Ways to Predict NFL Ownership

by Jon Bales (JonBales) Last Updated 2019-08-08 04:22:31 Premium Content Part of the product Using Game Theory in Daily Fantasy Football

As I’ve mentioned, I believe the biggest knock on the contrarian movement is that predicting ownership can be a difficult and even fragile practice. It is true that small shifts in player value can dramatically shift ownership. When that backup running back who is min-priced and set to see a full workload, we know he’s going to be highly owned, but it will also throw off the ownership of other players. Little pricing quirks and week-long narratives can also shift ownership in somewhat unpredictable ways.

But we don’t need to get it perfect. We just need to be able to predict ownership with a decent level of accuracy for it to hold value. And I think you can make a very strong case that ownership is much, much easier to project than player performances on a weekly basis. Certain weeks, we know beyond any doubt that a particular quarterback or running back is going to be in the top-three highest-owned players, and we can often predict which player will be the most popular. Try predicting player performance in the same way… good luck. We don’t really have a great idea how players are going to produce, but we can fairly accurately forecast how the crowd thinks they’re going to produce.

One thing that’s different about projecting ownership, though, is that it’s just as much of an art as a science right now. All the factors work together, so I don’t think it’s necessarily something we can model. Oh, don’t get me wrong, I’m trying. I’m working like a dog to try to create an algorithm to predict ownership, but it’s just a challenge to model things like when Adam Schefter talks highly of a player on a Sunday morning or how weird shifts in pricing from week to week will affect how the crowd views value.