After two weeks of right around 60% on all games, the GRT Statsy Preview Machine took a pie in the face this week, going only 7-12 (36.84%) overall, 1-5 (16.67%) in Category 2, and 0-4 ( let me see . . . 0%) in Category 3. For the season, the Machine is now 267-235 (53.19%) in Category 1, 100-83 (54.64%) in Category 2, and 54-43 (55.67%) in Category 3.

Because the Machine struck out on Category 3 games, it also struck out on both “Category 4” games. For the seven weeks we’ve been tracking Cat 4 games, they are 19-5 (79.16%).

The Machine had an identical result for all games using mid-week spreads: 7-12 (36.84%).

SP+, meanwhile, had a solid week, going 11-8 (58%) officially, using its own mid-week spreads. For the season, SP+ is 258-243-7 (52%). It had the same results against opening spreads (11-8, 57.89%) and is now 280-222 (55.78%) for the season against those spreads.

Below are the GRT Statsy Preview Machine’s picks for the 2020 college football bowl season. As always, if you’re wondering why we do this or what I mean when I refer to “confidence” and when I place game predictions into different categories, check out this post. Also, in case it’s not perfectly clear from the above results, spreads matter.

GRT SPM 2020 Bowl Picks

Ignore the Home and Away column headings again this week, as bowls are all neutral site games. I’ve accounted for them in the calculations, but haven’t denoted them in the table.

Bowl season features seven Category 3 games, those that the GRT Statsy Preview Machine likes particularly well. Of those, the following five make the Category 4 cut because they also agree with SP+:

  • Iowa vs. Missouri (Iowa -13.5)
  • Tulsa vs. Mississippi State (Tulsa -1.5)
  • Marshall vs. Buffalo (Marshall +3.5)
  • North Carolina State vs. Kentucky (North Carolina State +2.5)
  • Texas A&M vs. North Carolina (North Carolina +5.5)

I’m really interested to see how the Statsy Preview Machine does in this weird season where there were very few, if any, data points for cross-conference play. I’m hoping that its focus on how a team does relative to what its opponents usually do does a good job of predicting how it will do against any opponent regardless of conference affiliation or schedule strength. We’ll see.