Can you teach me how to do this stuff?
It covers Python, Web Scraping, Databases, SQL, Machine Learning and more.
Very. It's built using probabilistic and Bayesian modeling techniques, which fit thousands of curves to past data via
Markov chain Monte Carlo
techniques. The end result is something like the gray (the model) vs the blue (actual performance) below:
More on interpreting these plots
here
.
This is awesome. How do I run it all week?
Go to
fantasymath.com
and sign up there. This Monday site and the above correlations are just one game and 1k simulations. The full Fantasy Math model does EVERY game, up to 10k simulations each.
The bottom line is it lets you take into account player variability AND in-matchup correlations so you always start the guy who maximizes your probability of winning.