 |
APBRmetrics
The statistical revolution will not be televised.
|
| View previous topic :: View next topic |
| Who is the 2010-11 NBA MVP ? |
| Kobe Bryant, LAL |
|
2% |
[ 1 ] |
| Kevin Durant, Okl |
|
2% |
[ 1 ] |
| Pau Gasol, LAL |
|
0% |
[ 0 ] |
| Dwight Howard, Orl |
|
31% |
[ 11 ] |
| LeBron James, Mia |
|
31% |
[ 11 ] |
| Kevin Love, Min |
|
2% |
[ 1 ] |
| Dirk Nowitzki, Dal |
|
8% |
[ 3 ] |
| Chris Paul, NOH |
|
2% |
[ 1 ] |
| Derrick Rose, Chi |
|
17% |
[ 6 ] |
| Dwyane Wade, Mia |
|
0% |
[ 0 ] |
|
| Total Votes : 35 |
|
| Author |
Message |
EvanZ
Joined: 22 Nov 2010
Posts: 310
|
 Posted: Fri Mar 25, 2011 7:07 pm Post subject: |
 |
|
Well, the only way I'd give Rose the award is if value was defined in terms of productivity per $ of salary (he makes $5.5M this season). And if that were the criterion, then Westbrook would be in the running, too ($4M). And that's probably a very important part of it. I don't think Chicago could afford Boozer without Rose being so cheap.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir |
|
| Back to top |
|
 |
greyberger
Joined: 27 Sep 2010
Posts: 54
|
| Posted: Sat Mar 26, 2011 11:20 am Post subject: |
|
|
| Ric Bucher held anothed chat and gets a little worked up about the role of stats in public evaluation of players. |
|
| Back to top |
|
|
EvanZ
Joined: 22 Nov 2010
Posts: 310
|
|
| Back to top |
|
|
erivera7
Joined: 19 Jan 2009
Posts: 185
Location: Chicago, IL
|
| Posted: Sat Mar 26, 2011 1:58 pm Post subject: |
|
|
N/A
_________________
@erivera7
I cover the Orlando Magic - Magic Basketball
Last edited by erivera7 on Tue Mar 29, 2011 2:59 pm; edited 4 times in total |
|
| Back to top |
|
|
YaoPau
Joined: 28 Jan 2009
Posts: 35
|
| Posted: Sat Mar 26, 2011 2:39 pm Post subject: |
|
|
OKC net rating with Westbrook on the floor: +1.3
OKC net rating with Westbrook off the floor: +10.9
Seems like that could be fluky, as he's had decent plus minus numbers in past years. But I think you have to at least consider why that's happening, as elite players usually don't have those splits.
Most of us assume Westbrook is a very good defender, but his lineups have been terrible defensively this year. Offensively he's not a particularly efficient scorer, he doesn't spread the floor with a 3pt shot (which can really limit that offense with him and Thabo in the backcourt), and while he can run an offense, I've never thought of him as an elite passer.
I think Bucher handled it fine. He said stats are valuable, but you have to look at how a player is being used on his team, and how his style fits. |
|
| Back to top |
|
|
BobboFitos
Joined: 21 Feb 2009
Posts: 201
Location: Cambridge, MA
|
| Posted: Sun Mar 27, 2011 3:14 am Post subject: |
|
|
| erivera7 wrote: | Bucher's response to the question about Westbrook's numbers compared to Rose is flat-out arrogant: "Because statistics don't determine who the better player is."
So then what determines it? Fulfilling a narrative? I didn't know we're in the business of writing compelling children stories. |
The bigger issue that you/me/we don't seem to really dive on is that it's unbelievably presumptuous that the individual voter (we'll focus on Bucher, since he was the one who said that above) thinks they have that innate ability - that apparently 99.9% of other basketball viewers do not possess - to gauge true value. As in, they are allowed to completely disregard stats, because the stats paint a picture oftentimes completely different than the narrative they had planned/wanted to write about. So, simply create a new metric - we'll call it "what I think" - and if the player in question leads in that metric, well then, he deserves the award!
So for Bucher, well, Rose has a 107.4 WIT while Dwight is only 94.6 WIT, so hey, maybe if Dwight saved some orphans from a burning building he'd accumulate 15 WIT points to become the true MVP of this season.
--
The other issue as well is that the phrase above - "statistics don't determine the better player" - is actually not completely wrong. Well, on the surface it's ludicrous - but I'd counter by saying that the statistics that don't determine the better player are probably poor statistics! Or, at least, are useful in a different/appropriate context. For example, points per game is a statistic, and Monta Ellis averages 24.1 points. Rajon Rondo averages 10.2. Well then, points per game says Monta is "better" than Rondo! Well, not so fast. PPG is a statistic that has some value, but it's not a baseline for player A vs player B. etc etc.
_________________
http://pointsperpossession.com/
@PPPBasketball |
|
| Back to top |
|
|
BobboFitos
Joined: 21 Feb 2009
Posts: 201
Location: Cambridge, MA
|
| Posted: Sun Mar 27, 2011 3:31 am Post subject: |
|
|
| YaoPau wrote: |
I think Bucher handled it fine. He said stats are valuable, but you have to look at how a player is being used on his team, and how his style fits. |
Because the chat was linked I decided to read it, and he did not say this. Prefacing his comment ("statistics don't matter") by saying fit, style, context etc. are important when looking at said statistics, well, it would have gone a long way. Still doesn't make him right, but at least it salvages him somewhat and it's no longer completely ignorant. Instead he said:
| Quote: | Ric Bucher
(1:43 PM)
No, thank you. To be clear, stats can be very valuable - in gauging a player's improvement, efficiency, etc. But using them to compare players on different teams in different systems with different roles...well, a lot of people (who don't work in or around the NBA) do it. But it doesn't make it right. |
He is saying the statistics accrued on team A do not translate to team B. Last time I checked, teams are fairly similar, albeit with different levels of success. You need some people who can score. Need some people who can rebound. Need some people who can defend. Maybe you get some people who can do all 3. Maybe throw in some passing. Last time I checked, as well, roles translated fairly similarly from team A to team B - Dwight Howard is going to still be the best defender in the league if he's on the Minnesota Timberwolves. If he's on the Wizards, Jazz, or Nuggets, well, seems to me he'll still do the same things he does with the Magic.
I'm of the opinion the onus is on others who claim the opposite. What player has been wildly successful on one team, and then completely awful on another, in the same year? Sure, at the margins (again, due to fit, system, role, etc.) actual numbers will adjust slightly, but people, it's not drastic. It's the same game with the same rules whether you're in New York or California.
| Quote: | Ric Bucher
(1:50 PM)
This is hard to say without coming off as arrogant, but I'm going to try: the access I have, and have had for nearly 20 years, informs my opinion. You don't have that access. I understand that. Some who do, don't know what to do with it or don't utilize it, for whatever reason. More than anything, I remain teachable. When I write or say something, it's almost never without having checked it out with people in the league whose opinions I trust, and who will tell me when I'm off. I don't cite those people because they're usually multiples, but rest assured my understanding of who is good and who is not, who is doing what and who is not, is not based on my thoughts alone. |
I'll continue to stake my beliefs in numbers because they remain objective. I suppose Bucher clings to those beliefs - that his inside access has allowed him to have more correct and credible understanding of who's doing what in each game and for each team - pretty strongly, because there was never an alternative.
Everything can be measured. The holes in basketball analysis stem from those measurements not being accurate (or, more on point, not FULLY accurate) - not that they are immeasurable.
EDIT: Actually, there are certain things that can't be measured, I suppose. But those intangibles have no value when analyzing contribution towards wins and losses. If someone argued Derrick Rose has more "heart" than Dwight Howard, fine, whatever. Maybe he does... But does this matter? Seems to me it does for a nice narrative, but shouldn't when deciding value to a team.
_________________
http://pointsperpossession.com/
@PPPBasketball
Last edited by BobboFitos on Sun Mar 27, 2011 5:34 am; edited 1 time in total |
|
| Back to top |
|
|
YaoPau
Joined: 28 Jan 2009
Posts: 35
|
| Posted: Sun Mar 27, 2011 4:49 am Post subject: |
|
|
Huh? Bucher didn't mention how statistics translate, he just stated context is needed when working with stats.
Of course a player's value changes depending on his teammates and his team's system. Look at how LeBron's and Wade's clashing styles have affected their play. Look at how Carmelo's and Amare's bad defense have lead to Knicks losses. With better defensive teammates, both guys can have all-star value. Together their lineup's upside is limited. I mentioned Russell Westbrook's lack of a 3pt shot is a problem next to Thabo. I'd bet he'd be more valuable replacing Rondo/Jameer on the Celtics or Magic.
As for the days of the insider being numbered, no chance, although insiders will have to understand advanced stats to a degree. No matter how complicated your regression is in R, it won't match the complex regressions the brain can handle. I haven't seen an objective stat yet that has passed the eye test using 1 year of data to value players. |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3630
Location: Hendersonville, NC
|
| Posted: Sun Mar 27, 2011 7:56 am Post subject: |
|
|
Current chances of teams winning a title, according to b-r.com
| Code: | title% tm title% tm
.227 Chi .022 Okl
.185 Mia .011 NO
.157 SA .004 Por
.141 LA .003 NY
.084 Bos .002 Mem
.074 Orl .001 Hou
.046 Dal .001 Ind
.042 Den
|
A .001 probability means that a team won one 'title' in 1000 simulations, this morning.
http://www.basketball-reference.com/friv/playoff_prob.cgi
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
schtevie
Joined: 18 Apr 2005
Posts: 415
|
| Posted: Sun Mar 27, 2011 10:12 am Post subject: |
|
|
Wow, I'll take the Celtics to win it all, if someone gives me those odds!
With all due respect to the previous predictive power of the simulation, I don't think that these numbers well represent the current chances of teams winning a title. The current chances of teams winning a title are determined by the expected efficiency (and health) of players/lineups that are actually expected to suit up when the playoffs begin. (Does Justin's simulation include anticipated bench shortening?)
I am quite certain that Shaquille O'Neal's expected (?) return would substantively change these numbers. Duncan's ankle sprain, yes? What other teams are expected to see substantive lineup changes for the playoffs? Do Denver's listed odds include the apparently beneficial purging of Anthony? |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3630
Location: Hendersonville, NC
|
| Posted: Sun Mar 27, 2011 10:52 am Post subject: |
|
|
Heck, let's just multiply these team title probabilities by their stat amalgams on page 2 of this thread.
And take the square root of that product.
Have we now got a measure of 'player most likely to lead a team to a title'? | Code: | STM stat/team mvp tm smvp2 title%
8.90 James,Lebron Mia 428 .185
6.94 Rose,Derrick Chi 212 .227
6.79 Wade,Dwyane Mia 249 .185
5.77 Gasol,Pau LAL 236 .141
5.24 Howard,Dwight Orl 371 .074
5.11 Ginobili,Manu SAS 167 .157
5.05 Bryant,Kobe LAL 181 .141
3.56 Pierce,Paul Bos 150 .084
3.12 Garnett,Kevin Bos 116 .084
2.91 Nowitzki,Dirk Dal 185 .046
1.93 Durant,Kevin Okl 169 .022
1.87 Paul,Chris NOH 317 .011
1.67 Westbrook,Russel Okl 127 .022
0.75 Aldridge,Lamarcu Por 142 .004
|
Players for teams having zero chance to win it all, of course, are rated zero. Regardless of how good they are.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
huevonkiller
Joined: 25 May 2010
Posts: 15
Location: Miami, Fl
|
| Posted: Mon Mar 28, 2011 1:49 pm Post subject: |
|
|
| YaoPau wrote: | What are your definitions of value?
For me, the MVP should go to the player whose regular season performance helped his team's title chances the most. I think winning a title is the point of playing, so if you look back and say "that guy's performance in the regular season helped set his team up for a title run more than any other player", he should get MVP.
So stuff like luck, leadership, clutchness, and winning the important games against contending rivals makes a difference, and I think these analyses of overall efficiency are trying to fit an estimate to something we can break down further.
LeBron has cost the Heat several games by failing in the clutch, as they're 2-8 in close games. The Heat are 0-6 against the Celtics and the Bulls, and had they won those games they'd be the #1 seed easily. But instead they're on pace to have to win three road series to win the title. Unless you think his efficiency is so extraordinarily high that that stuff doesn't make a difference, LeBron isn't the MVP. And when I look at his 1-year APM and see it's just a hair above Rose's and below Dirk/Howard/Aldridge/Garnett, and his box score numbers are down from past years, I think he's in the discussion but likely not the best choice.
Howard/Aldridge/Nash/Paul have all been great this year, but have they increased their team's title chances dramatically? The way the playoffs usually play out, an elite player on a bad team is basically the same as an elite player on a #4 seeded team imo.
I think Dirk, Garnett, Ginobili are the three guys with elite efficiency and plusminus numbers who are on elite teams. I think Rose has to be in the discussion if the Bulls get the #1 seed, as he's been efficient enough, and he's been their undoubted leader all season, and beaten every top team. Kobe, Gasol deserve mention too, I just wonder if this is the year Gasol cancels Kobe out. It's between Dirk and Rose for me right now, with the other guys in the hunt if their teams move up in the standings. |
I'm sorry but this is one of the more immature responses I've seen here. I'm surprised considering you're an "apbr" guy or whatever, but this is full of nonsense. Support your opinion without relying on just team metrics.
The Heat have had some chemistry problems, by no means did they lose because of James. This might be one of his best seasons against elite defenses, Wade and Bosh simply are learning how to play in different ways. LeBron has to play differently too, but he's adjusted better than them.
| schtevie wrote: | Wow, I'll take the Celtics to win it all, if someone gives me those odds!
With all due respect to the previous predictive power of the simulation, I don't think that these numbers well represent the current chances of teams winning a title. The current chances of teams winning a title are determined by the expected efficiency (and health) of players/lineups that are actually expected to suit up when the playoffs begin. (Does Justin's simulation include anticipated bench shortening?)
I am quite certain that Shaquille O'Neal's expected (?) return would substantively change these numbers. Duncan's ankle sprain, yes? What other teams are expected to see substantive lineup changes for the playoffs? Do Denver's listed odds include the apparently beneficial purging of Anthony? |
Really the Celtics should be the favorites? I don't think so man, no metric has them at that level. Rondo is not doing well. |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3630
Location: Hendersonville, NC
|
| Posted: Mon Mar 28, 2011 2:15 pm Post subject: |
|
|
Getting Hollinger's PER and EWA from ESPN site, WS and WS/48 from b-r.com.
Then projecting EWA and WS to the full 82 games.
SMVP today = the cube root of these 3x2 win-estimating metrics and rates.
| Code: | smvp stat mvp tm e484 e82 PER EWA82 ws48 WS82
43.6 James,Lebron Mia 2.29 14.5 27.0 25.3 .238 15.3
39.0 Howard,Dwight Orl 2.15 13.2 26.3 23.1 .236 14.6
33.7 Wade,Dwyane Mia 2.08 12.4 25.5 21.6 .211 12.7
32.4 Paul,Chris NOH 1.79 10.7 24.5 19.4 .249 15.0
31.8 Rose,Derrick Chi 2.10 13.2 23.5 18.8 .204 12.9
31.1 Gasol,Pau LAL 1.76 11.2 23.6 18.4 .234 14.9
29.7 Love,Kevin Min 1.89 11.0 24.6 18.3 .218 12.9
28.2 Durant,Kevin Okl 1.80 11.3 23.8 20.0 .192 12.1
28.2 Bryant,Kobe LAL 2.09 12.0 24.1 18.9 .184 10.7
25.7 Nowitzki,Dirk Dal 1.89 9.6 24.1 15.4 .222 11.5
24.8 Westbrook,Russel Okl 1.95 11.6 23.9 18.5 .161 9.6
22.6 Aldridge,Lamarcu Por 1.60 10.8 21.6 16.4 .167 11.3
22.3 Randolph,Zach Mem 1.79 10.3 22.4 15.1 .176 10.2
21.8 Griffin,Blake LAC 1.83 11.8 21.7 15.7 .147 9.5
20.1 Stoudemire,Amare NYK 1.73 10.7 23.0 17.1 .134 8.3
|
LeBron is solidly 1st in everything but WS/48. With NO fading without West, that could change.
Voting at this moment: LeBron 8, Dwight 6, Rose 4, Dirk 3
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
mtamada
Joined: 28 Jan 2005
Posts: 377
|
| Posted: Mon Mar 28, 2011 2:27 pm Post subject: |
|
|
| huevonkiller wrote: |
| schtevie wrote: | Wow, I'll take the Celtics to win it all, if someone gives me those odds!
|
Really the Celtics should be the favorites? I don't think so man, no metric has them at that level. Rondo is not doing well. |
I don't think that's what schtevie is saying. There are two parts to his quote, and you're focusing on just the first part:
| Quote: | | Wow, I'll take the Celtics to win it all |
But the second part of his quote
| Quote: | | if someone gives me those odds! |
makes it clear that he's saying that the Celtics' probability of winning the title is the most underestimated amongst those figures. I.e. they're listed as having an 8.4% chance. If their true probability is say 16.8%, and if a bookie gives you those odds (and for simplicity let's assume he charges you nothing, i.e. there's no "vig"), then you can plop $1,000 on the Celtics and have an expected value of $2,000: about 5 times out of 6 you lose your bet, but 1 time out of 6 you collect about $12K. Net expected value of $2K.
It's how all financial moguls make their millions: invest in undervalued assets. IOW buy low and sell high. You bet on the Celtics not because they have the highest probability of winning the championship, but because their probability is the most grossly underestimated (if Schtevie is correct).
If the Lakers have a 50% true probability of winning the championship, but the oddsmakers list them at 90%, and you thus win only $1.1K when they win, and 0 when they lose, they are a very bad bet even though they're the favorites. |
|
| Back to top |
|
|
YaoPau
Joined: 28 Jan 2009
Posts: 35
|
| Posted: Mon Mar 28, 2011 3:02 pm Post subject: |
|
|
Huevon, I have a different perspective, that elite value comes from taking your team to a top playoff seed, since that's the top objective for teams in the regular season.
That's fine that you call that support immature, but I'm interested to hear why you think that way.
Last edited by YaoPau on Mon Mar 28, 2011 3:03 pm; edited 1 time in total |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
