Using a model developed with Wayne Winston, I post the NFL bets I would make against the spread each week. My model uses scores from previous weeks to predict the results of the current week of games. Here are my 2014, 2015, and 2016 betting results, where I am a combined 310-307 against the spread, including a terrible 2016-17 campaign that went 105-133 in an unpredictable year.

Now, it is week 2, so I only have one week of results with which to build my model. And it’s based on scores alone. So it only knows the scores of week 1, and, for example, it does not know that New England is typically a really good team. Nonetheless, the model has been profitable when run over the last 30+ years, even in week 2. Given that, here are the predictions for week 2: