Handicapper's Season Win Total Sabermetrics Model Cashes Again

Mark Lathrop

Wednesday, November 22, 2017 1:46 PM UTC

Wednesday, Nov. 22, 2017 1:46 PM UTC

Our expert MLB handicapper Mark Lathrop recaps another great season of results with his sabermetrics based win total futures model for the 2017 MLB season.

2017 MLB Record (with Futures): 242-209-12 (+17.46 units)

Another fantastic year is in the books for my sabermetrics based MLB win totals model here at SBR. While the 'Unders' have been cashing with regularity in past years, this year it was the 'Over' calls that performed significantly well. There was also a lesson learned about divisional parity that we will be using going forward.

To recall, my model takes the preseason projection of player WAR, or wins above replacement, and combines them into a roster score for each team. This is fed into an algorithm to produce a projected win total and then compared against available futures lines at Bovada. Over time, I have made adjustments to the model based on the optimism inherent in the projections – and no adjustments made for possible injuries. Also, the books have more wins in their futures lines than are available to be had by the league based on games played. The end result of the model is about a third of the games that are either 'Over' calls, 'Under' calls or passes.

There were 8 'Over' calls in the model this year, and 7 of them cashed. One of them was the lowest total available on the board in San Diego, which cleared its 66.5 game mark by 4.5 games. One of the stronger indicators was for the Dodgers, who were projected to beat their totals line by more than 7 games and ended up beating it by over 10. An 88% rate for the 'Overs' covered up some deficiencies on the 'Under' side, with only 7 of 13 of those picks cashing. It’s still a winning record, but brought the overall winning percentage of the model down to 67% at 14-7 ATS. That is the 3rd year in a row that this system has produced results better than 65%.

This leads us to discuss what went wrong in the 'Under' section of the model, and luckily for us, there is a lesson to use in future years. It has to do with the AL Central, and the terrible seasons for the Chicago White Sox, and surprisingly, the Detroit Tigers. The Tigers were a winning 'Under' call for me on the year, but they fell short of their projected win total by nearly 20 games! Those wins had to go somewhere, and with the unbalanced division schedule some of them ended up with the Minnesota Twins, Kansas City Royals and Cleveland Indians, who all exceeded their win totals against the 'Under' calls in the model.

In the future, we must make adjustments for when nearly an entire division is sitting in the top third or bottom third of the results, as those teams' records are interrelated. I look forward to building the model for 2018, and bringing these lessons to bear as we continue to profit off of sabermetrics projections.

Bet On MLB Futures At SBR's Top-Rated Sportsbooks
comment here