December 27th, 2022
This is it, the final week!
Before everyone disappears, I want to thank you for joining this first year at the website. It was really an adventure, and in the end I think it worked close to how it was imagined. We’ll be adding some new features to the existing tiers— including some extra foresight for the lowest tier and extra insight for the rest— so I hope you’ll enjoy next year even more. Plus, we are due for a year of better NFL predictability and added value from rational streaming planning!
Looking at accuracy performance, I’m really glad that at least the last 6 weeks straight have been fantastic for predictability. The predictive models were more able to more clearly show an advantage in this span.
As I’ve written before, my main disappointment from 2022 was that the first 10 weeks were so measurably more random than the other 5 NFL seasons that I’ve been ranking before this year. While all the best ranking sources did much more poorly than usual, I would have loved my own models to stand out. The good news— besides the successes of the last 6 weeks— is that (as always) I already have some modifications in mind for the off-season. The #1 change I will make will be to improve the speed for adjusting to new season information. This year, we thought certain teams (e.g. Seahawks) would be awful and others (e.g. Rams or Broncos) would be great— but 2022 made it a struggle to adjust quickly. I think this can be done better.
We will also be adding some new functions to improve the site for next year. I would like to bring back “closest historical match”. I would also like to add a feature that helps explain “why” a player is ranked higher or lower. I’m hoping we’ll also be able to include a way to ask whether certain injuries are accounted for. Meanwhile, the articles will finally get the updates that are due, and there will be some new articles as well. So there’s still a lot more work ahead, starting in January. Stay tuned!
Every year I’m asked if I will continue for the “junk week” of fantasy. The answer is the same as usual: No, the statistical models are not suitable to this week. (Your fantasy league should not be competing during this week!) However, I will also post a simplified list of D/ST and Kicker streamers, based solely on betting lines, when the time comes.
Tagged underAccuracy , Current Season , Expectations