2015-09-18

Superbowl 50 Projection, Week 2, Thursday

San Francisco over Denver

This particular point in the season is almost always the most misleading, when 2 teams have played their second game and everyone else has played one. I should really fix this, and I can think of how to do it. See, I calculate a team's strength based on points scored and allowed, then regress it back to .500 by an amount based on how early in the season it is. That keeps the uncertainty high until at least a few weeks in. My means of regression is based on the number of games the team has played. That means Denver (and KC, but they're already exactly at .500 because they've scored and allowed the same number of points) only get regressed 14/16 of the way to .500, and every other team gets regressed 15/16 of the way.

Basically, Denver's raw score before regressing is roughly equivalent to the other teams that are at .513, not .526, so more along the lines of Green Bay and New England, and actually lower than Buffalo and Miami within the AFC. They are 2-0, and have the division lead, but they perhaps shouldn't be as favored in the rest of their games as they are, which would lower their playoff and conference championship odds. The simple way to fix this would be to make the divisor not 16 but 256, and regress everyone based on the total number of games that have taken place in the league as a whole. This would even out the bumps and dips teams get for playing Thursday games, having late or early byes, or not playing until Monday, and by the time bye weeks are over, every week's final results would be the same as the current method.

TL;DR: Ignore my numbers today; they're garbage! But after Monday I'll be back to telling you exactly who's going to win this thing.


TEAMRecordStrengthDivPlyffWCQtrSemiChamp
SF1-00.53133.2654.7139.7321.0211.125.75
NYJ1-00.52927.4949.6234.9718.149.414.92
TEN1-00.52949.8458.5441.0521.2710.985.74
CIN1-00.52848.2057.6840.7321.0410.835.65
DEN2-00.52646.4464.1747.6724.6612.686.60
CAR1-00.52640.3755.6340.4521.1811.045.66
BUF1-00.52425.3746.6632.3816.598.514.40
MIA1-00.52124.8844.8830.9015.738.014.12
ARI1-00.52029.3351.0136.0318.589.604.86
GB1-00.51346.0356.3339.2019.9610.125.05
NE1-00.51322.2643.1729.2214.627.333.71
SD1-00.50828.2044.8330.5715.147.473.75
STL1-00.50427.0847.4732.5316.238.113.98
ATL1-00.50432.8749.5234.3717.148.534.18
DAL1-00.50241.7351.9335.1917.468.644.22
KC1-10.50015.7432.0420.249.854.792.37
NYG0-10.49819.0028.5417.758.734.292.08
PHI0-10.49621.2829.2518.028.834.322.09
SEA0-10.49610.3423.5614.657.193.531.70
DET0-10.49221.0630.4218.939.204.472.14
PIT0-10.48718.9625.4315.077.133.371.62
CHI0-10.48717.1426.7316.197.783.731.77
HOU0-10.48718.2824.8714.676.943.281.58
BAL0-10.48417.7624.3414.186.663.131.49
NO0-10.48014.2123.5514.416.833.241.51
WAS0-10.47917.9926.1015.797.463.521.64
IND0-10.47615.9922.2812.825.932.741.29
JAC0-10.47415.8923.0013.356.142.821.32
OAK0-10.4729.6218.0310.534.842.221.03
TB0-10.47112.5622.6313.666.352.951.35
CLE0-10.47115.0820.4711.655.332.431.13
MIN0-10.46915.7722.6213.106.062.801.28

[Week 1]

No comments:

Post a Comment