rogers park bryan wrote:
Boilermaker Rick wrote:
http://regressing.deadspin.com/chart-how-sabermetric-mlb-forecasts-stacked-up-against-1640931998/allLook how absolutely terrible about half the predictions are.
From what I can tell, it looks like 3 picks were spot on with a few others being pretty close.
Looks like the Saber side is a lot closer than Vegas overall.
Both were pretty bad though, but Vegas is not trying to predict actual results.
Quote:
While Fan Graphs and Davenport appeared to outperform its competitors and boasted slightly lower average errors (MSE, MAE) than sportsbooks, betting each team according to both Fan Graph’s and Davenport predicted totals would have finished just 13-17. In general, the number predicted by sportsbooks finished closer to the eventual win totals.
In summary, it’s pretty amazing that not a single one of the nine prediction methods I used was able to finish with an overall record above 0.500.
Lastly, I plotted the predicted and observed win totals, using both sportsbook predictions (abbreviated as “Las Vegas Prediction”) and the ensemble method (the average of the first four sabermetrics/statistics sites above, listed as “Steadheads Prediction.”)
This just goes to show you that the most correct answer does not mean the right answer.
Login
Register
FAQ
Search
