October 11, 2022

Items worth noting:

• I’m aware the betting lines give much less recs this week than usual. It seems like my models and Vegas agree more often this year than in the past.

• The betting lines display format will soon get an update. (1) I will make it more explicit about how much to spare (not wager). (2) There will be limited options for scaling the target bankroll amount.

• We’re in Bye season! Keep an eye on the future schedule.

Week 5 comments:

• It was the first significant loss on betting lines (50%), despite a good start on the first couple games of the week. This is normal / expected— read the last article if you haven’t, and brace for week 6! Still, I blame Taysom. But really, the downs are part of the… ups and downs.

• All other models showed high predictivity, excepting RB (and kicker). QB had very high predictivity.

Items not worth noting but anyway…

• Most of my time went into overhauling the back-end… With my model uploaded on servers, it works well for a couple weeks but then gets too slow after week 4. Now I can more quickly make changes and check the results instead of waiting minutes for the change.

• The rest of my time will go into integrating new variables to the models, which I never got the chance to complete in the off-season. Especially for Kicker and the game-score models. Subvertadown is constant evolution, and thanks for supporting that!

Have a good week.

/Subvertadown