ESPN wisdom of the crowd average win projections

Home for all your discussion of basketball statistical analysis.
Bobbofitos
Posts: 306
Joined: Sat Apr 16, 2011 7:40 am
Location: Cambridge, MA
Contact:

Re: ESPN wisdom of the crowd average win projections

Post by Bobbofitos »

J.E. wrote:
Bobbofitos wrote:Bawsh.
who?
Chris Bosh
I also don't believe in Avery Johnson
I don't think anyone could have done much with last year's roster. Do you not like him from his Dallas days? Postseason success was his problem there, not regular season success.
I'm not even sure the directive was to "win" last year, so I'm not holding him accountable for last year. I understand his undoing (at the end) was lack of success as far as winning a 'ship, but he misplays a lot of things around the margin. Sometimes that costs wins, other times you get lucky.
The Sixers, despite adding ONE player who is arguably better than anyone on their roster from last year, lost their THREE most important players, and are expecting to fill that with downright terrible players (hello Nick Young!).
Since when is Andrew Bynum downright terrible? They still have Doug Collins, who I think plays a big role in their success
Andrew Bynum is anything but downright terrible - he's very good. I said as much. Their 3 most important players (Brand, Iggy, LouWill) probably accounted for ~75% or so of their results, with a bit of Jrue and Thad Young sprinkled in. I also am fully on board the Doug Collins bandwagon, as he may even get guys (like Nick Young) playing a shred of defense. But they also lost Jodie Meeks (fine roleplayer/3pt specialist who filled a very real need for them), Nik Vuk (promising young bigman).

These minutes will get filled by "someone", and those someones appear to be a mixture of Evan Turner, (which I think is a mistake, since he's not good at basketball) Nick Young (duplicate in every way to Lou Will, just worse by a fair amount), Jason Richardson (mediocre player, depends how they shift minutes), and Kwame Brown. (He arguably wont get on the court, depends how much they plan on playing Hawes at the 4, which I also don't buy as workable)


The Pistons, Magic, and Bobcats are pretty much the "dreck" of the league. They are really awful basketball teams. ESPN projects them to combine to win 70. I would take the under in a heartbeat.
If the Magic still had VanGundy I'd pick them to win 40 games. I don't see why Nelson/Turkoglu/Afflalo/Harrington/Ayon should win 21 (which is reeeaally low) or less in the east.
You're right, 21 is very low. The Magic were 34-23 when Dwight went down, and finished 37-29. 3-6 without him to close the year - obviously just a small sampling of what to expect, but even directly translated, that's a 27 win pace. However, I think Gustavo Ayon is a sharp decline from Ryan Anderson (another very very underrated player who everyone thinks is tied up entirely in Dwight... This could very well be true. I believe the opposite, however) and the rest of their core is either not young, or downright old. Meaning Turkey & Al Harrington. It's hard to be optimistic about AAA outproducing JRich's production from last year, and less Dwight probably means more Big Baby. Tie in no Van Gundy (absence of him will cost them wins, I think) and you're left with a team that is going to be very bad. Further, they probably will be incentivized to tank, so you can't really not consider them downright losing games towards the end of the season to compete for the top lotto pot.
Detroit is pretty young, already won .38 last year and didn't change much. And then there's regression to the mean.
Should note they outperformed their pythag by 3 games. They also got positive returns by guys like Ben Wallace (that will be replaced by Andre Drummond - I see a net loss there) and are nearly bringing back the same other core. Regressing to the mean means there is a mean to regress to... For a lot of their players, their poor production is closer to their true value.

The Bobcats added Haywood and Sessions, two ~average players. BG is also better compared to what they had last year. 25 wins are possible
But this was a -15 PD team that was likely the worst team in history. It will take a fair amount to even move them to .25. Could they do it? Sure. I don't think they will, though.
Nuggets 50+
The one thing I don't like about the Nuggets is their center position, especially McGee.
Is McGee that bad?
TWolves
We don't know yet if Rubio will come back as good as he was before. The rest of the guards are garbage (I don't really expect Shved to be a positive, especially right of the bat, do you?), and so are their centers. It's the Timberwolves, and they're in the west. (did you say they might win 50?) I'd take under 41
They will have Pek for the full season. There is no more Millicic. They will have AK for the full season. There is no more Beasley.

Those two things alone will account for an amazing turnaround.
Crow
Posts: 10623
Joined: Thu Apr 14, 2011 11:10 pm

Re: ESPN wisdom of the crowd average win projections

Post by Crow »

Adelman and company were only -2 on actual wins - expected wins compared to Kevin McHale and company's -7 the previous season (McHale in Houston last season was even). But it was still -2 and Adelman's 5 year measure is an average of almost -1.

Counting the past for Roy and Kirilenko, the Timberwolves have 5 rotation guys with positive RAPM in their last season but Barea, Williams, Pekovic, Cunningham and Stiemsma had horrible RAPM estimates (-2 to almost -4) and they do have a few unproven guys.
EvanZ
Posts: 912
Joined: Thu Apr 14, 2011 10:41 pm
Location: The City
Contact:

Re: ESPN wisdom of the crowd average win projections

Post by EvanZ »

Bobbofitos wrote:
The one thing I don't like about the Nuggets is their center position, especially McGee.

Is McGee that bad?
-3.1 RAPM suggests he hasn't been very good.
EvanZ
Posts: 912
Joined: Thu Apr 14, 2011 10:41 pm
Location: The City
Contact:

Re: ESPN wisdom of the crowd average win projections

Post by EvanZ »

It wouldn't surprise me if Denver uses the following lineup quite often:

Lawson
Chandler
Iguodala
Gallinari
Faried

That could be an interesting small-ball lineup.
Crow
Posts: 10623
Joined: Thu Apr 14, 2011 11:10 pm

Re: ESPN wisdom of the crowd average win projections

Post by Crow »

Not sure what McGee's RAPM was in Denver alone. Someone tried to convince me he was better, especially in the playoffs, but I want to see his RAPM next season before I buy into too much.

Denver far better with Faried at Center than PF. http://www.82games.com/1112/11DEN12.HTM#bypos
Bobbofitos
Posts: 306
Joined: Sat Apr 16, 2011 7:40 am
Location: Cambridge, MA
Contact:

Re: ESPN wisdom of the crowd average win projections

Post by Bobbofitos »

EvanZ wrote:
Bobbofitos wrote:
The one thing I don't like about the Nuggets is their center position, especially McGee.

Is McGee that bad?
-3.1 RAPM suggests he hasn't been very good.
I'm curious how he performs this season. His conventional numbers look good.
J.E.
Posts: 852
Joined: Fri Apr 15, 2011 8:28 am

Re: ESPN wisdom of the crowd average win projections

Post by J.E. »

Bobbofitos wrote: However, I think Gustavo Ayon is a sharp decline from Ryan Anderson (another very very underrated player who everyone thinks is tied up entirely in Dwight... This could very well be true. I believe the opposite, however)
I'm one of Anderson's biggest fans, and http://stats-for-the-nba.appspot.com/pm/779.html seems to suggest that he can play well without Dwight (performing better than Dwight with the same 4 man units).
However, I think <25 wins is reserved for downright horrible teams full of rookies or players I don't know the names of. This isn't one of those teams.
Detroit is pretty young, already won .38 last year and didn't change much. And then there's regression to the mean.
Regressing to the mean means there is a mean to regress to...
It's usually a good idea to forecast all teams to converge towards _league mean_ (simply because it gives better prediction results), that's what I meant by that.

McGee's regular +/- was good in the playoffs, but it's not a large sample. Conventional numbers miss a lot of things; he led the league in goaltendings, for one. He's also the most overrated guy when comparing individual Reb% with actual rebounding impact (through rebounding RAPM) because he's not good at boxing out.
That being said, I think offensively he's a better fit in Denver (with Miller) than he was in WAS (with Wall)
Crow
Posts: 10623
Joined: Thu Apr 14, 2011 11:10 pm

Re: ESPN wisdom of the crowd average win projections

Post by Crow »

Does Portland win more than 35? I say no. Should they want to? I again say no given where they are.

Indiana to win 50? I think that is more likely 5+ too high than right.



After cutting 2 players recently, Houston will have to cut 4 more or trade them and take much fewer or no players back.
J.E.
Posts: 852
Joined: Fri Apr 15, 2011 8:28 am

Re: ESPN wisdom of the crowd average win projections

Post by J.E. »

On pinnacle you can now bet on "who will win the title". This probably doesn't perfectly correlate with their win predictions (once they come out) because ability to win a title is heavily influenced by willingness and (cap)room to make moves, but it should give us a good first guess on how Vegas ranks them

Code: Select all

Miami	Heat	3.28
Los	Angeles La.	4.12
Oklahoma	City	4.99
Chicago	Bulls	14.44
San	Antonio	16.22
Los	Angeles Cl.	28.84
Boston	Celtics	28.99
Dallas	Mavericks	33.65
New	York	36.52
Brooklyn	Nets	39.64
Indiana	Pacers	39.64
Memphis	Gizzlies	41.02
Philadelpia	76ers	61.25
Denver	Nuggets	79.82
Houston	Rockets	91.31
Atlanta	Hawks	97
Utah	Jazz	103.32
Minnesota	Timberwolves	109.33
Portland	Blazers	121.36
Orlando	Magic	139.38
Milwaukee	Bucks	151.41
New	Orleans	181.41
Golden	State	211.47
Phoenix	Suns	211.47
Cleveland	Cavaliers	241.5
Detroit	Pistons	241.5
Washington	Wizards	241.5
Toronto	Raptors	271.53
Sacramento	Kings	301.61
Charlotte	Bobcats	601.97
I like Boston @29, Dallas @33 and Indiana @40
Mike G
Posts: 6175
Joined: Fri Apr 15, 2011 12:02 am
Location: Asheville, NC

Re: ESPN wisdom of the crowd average win projections

Post by Mike G »

J.E. wrote:On pinnacle you can now bet on "who will win the title". This probably doesn't perfectly correlate with their win predictions (once they come out) because ability to win a title is heavily influenced by willingness and (cap)room to make moves, but it should give us a good first guess on how Vegas ranks them

Code: Select all

Miami Heat 	3.28
...
Charlotte Bobcats 	601.97
Are these numbers the odds against a team winning the title?
In other words, Miami's odds are 3.28-to-1, so probability of a repeat = 1/(3.28+1) = .234 ,
and Cha = .0017 ?

If so, the total likelihood of 30 teams = 1.04000
Does that 4% 'extra' represent the profit margin that Vegas anticipates?

I'd think the big difference between rankings by title probability and by season wins would be the Conference a team is in. The Lakers or Thunder could win more games than the Heat but have less chance to reach the Finals.
Crow
Posts: 10623
Joined: Thu Apr 14, 2011 11:10 pm

Re: ESPN wisdom of the crowd average win projections

Post by Crow »

Top 3 are a bit closer together here than I thought Vegas would have them based on previous information. Next 3-5 at first blush look a bit too long but probably aren't that bad estimates. Anybody that bets on a team beyond 14th except on the Hawks is foolish IMO. The Nets should probably be twice as long but if the bookies can get action from that fan base at those odds, I understand the motivation.
J.E.
Posts: 852
Joined: Fri Apr 15, 2011 8:28 am

Re: ESPN wisdom of the crowd average win projections

Post by J.E. »

If Rose can come back close to 100% they have 3 players in the RAPM top 20 and Tom Thibodeau. Not saying it's absolutely right to rank them #4, but I can sort of understand
Post Reply