About CFB Matrix

Predictions

Week 13 Pre-season Predictions

By  | 

Week 13 Pre-Season Picks and Preview

As you know, I make all the game winner/loser picks in the pre-season using just my TLC formulas to simplify college football.  It runs between 75-78% correct.  This has been the best model over the last 7 years of picking straight up (SU) game winners in all of college football. The best year, my picks, all done before the season with no changes, finished 7 games behind the gold standard, Vegas, for the 2014 season.  And I use 3 metrics. Talent, location, coaching.

Follow me on Twitter @CFBMatrix  email dave@cfbmatrix.com or text 971-217-8419.  All questions and requests will go into my Pod for the People each week.

Quick chart overview:
Talent Gap – My talent rank difference for the teams playing.  Anything within 10 is a close match up.  Ex. Alabama has a 6.25 edge over USC.
Coach Effect – My running average of a coach’s ability to coach a team up or down by games.  Ex. Satterfield is +3 games per year, Harsin is -1.5.  Talent rules the Matrix, but Coach Effect is a strong tie break in Power 5 and should be followed closely in the Group 5 teams
OC/DC Grade – A new metric that is rooted in scoring efficiency and graded by the talent a OC/DC has versus the talent of their schedule.  Use it for close picks and for forecasting OVER/UNDER leans

Week 13 TLC Picks (Done 3/22/2016)

YTD Pre-Season Picks Record (556-186 = 74.9%)  Vegas Weekly SU Record (586-146= 78.97%)
#UpsetAlert Record: 36-62  SU 57-40-1 ATS
Over/Under Leans on OC/DC grade: 41-21

REMINDER:  November is the time I stop tailing the Upset Alerts and ATS on the Upset Alert dogs.  The Vegas pencil gets sharper with each week in November and it becomes harder to get the upsets and covers.

RRR: Talent levels nearly the same.  Location neutral so this pick fell on coaching.  Stoops>Strong.  OU OC > UT OC.  DCs about equal. Pick: OKLAHOMA

2016-week-13-winner-picks

Leave a Reply

Your email address will not be published. Required fields are marked *