Sunday, November 6, 2011

न्फ्ल फोरेकास्ट्स - वीक #9

MARKGANZERSNFLPREDICTOR.BLOGSPOT.COM is a web site I developed to forecast American professional results, on a weekly basis, with the aid of a probability model and some statistical techniques.

Because this has been a VERY busy time of year for me, I've only forecasted three weeks out of the nine. Here are my picks for week #9

SEA @ DAL -- Dallas +8

ATL @ IND -- ATL +12

NYJ @ BUFF -- Buffalo by 4

CLE @ HOUS -- Hous by 11

SF49 @ WAS -- SAF49ers by 11

TBB @ NOS -- NOS by 1/2 points

MIA @ KC -- KCC BY 4.5

CIN @ TN -- CIN BY 4.5

DEN @ CAR -- CAR by 8

NYG @ NEP -- NEP BY 6

STL @ AZC -- AZC by 10

GBP @ SDC -- GBP BY 6.5

BAL @ PIT -- BAL by 7

CHI @ PHI -- PHI by 3


The model is only of academic interest, UNLESS, you are a bettor (as opposed to a gambler). A bettor is one who week in and week out makes cold, rational, calculating bets. For example, based on this analysis, the three surest victories (if one were to bet only victories, and ignore what the published spreads are) are:

Atlanta (by 12)
Houson (by 11)
San Francisco (by 11)
Arizona (by 10)
Dallas (by 8)
Caroline (by 8)
Baltimore (by 7)
Green Bay (by 6.5)
New England (by 6)
Cincinnati (by 4.5)
Kansas City (by 4.5)
Buffalo (by 4)
Philadelphia (by 3)
New Orleans (by 0.5)

NOTE: The spreads I've shown here are the "crude" model data for each time, but unadjusted for the "17 points slaughter rule" which says that for assessing "adjusted" margins of victory, all victories in excess of 17 points are to be treated as 17 point victories.

FURTHER NOTE: I've imputed "home field advantage" as being worth 3 1/2 points. Further work and analysis are needed in this arena.

FINAL NOTE: The "crude" margins of victory I've used for this analysis are about TWO POINTS PER GAME GREATER than those shown here, a mildly arbitrary choice to "adjust" for "the slaughter rule," noted above.

No comments:

Post a Comment