBron and Dirk also have the two best career ON/OFFs since that stat has been tracked by basketball-reference. I think there's a strong case to be made that they are the two most impactful stars of their generation (with LeBron first, and Dirk second). It's hard to put Dirk about Duncan, and I still have Duncan ranked higher on my list at the moment, but some of these stats that are emerging are at least making me consider that question.

They have the two best career on/offs in the regular season. According to BBR, Kevin Garnett is the best in the playoffs for any player career 2001-2012.

[Vinsanity420]

I will never understand why people think that clutch is only during the last few seconds of a game. You can be just as clutch earlier on when your team needs a basket to end a run by the other team.

Generally associated with pressure when the game is winding down. A ton of casual fans only tune in towards the end, too. Everyone's watching the last few plays, etc.

[Gideon]

good work

according to +/-, Lebron and Dirk are the greatest clutch performers in the NBA for their career. do I read that right?

LeBron and Dirk also have the two best career ON/OFFs since that stat has been tracked by basketball-reference. I think there's a strong case to be made that they are the two most impactful stars of their generation (with LeBron first, and Dirk second). It's hard to put Dirk about Duncan, and I still have Duncan ranked higher on my list at the moment, but some of these stats that are emerging are at least making me consider that question.

They have the two best career on/offs in the regular season. According to BBR, Kevin Garnett is the best in the playoffs for any player career 2001-2012.

I focused on regular season on/off, because I think sample size is a huge factor with on/off... for that reason I would put much more stock in regular season on/off (which in the case of KG incorporates almost exactly 10 times the number of games as the playoffs... I assume this is similar for most top players). I've seen bizarre on/offs with even 2 or 3 full seasons of sample size, but once you're looking at multiple seasons up to a career, it starts to seem like a pretty good indicator.

Obviously, KG is a great player (I think he might be 3rd in RS on/off, anyway after Dirk and LeBron), so I'm not saying it's an anomaly he's the leader in on/off for the playoffs... but for a stat that is SO dependent on sample size, I would be much more inclined to focus more on wherever there's a significant sample size (which is clearly RS in this case). Garnett has played 114 playoff games since 2001... that's barely more than a single season, which just is not a reliable sample size for on/off.

*Actually, thinking about this more... the very best solution seems like it would be to just combine numbers from RS and PS and get the largest sample size possible for on/off. That seems like it would be the best indicator. I don't have time to do that for LeBron, Dirk, and KG at this moment, but judging by a quick glance at the numbers, I would expect those to be the top 3 guys using that metric, with LeBron first and Dirk and KG very close for 2nd.

[Chicago76]

Interesting stuff. I haven't reviewed the PBP in any detail, but a few things stick out to me as reasons for the extreme variance among players and for the same player year over year (apart from sample size):

1-Pts per 48 is extremely clock and score sensitive. What I mean by that is that a team up 5 with 1:30 left will end up getting fouled a lot, which means a lot of FTs. If you are a big, mobile target (like a James), you're going to be seeing a lot of the ball off inbounds passes and probably getting a disproportionate share of the FTs. You still have to hit 'em, but going 5-6 in the last couple minutes of a ball game in a FT war will boost the pts/48 quite a bit. On the other end, teams playing from behind will apply a high risk/high reward strategy to make up the differential. They'll take more 3s and shots early in the clock. On average, this reduces their Ortg, but it is the best strategy to attain a W.

2-Individual pts per 48 is also somewhat dependent upon how good your team is. I'm not picking on James here (because I do think he gets a bum wrap in clutch discussions), but let's take his 08-09 cavs. They were a 66 win team with an avg pt differential of +9. If we assume their MOV is normally distributed around that +9, then that means clutch situations (5 pts or less either way within 5 minutes) are going to be skewed toward them having the lead up 4 or 5, which again means more FTs and a nice Ortg on a team level or on an individual level, a nice pts/48.

Probably a cleaner way to look at this is from an expected value/expected outcome approach. Based upon the 5 pt/5 min rule, the following would need to be recorded:
-home team v. away team (2 scenarios)
-time of first possession inside of 5 minutes (could probably be ignored, because it isn't that important if there is 4:50 left or 4:59 left in a close game).
-score at that point in time (11 scenarios of -5 to +5)
-who has the ball (2 scenarios)

If you knew this for every game, you would be able to provide odds for all 44 scenarios across the league. Example, when the home team gets the ball and a 4 pt lead on first possession inside of 5 min remaining, maybe they win 80% of the time. Then compare that to what the Cavs actually did. Apply odds vs. outcome to each scenario and you can measure teams based upon actual wins v. expected wins. This is basically what the advanced stat football crowd does for QBs, but IMO, it is probably more fair to apply this to star players in basketball because the sample size will be greater and they can influence both offense and defense. There is also a lot less variance in team defensive ability in the NBA than the NFL, so you're not stuck hypothesizing about how the offensive strategy is influenced by the ability of the defense to make stops.

[ElGee]

LeBron and Dirk also have the two best career ON/OFFs since that stat has been tracked by basketball-reference. I think there's a strong case to be made that they are the two most impactful stars of their generation (with LeBron first, and Dirk second). It's hard to put Dirk about Duncan, and I still have Duncan ranked higher on my list at the moment, but some of these stats that are emerging are at least making me consider that question.

They have the two best career on/offs in the regular season. According to BBR, Kevin Garnett is the best in the playoffs for any player career 2001-2012.

I focused on regular season on/off, because I think sample size is a huge factor with on/off... for that reason I would put much more stock in regular season on/off (which in the case of KG incorporates almost exactly 10 times the number of games as the playoffs... I assume this is similar for most top players). I've seen bizarre on/offs with even 2 or 3 full seasons of sample size, but once you're looking at multiple seasons up to a career, it starts to seem like a pretty good indicator.

Obviously, KG is a great player (I think he might be 3rd in RS on/off, anyway after Dirk and LeBron), so I'm not saying it's an anomaly he's the leader in on/off for the playoffs... but for a stat that is SO dependent on sample size, I would be much more inclined to focus more on wherever there's a significant sample size (which is clearly RS in this case). Garnett has played 114 playoff games since 2001... that's barely more than a single season, which just is not a reliable sample size for on/off.

*Actually, thinking about this more... the very best solution seems like it would be to just combine numbers from RS and PS and get the largest sample size possible for on/off. That seems like it would be the best indicator. I don't have time to do that for LeBron, Dirk, and KG at this moment, but judging by a quick glance at the numbers, I would expect those to be the top 3 guys using that metric, with LeBron first and Dirk and KG very close for 2nd.

Gideon I don't actually see on/off that way. I don't think you are increasing accuracy by increasing sample at all. I haven't tested variation after 50 games say, versus 80, but I don't see that as the issue. I think the stat just isn't particularly accurate for measuring goodness. It is what it is, which is first and foremost an indicator of what's happening with a guy on the court versus off it. That's interesting. It's valuable. It's also a conditional measure, and for the purpose of measuring player goodness, is subject to some major confounding variables like teammates and opponents (this is why people try and adjust it).

I don't see increasing the sample as changing these issues at all. And similarly, I don't see the samples for multi-year PS, which I can post tomorrow, as being subject to a small sample. For his career, KG has 4800 minutes on and 1200 minutes off -- what sample-size argument can you come up with to object to that?

[SideshowBob]

Note that I only adjusted to 48 minutes to remain consistent with the 82games numbers we had before.

2-Individual pts per 48 is also somewhat dependent upon how good your team is. I'm not picking on James here (because I do think he gets a bum wrap in clutch discussions), but let's take his 08-09 cavs. They were a 66 win team with an avg pt differential of +9. If we assume their MOV is normally distributed around that +9, then that means clutch situations (5 pts or less either way within 5 minutes) are going to be skewed toward them having the lead up 4 or 5, which again means more FTs and a nice Ortg on a team level or on an individual level, a nice pts/48.

True, but we're also seeing the same from him in 08, when Cleveland had an average MOV of -.35, and FWIW, his FTA/48 were higher in both surrounding seasons (08 and 10), both weaker years for Cleveland than 09. Still, the logic holds, that's definitely something more than consideration.

Probably a cleaner way to look at this is from an expected value/expected outcome approach. Based upon the 5 pt/5 min rule, the following would need to be recorded:
-home team v. away team (2 scenarios)
-time of first possession inside of 5 minutes (could probably be ignored, because it isn't that important if there is 4:50 left or 4:59 left in a close game).
-score at that point in time (11 scenarios of -5 to +5)
-who has the ball (2 scenarios)

If you knew this for every game, you would be able to provide odds for all 44 scenarios across the league. Example, when the home team gets the ball and a 4 pt lead on first possession inside of 5 min remaining, maybe they win 80% of the time. Then compare that to what the Cavs actually did. Apply odds vs. outcome to each scenario and you can measure teams based upon actual wins v. expected wins. This is basically what the advanced stat football crowd does for QBs, but IMO, it is probably more fair to apply this to star players in basketball because the sample size will be greater and they can influence both offense and defense. There is also a lot less variance in team defensive ability in the NBA than the NFL, so you're not stuck hypothesizing about how the offensive strategy is influenced by the ability of the defense to make stops.

That'd take quite some time though, even within a single season. And at that point I'd be left wondering whether its even worth it. Even whilst posting all this, one of Elgee's blog posts comes to mind, the essential idea of which was that late game play doesn't have significant impact on title odds anyway. Still, I wouldn't brush the idea off.