View previous topic :: View next topic |
Author |
Message |
Mountain
Joined: 13 Mar 2007 Posts: 1527
|
Posted: Thu Nov 27, 2008 3:58 pm Post subject: The Lakers and hot November Projections by Neil Paine |
|
|
Analysis from the Basketball-Reference.com blog
http://www.basketball-reference.com/blog/
http://tinyurl.com/5z8nmr
More on the value of small early samples (at least for the very hottest out of the gate):
"All told, the correlation of a team’s November SRS and its final winning % is 0.81, and the correlation of November SRS and final SRS is 0.84, which isn’t bad at all for just a month’s worth of games (it also backs up what we observed earlier in terms of each month’s predictive value). This is good news for Laker fans, as it basically ensures (barring injury) that Los Angeles will be a title favorite by season’s end — in NBA history, 84% of the teams leading the NBA in SRS after November went on to finish in the top 3 in final SRS, and 55% finished #1 overall. In other words, this is definitely one case where a small sample can tell you a lot about the NBA’s big picture." |
|
Back to top |
|
|
NickS
Joined: 30 Dec 2004 Posts: 384
|
Posted: Thu Nov 27, 2008 4:16 pm Post subject: |
|
|
I hate to admit it, since I'm not a Lakers fan, but what's not to like about the Lakers?
They're well coached, they have a star player who can win games in the 4th quarter. They're big and athletic. They're deep. The have solid 3-point shooting. They have a core of players that have experience playing together, but they aren't an old team.
They look scary this year. |
|
Back to top |
|
|
deepak
Joined: 26 Apr 2006 Posts: 664
|
Posted: Thu Nov 27, 2008 8:17 pm Post subject: |
|
|
He says that the correlation between Nov. SRS and the final win% is 0.81.
What about Nov. win% and final win%? |
|
Back to top |
|
|
Neil Paine
Joined: 13 Oct 2005 Posts: 774 Location: Atlanta, GA
|
Posted: Thu Nov 27, 2008 9:28 pm Post subject: |
|
|
deepak_e wrote: | What about Nov. win% and final win%? |
0.802 |
|
Back to top |
|
|
deepak
Joined: 26 Apr 2006 Posts: 664
|
Posted: Fri Nov 28, 2008 1:22 am Post subject: |
|
|
davis21wylie2121 wrote: | deepak_e wrote: | What about Nov. win% and final win%? |
0.802 |
Maybe I'm not understanding these numbers correctly. Doesn't that essentially mean that Nov win% and Nov. SRS are about the same in predicting final win-loss record? I would have thought that taking into account strength of schedule would increase the correlation coefficient more substantially. |
|
Back to top |
|
|
Neil Paine
Joined: 13 Oct 2005 Posts: 774 Location: Atlanta, GA
|
Posted: Fri Nov 28, 2008 12:44 pm Post subject: |
|
|
I think it speaks more to the fact that the schedule a team faces in November is representative of the schedule they'll face all season long. |
|
Back to top |
|
|
Ilardi
Joined: 15 May 2008 Posts: 263 Location: Lawrence, KS
|
Posted: Fri Nov 28, 2008 2:01 pm Post subject: |
|
|
For what it's worth, I decided to see how well last season's adjusted plus-minus ratings (http://82games.com/ilardi2.htm) would predict the Lakers' team efficiency this season, based on the actual player minutes distributed through the team's first 13 games. Assuming no improvement (or decline) for any player from 07-08, we get a projected per-game point differential of roughly +10.1 ppg, which is considerably below the team's actual margin of +14.3 ppg. However, given that players tend, on average, to improve up until roughly age 25 (and to decline starting around age 30), we could expect improvement this year from Bynum, Ariza, Farmar, and Vujucic (and decline from Fisher) . . . depending upon our assumptions there, we might get an adjusted projected ppg differential for the Lakers of up to +11 ppg. And, given an inevitable "regression to the mean" effect for the team over the next 69 games (there's no way they'll actually sustain a margin of +14.3 ppg), this projection (an aggregate margin of somewhere between +10 and +11 ppg) sounds just about right to me. |
|
Back to top |
|
|
gabefarkas
Joined: 31 Dec 2004 Posts: 1313 Location: Durham, NC
|
Posted: Mon Dec 01, 2008 11:57 am Post subject: |
|
|
Ilardi wrote: | ... depending upon our assumptions there, we might get an adjusted projected ppg differential for the Lakers of up to +11 ppg. And, given an inevitable "regression to the mean" effect for the team over the next 69 games (there's no way they'll actually sustain a margin of +14.3 ppg), this projection (an aggregate margin of somewhere between +10 and +11 ppg) sounds just about right to me. |
I know you started with the caveat "depending upon our assumptions", but I'm curious what kind of assumptions/adjustments you made? Did you just look at the average (assumed) rise in APM for age X to age X+1 for the four risers and use that instead? |
|
Back to top |
|
|
Ilardi
Joined: 15 May 2008 Posts: 263 Location: Lawrence, KS
|
Posted: Mon Dec 01, 2008 1:18 pm Post subject: |
|
|
gabefarkas wrote: | Ilardi wrote: | ... depending upon our assumptions there, we might get an adjusted projected ppg differential for the Lakers of up to +11 ppg. And, given an inevitable "regression to the mean" effect for the team over the next 69 games (there's no way they'll actually sustain a margin of +14.3 ppg), this projection (an aggregate margin of somewhere between +10 and +11 ppg) sounds just about right to me. |
I know you started with the caveat "depending upon our assumptions", but I'm curious what kind of assumptions/adjustments you made? Did you just look at the average (assumed) rise in APM for age X to age X+1 for the four risers and use that instead? |
Yes, more or less, but I haven't yet conducted a systematic study of APM as a function of age (it's on my to-do list!), so they were simply educated guesses . . . |
|
Back to top |
|
|
|