Fantasy Football — Week 12 AI Performance and Recap
Clustered at the top
TLDR: We had a slightly higher score then ESPN and 4for4, not a ton of variance between the three sites.
Recap of Big Miss and Hits
- James Conner and Alvin Kamara — We were low on these guys and it turned out to pay dividends
- JuJu Smith-Schuster /Antonio Brown — I didn’t watch the game so I am not sure what went wrong
- Sterling Shepard — I thought he might be a sneaky play in DFS due to the Eagles secondary being a train wreck. He didn’t do anything.
About the Graphs
- AI Closer are the players we had the closer value at the end of the week
- ESPN Closer are the players ESPN had the closer value at the end of the week
- Read the legend for info on the colors
I will be looking at the performance between fantasy sites and Gridiron AI.
Performance Baseline — I am going to use the following player pool to try to measure week to week performance in our model against ESPN:
Finished in Top 50 WR and RB | Finished in Top 16 TE |Finished in Top 20 QB
What does the table below mean:
- Score — Total point difference between projection and actual
- Correct — Projected Rank equals the actual rank
- Hybrid — 1 point if the Rank is correct, .5 points if Rank is within five positions of the actual Rank
- We’re all pretty clustered together, our score is slightly worse but that isn’t always a bad thing. We are predicting outcomes in different ways. Next season, I’ll try to automate a way to highlight what each site is doing well and doing poorly.