QB by coverage updated to this week. We adjusted the scales to make the MARGINAL rates more relevant to the model. Otherwise we are very confident in this ASPECT to the model. Also play DEN/LAC Under (will explain when full model is released).
We selected 12/16 match-ups last week, and the model correlated at 60.3% to outcomes (NOT point total/ATS, that hit 37%). There are still a few key injury adjustments to be made to this week's final selections (DET and NYG to name a few), but about 90% of the alterations have been accounted for/adjusted in the … Continue reading NFL Week 15 Picks
Here is how the Turnover likelihood adjustment to the model looks like for week 15. Full model to come after practice injury reports come out on Friday.
We need to wait for practice injury reports to release the full slate of selections, but below are the match-ups (NOT included in the objective model) to help us identify QBs mismatches. As a reminder the first set of columns indicate how frequently a team uses the respective coverage. The second set is a QB's … Continue reading NFL QB vs. Coverage Week 15 Model
By request, we added a few more columns. Mind you nDAVE is our base model using Football Outsiders DVOA, add 3 adjustments for Coach, Injury and Turnover Likelihood. Other columns included are subjective, yet something NOT reflected in the model. For example, net DVOA (or DAVE) gives us how well you expect a team to … Continue reading NFL Week 14 Picks
Here we go again, correlated at almost 46% last week. Remember "TO Likely" relates to THAT QB, and "Opp TO Li" relates to your opponent turning the ball over.
We are testing what we call the "Mayfield/Goff Adjustment" this week: adding a multiplier for QBs most sensitive to the pressure, relative to opponent's BLITZ (not pressure rate, on purpose). This, along with the product are the two columns to the right. Again players in yellow are either questionable and/or low volume, so take them … Continue reading NFL QB vs. Coverage Week 14 Model
Week 12 worked out quite well. As we have been, and will be, we will let our numbers (for the most part) do the writing for us. We are still planning to provide a deep dive on weights/how we got to this variables, but for now you can find a simple breakdown of the numbers … Continue reading NFL Week 13 Picks
As outlined last week in Turnover Likelihood, here is the breakdown of this adjustment we add to the of aggregate weekly model. This has correlated between 20-30% to actual turnovers. Full model to come shortly. Note, players with colored cells had adjustments made because of lack of volume.
Last week, despite multiple surprises at the QB position, we released our QB v Opponent Coverage Model. This model, although not included in our weekly aggregate model (yet), helps predict QB outcomes for the week by stratifying QB success by coverage and comparing it to their opponents likelihood to play x coverage. Below you will … Continue reading QB v Opponent Coverage Model – Week 13