Is Westbrook Exposing BPM ?

[huevonkiller]

Westbrook does have a TS% of 0.546, which is about average. BPM only looks at TS% in terms of efficiency. So it sees ridiculous volume at league-average efficiency.

Given his level of usage, the turnovers are reasonable. The team averages fewer turnovers with him on the floor than off. ( http://nbawowy.com/#/l9ioib5uhlp )

And you conveniently forgot his monster offensive rebounding, and assist rate that is higher than any of Chris Paul's seasons. Or looking at it another way, Hollinger turnover rate says he's even better than LeBron (sorry Mini-Shaq).He's what Kobe Bryant should have been, or what Kobe's fans thought he was.

If Westbrook had those low-IQ ball coaches from Jordan's era not spacing the floor correctly (see the previous decade), his TS% would be above average. So it's because the game is so much smarter now that we're even having this discussion. Even so his season this year might still be the best ever because of his dominance offensively.

http://www.basketball-reference.com/lea ... _1991.html

We are told than BPM's use of offensive stats for defensive ratings helps the model or overall model but Westbrook is a case where this choice probably hurts the credibility of his DRPM. I don't know how rare it is that it matters much but from the outside I worry that it is a generalized problem.

Point guard is the most integral position offensively, so it is zero surprise that someone with Westbrook's style dominates OBPM. His OBPM is the highest in the history of basketball for someone not named Steph Curry (and Steph Curry doesn't seem capable of playing like that in the Finals, nothing personal). See my previous post for why this is the case. You clearly didn't take the time to look at the quality of his season.

Westbrook's OBPM is higher than James Harden total BPM, so he's not being sustained by that DBPM. West's OBPM might finish higher than any of Jordan's seasons. Defense is not really the reason Westbrook is in the discussion for MVP, but it doesn't hurt that the Thunder defend much better with him on the court.

http://www.sports-reference.com/blog/20 ... -mvp-race/

In fact, I'd trust Westbrook in his prime in the Finals instead of LeBron, who seems to have psychological/playing-style issues (2007, 2011), stamina issues (2014, 2015), injury issues (2010), getting bailed out by Irving issues (game 7 2016), and playing in the Semi-Pro league issues (obviously) . I don't think Jordan's era was as good as this one, and Westbrook's offensive stats are comparable to the very best ever, at worst. It is pretty easy to see why someone would think West is this season's MVP, since he's only up against a barely-dunking James Harden, and an older, Charles-Barkley-hurt-my-feels LeBron James which is surrounded by elite shooters like Harden. James Harden doesn't have the two-way potential West has, or BPM in the playoffs, either. I'd bet money on James Harden winning the MVP though.

[Mike G]

.. I'd trust Westbrook in his prime in the Finals instead of LeBron, who seems to have psychological/playing-style issues (2007, 2011), stamina issues (2014, 2015), injury issues (2010), getting bailed out by Irving issues (game 7 2016), and playing in the Semi-Pro league issues (obviously) .

Didn't LeBron dominate the last 2 NBA Finals? Hasn't he become more consistently dominant in recent years?
He's the all-time leader in playoff Win Shares and VORP; but it's all about past shortcomings?
Obviously what?

[DSMok1]

I'm an Oklahoma City Thunder fan, and I developed Box Plus/Minus.

I believe Westbrook is breaking my regression, not that he's better than LeBron ever was.

[Mike G]

I think it may be Westbrook's Ast% that is breaking stuff. In one game, it was 100% --
http://www.basketball-reference.com/box ... 70OKC.html

In ten other games, it's been over 70%. He's not assisting on anywhere near 70% of the FG, since he's doing so much of the scoring.
Maybe a term like (1-Usg%)*Ast% would be useful, even if it does start to look like just Ast/x poss.

It doesn't seem right for a hypothetical ball hog who scores every FG but two -- during which he got one assist -- to have the same Ast% as someone who assisted 15 of his teammates' 30 FG.

[Crow]

hk, my comment that you highlighted is entirely about DRPM and your comment and comment on my comment is not specifically about DRPM. I have never shown any interest in the MVP race here or anywhere and don't plan on starting.

[Statman]

I think it may be Westbrook's Ast% that is breaking stuff. In one game, it was 100% --
http://www.basketball-reference.com/box ... 70OKC.html

In ten other games, it's been over 70%. He's not assisting on anywhere near 70% of the FG, since he's doing so much of the scoring.
Maybe a term like (1-Usg%)*Ast% would be useful, even if it does start to look like just Ast/x poss.

It doesn't seem right for a hypothetical ball hog who scores every FG but two -- during which he got one assist -- to have the same Ast% as someone who assisted 15 of his teammates' 30 FG.

I think the same issue would "break" bmp at the other, low end - if a guy was 40 pts & 20 reb per 36, with a 75% TS%, low turnover rate, yet had 1 assist on the season - I believe he'd have a crazy low (ie very negative) bmp, & be considered one of the very worst in the league by the metric. We just don't have a guy at that end, but we do have RWB on the positive BMP end. Almost ALL of the "problems" I tend to have with bpm is tied to that Ast%*Dreb% part or whatever the formula is.

[bchaikin]

I think the same issue would "break" bmp at the other, low end - if a guy was 40 pts & 20 reb per 36, with a 75% TS%, low turnover rate, yet had 1 assist on the season - I believe he'd have a crazy low (ie very negative) bmp, & be considered one of the very worst in the league by the metric. We just don't have a guy at that end,

last 2 years (1415-1516) hassan whiteside - 19.6 pts/40min, 61.2% ScFG%, 16.5 reb/40min, 4.8 bs/40min, just 0.4 ast/40min...

[Crow]

PER, WS/48, RPM and BPM all have Whiteside down this season compared to last but BPM has the biggest decline and has him below league average (and negative for his whole career except last season).
The outlier metric deserves consideration but it may have more to prove.

These and other metrics are lenses (of different qualities and with different quirks), not necessarily THE answer. Using the ensemble for thinking and / or an explicit blend of them makes sense to me.

[Mike G]

Last year, Whiteside's BPM was greater than his RPM, by 2.5 to 1.4. His Usg/Ast% ratio was (207/24) 8.6
This year, his Usg/Ast is down to (223/34) 6.6; and his BPM trails his RPM by -1.0 to 0.9

I ran fauxRPM after 62 games for Miami, and their top 10 looked like this:

Code: Select all

Min     Heats          fRPM   BPM   PER   WS/48
2035  Hassan Whiteside 2.6   -1.3   21.8   .168
1941   Goran Dragic    1.7    3.0   20.1   .154
1743   Tyler Johnson    .8    1.9   16.0   .133
1623   James Johnson    .4    2.6   16.6   .113
1599  Rodney McGruder -2.0     .0    9.3   .080
1343   Dion Waiters   -1.0    -.6   14.6   .058
1201  Wayne Ellington -1.1     .1   12.9   .107
1100  Josh Richardson -2.1   -1.2    9.3   .042
906    Luke Babbitt   -2.3   -1.5    8.7   .086
809    Willie Reed     -.2    -.6   17.2   .164

   correl. With min.   .82    .52    .73    .69

fRPM formula has 2 terms involving Ast%. After converting Whiteside's 3.4 Ast% to .034, Usg to .207, etc, the terms are:
Ast*TRb*92.5 . . and . . (1-Usg)*Ast*(-8.06)

If we change his Ast% from 3.4 to zero, his fRPM drops from 2.6 to 2.1
If we triple Ast to 10.0, fRPM rises to 3.7

[Crow]

With any discrete stat we can see what a player does (on average or in lineups) but if one had the time and rationale for using the time, one could look at average total demonstrated capacity of any lineup (or all of them) for that stat, actual total production and whose capacity got utilized. If a player does not do something but the lineup is ok or better on that or in spite of a deficiency there then the low individual production doesn't bother me as much. If utilization of individual capacity is high but lineup production isn't as good, it impresses less. RPM at least tries to get at net team impact. Individual box score metrics are blind or agnostic about many team stats while on court or indirect impacts on team stats. One may think of individual box score stats and metric as blind to team or random noise but I don't think they really are. Almost everything is affected by lineup context, indirect impacts and noise.

[bchaikin]

PER, WS/48, RPM and BPM all have Whiteside down this season compared to last but BPM has the biggest decline and has him below league average (and negative for his whole career except last season).

whiteside this season is currently averaging 16.7 pts/g, 14.2 reb/g, and 2.1 bs/g. over the last 4 decades only 3 players have averaged 16+ pts/g, 14+ reb/g, and 2+ bs/g in a single season - dwight howard (3 times), moses malone (1), and hakeem olajuwon (1)...

that's 6 times a player had done this in 4 decades, and while whiteside did not score as well as these others did, he was more efficient on offense (pts/0ptposs), committing just 2.1 to/g (each of the others was at 3+ to/g)...

[Statman]

I think the same issue would "break" bmp at the other, low end - if a guy was 40 pts & 20 reb per 36, with a 75% TS%, low turnover rate, yet had 1 assist on the season - I believe he'd have a crazy low (ie very negative) bmp, & be considered one of the very worst in the league by the metric. We just don't have a guy at that end,

last 2 years (1415-1516) hassan whiteside - 19.6 pts/40min, 61.2% ScFG%, 16.5 reb/40min, 4.8 bs/40min, just 0.4 ast/40min...

He is the most obvious current example, it just isn't the magnitude negatively that Westbrook is on the positive end.

Hassan was below average bpm in 2015. In 2016 he had a big uptick in BPM to merely a meh 2.5 (from -0.9). This is with almost all his rates remaining the same or getting worse slightly - except he averaged about a half an assist per 48 more in 2016 (with about a .7 increase in TO per 48 mind you). He went from well below average to a decent player pretty much because of 1/2 an assist per 48 - as far as I can tell.

[bchaikin]

He went from well below average to a decent player pretty much because of 1/2 an assist per 48 - as far as I can tell.

whiteside was the ages of 25-27 this and the past 2 seasons (1415-1617). during that time (5462 minutes) he averaged 20.0 pts/40min, 16.8 reb/40 min, and 5.6% BS...

you know how many players in that age range over the past 4 decades (1000+ minutes) averaged 20+ pts/40min, even 13+ reb/40min, and 5.0%+ BS?...

just 2 other players - david robinson and hakeem olajuwon...

for comparison dikembe mutombo in that age range was at just 15.3 pts/40min, 13.5 reb/40min, and 5.5% BS...

[Statman]

He went from well below average to a decent player pretty much because of 1/2 an assist per 48 - as far as I can tell.

whiteside was the ages of 25-27 this and the past 2 seasons (1415-1617). during that time (5462 minutes) he averaged 20.0 pts/40min, 16.8 reb/40 min, and 5.6% BS...

you know how many players in that age range over the past 4 decades (1000+ minutes) averaged 20+ pts/40min, even 13+ reb/40min, and 5.0%+ BS?...

just 2 other players - david robinson and hakeem olajuwon...

for comparison dikembe mutombo in that age range was at just 15.3 pts/40min, 13.5 reb/40min, and 5.5% BS...

You do realize we appear to be on the same page here. Preachin' to the choir.

[Mike G]

Hassan was below average bpm in 2015. In 2016 he had a big uptick in BPM to merely a meh 2.5 (from -0.9). .. He went from well below average to a decent player pretty much because of 1/2 an assist per 48 - as far as I can tell.

Another possibility is that the BPM team adjustment changed significantly from one year to the next.

Without a team adjustment, BPM would have to try harder to make the stat rates work.