Third Pick Week Selections
Third Pick Week Selections and Post-Mortem on Last Week
Last Week Post-Mortem: Last week could not have been uglier from a picks standpoint. My analytics went 1-7 on official picks and 2-8 if the two games that went off the board were included. Monkeys can throw darts at a pick sheet and do better than that! Any confidence that many of you had in my methodology likely evaporated during last week’s debacle. That being said, I spent much of Sunday dissecting what went wrong and I believe I found an answer.
One could argue that injuries were a factor or unbelievably unlucky bad play, especially in the Boise State and Texas A&M games. Bowling Green had a back-door defeat via a late, meaningless TD by Massachusetts. But these would be just excuses for a bad week. During testing I never had the benefit of an injury report and the results were still outstanding. So what happened?
During 13 weekly observations (albeit 12 in testing) my analytics never had a week worse than 7-5. Further, my picks lost last week at virtually the same percentage that they had previously won by during 13 weekly observations. However, something very subtle happened last week that had never happened before and I didn’t pick up on it until after the fact. Now that I know how to handle circumstances similar to last week I am confident that I won’t experience another debacle similar to last weekend.
My analytical process is two-fold: 1) screen first for games where the actual outcome will likely be a wide miss from the line. The output for line misses is what I call a “strainer number”. It determines which particular games I find attractive. Last week the majority of games had a negative strainer number, which was unprecedented in my experience with this system. By doing a quick look-back at 2013, there were on average only two negative strainer numbers per week in 2013. I have been using the absolute value of the strainer number then moving to step 2) which is the art of determining the most consistent team.
Why is this significant? As I stated in the paragraph above, during the 2013 season only two games per week had negative strainer numbers…and I lost almost ALL of those games. As a result I now know that the actual value of the strainer number is very significant piece of information. If the strainer number is positive, take the most consistent team. If, however, the strainer number is negative, then take the team with a history of more volatile performance. In the future I will operate under these new guidelines.
For many of you who have been following this site I’m sure your trust has been dented at a very minimum. You can wait-and-see, but I still have a high degree of confidence in this methodology.
This week’s picks
There are 14 total picks this week with a breakdown of 9 favorites and 5 underdog or pick ‘em games. This favorite to dog mix is consistent with what I experienced during my 2013 back-test. That being said, four of the underdog picks were by small margins so I am putting them at the bottom of my confidence scale. Nevertheless here is list in its entirety.
10/2/2014 -6.5 Florida Atlantic @Florida Intl 7:00 PM EDT
10/4/2014 -6.5 @Virginia Pittsburgh 7:30 PM EDT
10/4/2014 -7 @Michigan State Nebraska 8:00 PM EDT
10/4/2014 +12 UTEP @Louisiana Tech 7:00 PM EDT
10/4/2014 -14.5 @Colorado State Tulsa 3:00 PM EDT
10/4/2014 Pick Miami-FL @Georgia Tech 7:30 PM EDT
10/4/2014 -17 Baylor @Texas 3:30 PM EDT
10/4/2014 -4.5 Navy @Air Force 3:30 PM EDT
10/4/2014 -3.5 @Washington State California 10:30 PM EDT
Slightly lower margins/confidence picks but are still official picks
10/4/2014 +14 North Carolina State @Clemson 3:30 PM EDT
10/4/2014 -17 @Louisiana-Lafayette Georgia State 7:00 PM EDT
10/4/2014 -3 Central Michigan Ohio 3:30 PM EDT
10/4/2014 +3.5 @Syracuse Louisville 7:00 PM EDT
10/4/2014 +1 @North Carolina Virginia Tech 12:30 PM EDT
*10/4/2014 +12.5 Texas Tech Kansas State 7:00 PM EDT
*Pending status of Texas Tech QB Webb before game
Year To Date ATS
8-9
