My first pass is actually attacking daily fantasy (i.e. given data about past NHL games, predict a player's performance in an upcoming game). I'm scraping some

My first pass is actually attacking daily fantasy (i.e. given data about past NHL games, predict a player's performance in an upcoming game). I'm scraping some NHL stats sites to build input vectors based on a couple dozen manually-chosen features (e.g. "number of goals scored last game", "opposition goalie's save percentage in his last 5 games") and passing them through a neural network to see if there's any usable correlation between the input features and a player's eventual performance (goals, assists, +/-, etc.). It's all in good fun -- mostly an excuse to take a look at TensorFlow and regain some of my machine learning knowledge from undergrad :) I'd be curious to hear about anyone else mucking around with machine learning for fantasy sports!