2012 predictions

[Mike G]

At 1/3 of the season, with Pythagorean irrespective of current wins.

Code: Select all

Current wins + pythagorean W%
4.72   JohnH      6.24    zRA
5.05   Crow       6.25    ez2
5.32    J.E.      6.42    ezPM
5.43   MikeG      6.55   DSMok1
5.67   Vegas      6.59   KevinP
5.80    dis       6.73   11Pyth

Pythagorean only
4.77   JohnH      6.00    zRA
5.07   Crow       6.20    ez2
5.30    J.E.      6.38    ezPM
5.40   MikeG      6.47   11Pyth
5.80    dis       6.57   DSMok1
5.86   Vegas      6.72   KevinP

[Mike G]

... one more argument for the peak basketball age being later than 26

... doesn't mean they became "better" after 28, but that their priorities changed.

On the other hand, I just noticed that b-r.com now lists in their "Miscellaneous Statistics" the minutes-weighted Age of each team.
http://www.basketball-reference.com/lea ... _2012.html

The NBA average minute is being played by a 26.8 yr old.
Of 12 teams older than that, 11 have winning records (Phx, at 28.9, is 8-14), and 10 have positive SRS (Orl is -1.1)

Of 11 teams with SRS > 3, only 2 are younger than avg -- Philly and OKC.
Both teams play around 25 y.o.

[Crow]

Thanks for mentioning the B-R minutes weighted age data. That is handy to have available.

[Crow]

Orlando the worst FT% team in 33 years and just outside the bottom 10 of all-time. Near bottom third or in it on at the rim shots and inside FG%. 2nd lowest on mid-range attempts but the worst at them. The offense is mainly just 3 point shooting.

At lineup level the Magic's starting lineup is one of the worst performing for lineups used 150+ minutes on raw and adjusted +/-. There are a number of next most used lineups doing terrific. Make a change? Nelson's injury forced one and that may help them.

[Mike G]

The avg error is down to 5.56

Code: Select all

4.48   JohnH      5.89    ez2
4.75    Crow      5.98    zRA
5.02    J.E.      6.10   KevinP
5.25   MikeG      6.12    ezPM
5.49    dis       6.41   DSMok1
5.68   Vegas      6.58   11Pyth

There are 7 teams doing better than any of us guessed, and 9 which are doing worse.
This leaves 14 teams that are within the range of our predictions.
Ranked by avg error (either too high or too low) -

Code: Select all

All predicted too high             All predicted too low
tm    pyth+  closest     Avg      tm    pyth+   closest   b   Avg
Cha    8.9   4   JH     12.1      Phi   49.4   10   JE       13.5
Det   14.4   7   ezPM    9.6      Atl   42.9    6   ezPM      8.9
Was   12.5   4   ezPM    8.4      Min   34.4    1   KP        8.4
NOH   17.3   4   MG      7.9      Uta   35.0    5   MG,KP     7.9
NYK   28.8   4   ezPM    7.9      Ind   41.3    3   MG        7.0
Sac   17.3   3   JE      7.7      Por   42.3    3   Crow      5.1
NJN   20.0   2   dis     5.3      Hou   35.3    1   JE,Crow   3.6
Orl   36.8   1   Crow    3.7                  
Mil   29.1   1   MG      2.9                  

EDIT: This has been corrected since originally posted.

[Mike G]

After a 10-game slate, avg error down to 5.45

Code: Select all

4.52   JohnH      5.80   KevinP
4.74    Crow      5.92    ezPM
4.95    J.E.      5.94    ez2
4.97   MikeG      6.14    zRA
5.31    dis       6.22   DSMok1
5.44   Vegas      6.40   11Pyth

Season at 37% (of 990 games).

[Mike G]

Season at 39%, avg error down to 5.26 per participant, for both Pythagorean wins projections and for current wins + Pyth% of remaining games.

Code: Select all

Current wins + Pythagorean remainder
4.43   JohnH      5.69    ez2
4.60    Crow      5.71   KevinP
4.74   MikeG      5.73    ezPM
4.82    J.E.      5.88    zRA
5.17    dis       6.07   DSMok1
5.29   Vegas      6.13   11Pyth

Pythagorean only
4.45   JohnH      5.57    zRA
4.58   MikeG      5.60    ez2
4.68    Crow      5.77    ezPM
4.73    J.E.      5.81   KevinP
5.24    dis       5.85   11Pyth
5.54   Vegas      5.99   DSMok1

[Crow]

Lots of contestants doing well this season.

[Mike G]

Just 4 games last night, and things got interesting.

Code: Select all

Current wins + Pyth% * remaining         Pythagorean% * 66
4.46   JohnH     5.61    ez2          4.49   MikeG    5.59    zRA
4.59    Crow     5.65    ezPM         4.53   JohnH    5.60    ez2
4.62   MikeG     5.68   KevinP        4.71   Crow     5.77    ezPM
4.81    J.E.     5.80    zRA          4.81   J.E.     5.80   KevinP
5.06    dis      5.93   DSMok1        5.20   dis      5.89   11Pyth
5.25   Vegas     6.12   11Pyth        5.56   Vegas    5.92   DSMok1

[Crow]

Given you have standings by the above two methods, it might be interesting to see the standings based on a straight projection of actual wins to date as well.

[Mike G]

With season at 41.7%, avg error is 4.89

Code: Select all

Current + Pyth% of remaining         Pythagorean % of 66 games
4.20   Crow     5.19   ezPM         4.38   Crow     5.32   Vegas
4.27   JohnH    5.28   ez2          4.39   MikeG    5.41   KevinP
4.32   MikeG    5.38   KevinP       4.42   JohnH    5.51   zRA
4.42   J.E.     5.56   DSMok1       4.49   J.E.     5.53   ezPM
4.57   dis      5.62   zRA          4.72   dis      5.62   DSMok1
4.96   Vegas    5.88   11Pyth       5.30   ez2      5.86   11Pyth

The worst of our guesses, listing just the worst for any team:

Code: Select all

err   tm   prediction  proj.
20   Cha     KP    28    8
15   NOH     KP    31   16
15   Phl     MG    33   48
11   Det     DSM   28   17
11   LAL    Vegas  48   37
11   Min   ezPM,JE 22   33
11   NJN     KP    29   18
10   Atl     KP    31   41
10   Chi     DSM   41   51
10   Sac     MG    29   20

While Kevin P seems to dominate the worst predictions, he's also got the best guess (or virtual tie) on Mem, Mia, Min, Phx, and Uta.
Crow has only one worst guess: picking Mem to win 40, and they're headed for 33-34.

[Mike G]

With .445 of season gone, avg absolute error is 4.51; pyth is 4.56

Code: Select all

Current + Pyth% remaining games      Pythagorean% of 66 games
3.80   JohnH    4.86   ez2          3.94   J.E.      4.73   ez2
3.91   MikeG    4.88   ezPM         3.98   JohnH     5.03   zRA
3.98   Crow     5.18   DSMok1       4.02   MikeG     5.05   KevinP
4.02   J.E.     5.18   KevinP       4.06   Crow      5.07   ezPM
4.16   dis      5.22   zRA          4.11   dis       5.21   DSMok1
4.44   Vegas    5.52   11Pyth       4.72   Vegas     5.58   11Pyth

[Mike G]

Season is at 50.3%, and teams are still sliding toward their expectations, in general.

Code: Select all

Current wins + Pythagorean             Pythagorean only
3.77   MikeG    4.54   ez2          3.65   Crow     4.49   Vegas
3.79   JohnH    4.67   ezPM         3.85   dis      4.63   KevinP
3.81   Crow     4.89   KevinP       3.97   JohnH    4.78   zRA
3.95   dis      4.91   DSMok1       4.00   J.E.     4.83   ezPM
4.03   J.E.     5.06   zRA          4.03   MikeG    5.02   DSMok1
4.34   Vegas    5.50   11Pyth       4.44   ez2      5.63   11Pyth

If you're higher in the left columns than in the right, then you've been lucky (or you're influencing games).
Avg error now 4.34 in Pyth+

The most shocking team is Cha, headed for 9 wins. Hollinger is closest at 13. On avg, we are off by 12.4
The only other team with such a large best-guess error is Hou, headed for 38. Crow and JE guessed 34.

Philly's projection (41) is now within 3 of JE's guess.
The sinking Celts (33) are 3 below KP's bottom guess.
The recovering Pacers (41) are 3 better than my high guess. I also had lowest hopes for Mil, and they're 3 wins below that (27).
Nets (19) look 3 wins worse than dis guessed, and Okl (48) looks 3 wins better than any but Vegas.

With an 0-5 night, the East is now less than .400 vs the West, for what it's worth.

[Mike G]

At .543 of the season's schedule:

Code: Select all

Current W + Pyth% of remainder        Pythagorean W% X 66
3.76   JohnH    4.52   ez2         3.49   Crow     4.44   ez2
3.80   Crow     4.75   ezPM        3.74   JohnH    4.44   KevinP
3.99   MikeG    4.75   DSMok1      3.85   J.E.     4.74   zRA
4.00   dis      4.79   KevinP      3.91   dis      4.78   DSMok1
4.10   J.E.     4.98   zRA         3.98   MikeG    4.88   ezPM
4.30   Vegas    5.36   11Pyth      4.33   Vegas    5.53   11Pyth

[Crow]

If Hollinger beats me it could be considered to come from my not expecting Charlotte and New Orleans to be as awful as he did. My low expectation of Orlando might figure into it as well or make the ultimate difference.

Of course several others are still close enough to beat either of us.