Comments
There's unfortunately not much to read here yet...
Best of Machine Learning
Discover the best guides, books, papers and news in Machine Learning, once per week.
Follow the full discussion on Reddit.
I have a stacking ensemble to predict probability of the total going under and a separate model that’s practically the same for the over. Right now on the under my test set size .2 which is 1400 game logs of my dataset and in order for my model to make a pick it has to meet a certain threshold. On the test set it’s precision on games that met the threshold (predicted prob > threshold) is .56, but the issue is it only made 132 picks. Is this enough to even consider using? I also have the average implied probability of the odds for the correct picks which was .5258 so roughly -111. I feel like 132 picks in 1400 games for just under is definitely enough in terms for the actual application too sports-betting, but I’m not sure if it’s enough to consider the results reliable. Essentially, my question is do I need to make the test set bigger or make my dataset itself bigger?
There's unfortunately not much to read here yet...
Ever having issues keeping up with everything that's going on in Machine Learning? That's where we help. We're sending out a weekly digest, highlighting the Best of Machine Learning.
Best of Machine Learning
Discover the best guides, books, papers and news in Machine Learning, once per week.
Join 800 others in curating the best parts of Machine Learning!
Just one more step. We use social logins to make sure that real humans are curating the list!
or contribute on GitHub directly.