LBJackal wrote:GotowarMissAgnes wrote:Predicting everyone would be a pain, but I'd go further than just predicting a handful of tough to predict players. I'd do something like predicting the top 100 or 200 players and maybe include another 25 or 50, youngsters, maybe.
There are several different tests of prediction you could use. A common one is the Mean Squared Predictive Error, which is a fancy term for "Take how far off you were, square it, add it up, and divide by the number of observations"
Like Tavish said it would be good to have some sort of weighted system where players who were tough to predict are given more weight. Or maybe not more weight, but if you get them right you get more than a normal "right" prediction. And if you get a simple one wrong, you get less points than a normal "wrong" prediction. Just an idea, and really any system would be fine by me.
Unless you could develop an objective system ahead of time that fairly weighted players based on difficulty, I'd pass on that. The idea has merit, but creating the system, getting everyone to agree that it's fair, and then applying it to all the players is pretty messy.