The Andrew Extravaganza

The system had its best performance to date in predicting straight up winners, 41-14 or 74.5% correct. Against the spread, the system continued the trend of improvement and is approaching the excellent results of its debut pick set. Here's the table of results so far:

Week	Straight Up	SU%	Against the Spread	ATS%
9-22-07	33-15	68.8	27-20-1	57.4
9-29-07	37-16	69.8	23-29-1	44.2
10-06-07	35-19	64.8	28-24-2	53.8
10-13-07	41-14	74.5	30-24-0	55.6
Total	146-64	69.5	108-97-4	52.7

You can find a complete listing of this week's picks here.

The system for identifying best bets continues to evolve. To determine this week's best bets I have isolated 13 angles, rated each based on win percentage and sample size, and then applied them to every game on the schedule. This week 21 out of 52 games meet the base confidence threshold. Those 21 games have been divided into the familiar four levels of confidence.

I went back and applied this new system to the past four weeks of games and found that selected games went 53-22 against the spread. To some degree this is not surprising, the group was self-selecting as those games formed the foundation for the 13 angles. At the same time, it cannot be denied that the application of multiple angles has created a money making method of bet selection.

The challenge now is to see if this thing actually predicts the winners as well as it identifies past results. Also, new this week is a complete listing of qualifying bets; in weeks past I limited the number of best bets to eight. Here are the best bets for week eight:

Confidence	Away	Home	Point Spread	Prediction
	Florida	Kentucky	Florida by 6.5	Kentucky by 1
	Wake Forest	Navy	Wake Forest by 3	Navy by 1
	Michigan	Illinois	Michigan by 2.5	Illinois by 5
	Miami of Ohio	Temple	Miami of Ohio by 5.5	Miami of Ohio by 5
	Tennessee	Alabama	Pick'em	Alabama by 1
	Idaho	New Mexico State	New Mexico State by 9.5	New Mexico State by 10
	Vanderbilt	South Carolina	South Carolina by 13	South Carolina by 14
	Nevada	Utah State	Nevada by 7	Nevada by 6
	Kansas State	Oklahoma State	Oklahoma State by 3	Kansas State by 2
	Tulsa	UCF	UCF by 3	Tulsa by 4
	Utah	TCU	TCU by 3.5	TCU by 3
	Texas Tech	Missouri	Missouri by 3.5	Missouri by 2
	Auburn	LSU	LSU by 10.5	LSU by 17
	San Jose State	Fresno State	Fresno State by 12.5	Fresno State by 14
	Northern Illinois	Wisconsin	Wisconsin by 23	Wisconsin by 19
	Army	Georgia Tech	Georgia Tech by 24	Georgia Tech by 21
	Texas	Baylor	Texas by 25	Texas by 16
	Boise State	Louisiana Tech	Boise State by 16.5	Boise State by 12
	Houston	UAB	Houston by 13	Houston by 8
	Penn State	Indiana	Penn State by 7.5	Penn State by 2

FYI, this is how the different confidence levels performed when I went back and evaluated the 13 angles over the past 4 weeks:
Level 4, 8-1
Level 3, 17-3
Level 2, 21-13
Level 1, 7-5


Update: Results can be found here.

Kentucky by 1 as a Level 4 confidence pick. Nice.

In Kentucky's favor:
Home Dogs of less than 7 points are 20-7-0
Predicted outright upset winners are 15-11-1
Predictions that differ from the spread by 1.5 points to 10 points are 70-58-1

Three of the four 4 confidence level picks follow that same logic, which is strange because there have been only four such cases from the prior four weeks.