BIG BLUE FANS FOR
2010-11 Season Analytical Writings The 2010-11 Regular Season is now in the books, and it is time to evaluate the effectiveness of the NGE predictions vs actual outcomes. Each UK game prediction included a UK score, an Opponent score, a margin, and an Over/Under prediction. The prediction methodology has been the NGE Methods applied to the team statistics as they existed immediately prior to the game, including the use of Pomeroy's Strength of Schedule measurement, also taken the day of the game. The UK predicted scores ranged from a high of 92 points against Mississippi Valley State to a low of 68 points at Florida. The average predicted UK score was 77.3 points. The actual UK scores ranged from a high of 91 points against Coppin State to a low of 64 points at Tennessee in the last game of the season. The average UK points in the 30 games was 76.8 points. The variation between predicted and actual scores had an average difference of 0.5 points. The minimum variation between predicted and actual was -13 points at Georgia in the first SEC game of the 2011 season, and the maximum variation was 10 points against Boston University, and against South Carolina in Rupp. The Standard Deviation of this difference was 5.9 points. The UK Opponent predicted scores ranged from a high of 79 points against Washington at Maui to a low of 50 points against Mississippi Valley State. The average predicted UK Opponent score was 62.5 points. The actual UK Opponent scores ranged from a high of 84 points by Connecticut at Maui to a low of 44 points by LSU. The average UK Opponent points in the 30 games was 64.3 points. The variation between predicted and actual scores had an average difference of -1.8 points. The minimum variation between predicted and actual was -20 points at Portland in the second game of the season, and the maximum variation was 23 points against Mississippi State in Rupp. The Standard Deviation of this difference was 9.7 points. The predicted game margins ranged from a high of 42 points against Washington in Maui to a low of -4 points at Louisville. The average predicted game margin was 14.8 points. The actual game margins ranged from a high of 38 points against LSU to a low of -17 points against Connecticut in Maui. The average game margin in the 30 games was 12.5 points. The variation between predicted and actual game margins had an average difference of 2.2 points. The minimum variation between predicted and actual was -25 points against Connecticut in Maui, and the maximum variation was 23 points at Portland in the second game of the season. The Standard Deviation of this difference was 11.2 points. The predicted game total points scored [Over/Under] ranged from a high of 156 points against Mississippi Valley State to a low of 133 points at Alabama. The average predicted Over/Under was 139.8 points. The actual Over/Under ranged from a high of 164 points against Mississippi State to a low of 122 points at Tennessee in the last game. The average Over/Under in the 30 games was 141.1 points. The variation between predicted and actual Over/Under had an average difference of 1.3 points. The minimum variation between predicted and actual was -26 points against Mississippi State, and the maximum variation was 18 points at Tennessee in the last game of the season. The Standard Deviation of this difference was 11.5 points. The NGE Methods predicted the correct winner in 22 of the 30 games. In two games, the NGE Analysis indicated a likely UK loss, but UK won twice, and six times, the analysis indicated a UK win, and UK lost. The NGE analysis indicated game margins of 10 points or less for 14 of the 30 games, and the results were 6 right and 8 wrong for those 14 games. The NGE analysis indicated game margins of more than 10 points for 16 games, and the results were 16 right and no wrong for those 16 games. Submitted by Richard Cheeks
Submitted by Richard Cheeks
Go Back Copyright 2010 |