Which MLB Teams Do Our Models Like More Than Vegas Does?
Last week, we took a look at what MLB teams our model thinks will underperform their Vegas over/under total.
Here, we’ll look at the flip side of that. Before going on, it is important to remember that we do not necessarily think these three teams will be good in 2018; we forecast all of them to be below .500.
However, our numbers think they’re all a decent bet to hit the “over,” because, as we covered in the first installment, projection models tend to be heavily regressed to the mean, with a smaller spread than the end-of-season outcomes.
Last year, for example, five teams lost more than 70 games, but our models do not see any squad losing 70 games this year, while FanGraphs projects there to be two such teams.
That is not to say there will not be any teams that hit these lows, but often times when a team is an outlier in this way, there is a bit of bad luck involved that a projection model would not forecast. For example, Detroit tied for the worst record in baseball last season (64-98), but had a Baseruns record of 69-93. Baseruns strip away sequencing to look at a team’s underlying level of performance, meaning the Tigers were five games worse than expected due to cluster luck.
Detroit was still bad, but they weren't as bad as their record suggested, which was true of all six teams that lost at least 70 games last year. This is instructive when looking at the following three teams, as like Vegas, we think each of them will be bad in 2018 to varying levels, but we differ as to the degree.
For clarity, we'll be using the over/under numbers from OddsShark.com.