2012 predictions

[EvanZ]

I should try plugging in my A4PM rating into this.

[Crow]

Might be worth a look.

I know you proposed on blend of EZPM and RAPM in this contest but it might be worth looking back instead of ahead to find what the optimal blend of those 2 would have been this season so far, or what an optimal blend of various forms of RAPM (prior-informed, not, multi-year, A4PM) would be.

I know Alex at SportsSkeptic looked at blend performances for past seasons.

[Mike G]

Evan's blend of ezPM and RAPM is listed as ez2, and it's fared the best of his 3 entries.

[Mike G]

Season at 59%

Code: Select all

Current W + Pyth% remaining          Pythagorean% x 66
3.85   Crow    4.59   ez2          3.61   Crow    4.43   KevinP
3.91   JohnH   4.75   ezPM         3.77   JohnH   4.57   ez2
4.12   MikeG   4.77   KevinP       3.98   dis     4.64   DSMok1
4.13   dis     4.79   DSMok1       4.04   J.E.    4.72   zRA
4.18   J.E.    4.94   zRA          4.06   MikeG   4.94   ezPM
4.42   Vegas   5.59   11Pyth       4.40   Vegas   5.51   11Pyth

[Crow]

Does pyth just suck relatively for this year for this use or is it not so good in general? Has anyone check other seasons?

[Mike G]

11Pyth is the avg team expected wins this year, minus last year's.
In other words, it's the avg error from just predicting every team to repeat last year's performance (in point differential).

[Crow]

Ok, thanks for the clarification. I didn't check or think that thru.

[Jeff Fogle]

Any update on these recently? I got in the habit of checking in on them every few days, and I'm having withdrawals (lol). Thanks in advance if Mike G. gets a chance to post an update...

[Mike G]

Wow, I just keep putting it off one more day. Lots of changes:

Code: Select all

3.56    JohnH       4.23    ezPM
3.63    Crow        4.29    DSMok1
3.74    dis         4.33    Vegas
3.75    MikeG       4.45    zRA
3.92    ez2         4.51    KevinP
3.94    J.E.        5.01    11Pyth

That's avg error from current wins plus pythagorean remaining games.

Here's the pure pythag% X 66

Code: Select all

3.43    Crow         4.02    KevinP
3.48    dis          4.10    zRA
3.61    MikeG        4.12    DSMok1
3.62    JohnH        4.22    Vegas
3.71    J.E.         4.46    ezPM
3.89    ez2          5.14    11Pyth

[Crow]

Could be a photofinish, affected by in some or many cases by "lower meaning" late season games and the posture of certain teams towards those games.

[Mike G]

At 74% of the season.

Code: Select all

Pythagorean% X 66 avg error        Current wins + pyth% remaining
3.62   dis      4.03   KevinP       3.87   MikeG    4.31   ezPM
3.62   Crow     4.17   DSMok1       3.92   JohnH    4.51   DSMok1
3.65   MikeG    4.37   Vegas        3.97   Crow     4.63   KevinP
3.78   J.E.     4.42   ezPM         4.01   dis      4.65   Vegas
3.79   JohnH    4.43   zRA          4.16   ez2      4.85   zRA
3.98   ez2      5.23  11Pyth        4.18   J.E.     5.09  11Pyth

[xkonk]

Just an off year for Vegas, or are you all quitting your jobs and gambling full-time?

[EvanZ]

Just an off year for Vegas, or are you all quitting your jobs and gambling full-time?

Good idea.

[Mike G]

Season 81% completed.

Code: Select all

Current wins + Pyth% remaining      Pythagorean% X 66
3.43   Crow    4.04   DSMok1       3.11   Crow    3.83   ez2
3.56   JohnH   4.10   ezPM         3.34   dis     3.87   KevinP
3.79   J.E.    4.15   Vegas        3.38   JohnH   4.07   Vegas
3.87   dis     4.37   KevinP       3.39   J.E.    4.24   ezPM
3.89   MikeG   4.61   zRA          3.56   MikeG   4.31   zRA
3.95   ez2     5.00   11Pyth       3.80   DSMok1  5.11   11Pyth

[Crow]

.2 bounces have been occurring in a week or less.