Least Squares

With the MLB Playoffs beginning this evening, I figured it was time to test my rankings and pull out the old probability calculator. I created the MLB Ratings based on a simple least squares NLP Optimization that I have discussed before. Oakland at Kansas City The Royals are in the playoffs for the first time in ages and they get to host a game. Unfortunately, they didn’t seem to have a home field advantage during the regular season, so I am not sure how much this helps (although in reality we can assume it does, at least a little). The numbers say the A’s are the better team by almost 0.7 of a run (per game, for the season). I show them as a 63.5% favorite. ...

The first weekend of the NCAA Tournament was a wild one. In our competition, we chose models with high variance, knowing full well we could be in a world of hurt if a game or two did not go our way. Being scored on a log loss scale was new to us, and we knew of the risks, but did not really think things could get too bad. ...

With March Madness upon us, I have been in a solid state of sleep-deprivation. It all started with a class project assigned in late February that suggested we enter the Kaggle competition of our choice or create a similar type project. I was immediately drawn to the March Machine Learning Mania being hosted by Kaggle and Intel. For the past three weeks, in any spare time, I have been trying to find and clean data to run models. I thought things were slowing down last week until I decided to try some new data I had found. ...

With the first weekend of the NFL Playoffs completed, it seemed like a good time to catch up on how well the Least Squares Optimization predictions did the past two weeks. If this is your first time reading about this, please refer to my initial article here. First, let’s recap the final week of the regular season. Using only the games where the percentage difference between the expected line and the actual lines (from by sportsbook.com when published) was greater than 100%, the predictor went 3-1. In games where the raw absolute value of the expected and actual lines was greater than 2.5, the predictor went 5-3. Overall the predictor went 12-3 (the Bears-Packers game did not have a line sure to the unsure status of Aaron Rodgers. ...

If you have not read any of my previous Least Squared posting, please refer to the initial post here. Week 16 in review: Only one game was in the range where we have been fairly confident on the selections. New England was a favorite in the system, but getting 2.5 points from Vegas. New England destroyed the Ravens, so the >100% difference between the expected and Vegas lines, bring that rule to 12-3 over the past four weeks. ...

There’s no shortage of data out there when it comes to college football, so I decided to take the time to create a least squares model, based on the same principals I have been using for the NFL, and outlined here. The idea was to take all 752 schools that played college football, cross that with all 4138 games that were played (up to last weekend) and see how they predict the games (especially versus the Vegas lines). ...

Sometimes you just need to slow down and look at the data a little closer. That was the case last week when I mistakenly posted the wrong side of the New England-Miami line. I mentioned it was probably best to stay away from it all together due to the raw difference being so small, but I also stated the wrong side to take. Oh well. Lesson learned. Week 15 review – The least-squares method and choosing only those lines where the percentage difference of expected and actual Vegas line over the actual line was greater than 100% went a shocking 4-0. Did I bet it this way? Nope. The three lines between 80-100% went 1-2. In games where the absolute raw was greater than 2.5 went 3-2. Overall, the LS method went an incredible 11-4-1. ...