Computer Rating System Prediction Results for College Football (NCAA IA)

2014 Retrodiction Results

Through
Rank System Pct. Correct Against Spread Absolute Error Bias Mean Square Error games suw sul atsw atsl
1Logistic Regression0.834210.0000015.3849-1.7076639.42376063412600
2Sagarin Elo0.832890.0000011.38970.7004208.56276063312700
3Least Squares w/ HFA0.826320.000009.28850.0512135.15176062813200
4Payne Power Ratings0.815790.0000011.0928-0.2774195.87676062014000
5Massey Concensus Rank0.814470.0000011.02120.9871197.71776061914100
6Sagarin0.809210.0000010.55210.5754179.99576061514500
7System Average0.806580.0000010.70800.7210184.02076061314700
8system Median0.802630.0000010.64450.7746182.14276061015000
9Beck Elo0.802630.0000011.14330.3585199.18576061015000
10Least Squares0.800000.0000010.45060.8218173.85476060815200
11Stortrends0.766120.0000011.50450.2327208.12166751115600
12CPA Retro0.793420.0000011.74210.7469229.05176060315700
13CPA Rankings0.792110.0000010.53700.7913176.04076060215800
14NutShell Combo0.790790.0000011.55420.4336223.36376060115900
15Born Power Index0.790790.0000010.98841.4090192.65976060115900
16Sagarin Predictive0.789470.0000010.62230.5138179.15276060016000
17PerformanZ Ratings0.789470.0000011.28910.8602198.88476060016000
18NutShell Sports0.789470.0000011.96730.1682231.62576060016000
19Sonny Moore0.786840.0000011.16380.9848196.18676059816200
20Edward Kambour0.780260.0000010.67161.1649182.11976059316700
21Laz Index0.778950.0000010.85800.6309189.53876059216800
22Covers.com0.778950.0000011.82710.0870224.36976059216800
23Ashby AccuRatings0.773680.0000010.93951.7062193.15776058817200
24Nutshell Girl0.771050.0000011.80990.6992237.48476058617400
25Pigskin Index0.765790.0000011.16970.8485199.92576058217800
26Stat Fox0.760530.0000011.12371.9051192.92876057818200
27SuperList0.756580.0000012.54720.7683256.46976057518500
* This system does not make predictions.  I make predictions for this
  system by translating it to a new scale that allows for making predictions.



Retrodictive records are found by taking the ratings from the current week
and applying them to the entire season to date.

The ideal system would be one that has the highest correct game decisions,
has the smallest mean error(deviation from the actual game result), and has
a bias of zero.

Mean Error = average[abs(prediction-actual)]

      Bias = agerage(prediction - actual)

      Std. = Standard Deviation of individual game biases