Which MLB Teams Do Our Models Like Less Than Vegas Does?

Los Angeles Dodgers

Projected Wins By Our Models: 90

2018 Over/Under: 96.5

Our Dodgers forecast is illustrative of the conservative nature of projections. Most team projections are heavily regressed to the mean, and that means the projected totals tend to be more bunched than the actual results at the end of the year (the standard deviation of our team projections is 7.4 wins, while it was 11.5 for actual wins at the end of the 2017 regular season).

By our model, 90 wins is still good enough to tie for the fifth-best record in baseball for 2018, while it would have been ninth in reality at the end of last year. Granted, even knowing what we know about projections, 90 does seem a bit low for a team that had 102 Pythagorean wins and 101 BaseRun wins in 2017, but there are still some reasons to assume the arrow is pointing downwards.

Los Angeles allowed a team BABIP of .281 last season, tied for the second-lowest in MLB, and this is not something a model would reasonably expect to repeat itself. While Clayton Kershaw has a history of running super-low BABIPs, we should probably be less confident in Alex Wood (.267 in 2017), Rich Hill (.261) and Kenta Maeda (.278) doing the same for another season.

The Dodgers' pitching staff also lost Brandon Morrow, Tony Watson and Yu Darvish (2.9 combined fWAR) this offseason.

Also, while our model might still be a bit too low on LA, it’s worth noting that FanGraphs also has the Dodgers winning fewer than 96.5 games. PECOTA has them hitting the over by 0.5 wins.