Perhaps the best way to think about deploying analytics in trying to predict the future is imagining it as card counting. When gamblers count cards they aren’t going to win every hand, even when they know that the deck is long on face cards. They’re just trying to tilt the odds in their favor. The graph below on Cam Newton’s upcoming schedule can be viewed in the same light. It looks like the next 5 weeks (including the bye) could be difficult for Cam. But that doesn’t mean that he couldn’t exceed the projected points that I have for him. If I go back to the blackjack analogy, it’s like Cam is facing a deck short on face cards over the next 5 weeks. His schedule goes: ATL, SEA, BYE, DAL, CHI.
If I had to offer one area where there exists the most room for error in my forecast, I would say that it’s probably related to the forecast drawing from Cam’s season to date, and he may have played below the level he’s capable of playing. If he had played better to date (closer to last year’s Cam) then the projection over the next 5 weeks would still be lower, just not as low as I currently have him forecast. My forecast is based on comparables and Cam’s comps aren’t looking great right now. He went from being a 2004 version of Duante Culpepper, to being a 2005 version of Duante Culpepper. That’s not the direction he wants to be going.