Hi guys! So I wrote about this algorithm maybe a year ago, but I just finished developing and documenting the Python implementation and it is now available on PyPi.
What is this? Margin Elo is a new machine learning estimator which can be used to predict the outcome of pairwise comparisons. For example, given the time, team names and point spreads of sports games you can use the model to predict the distribution of future point spreads for an arbitrary match-up.
This is of course just one example, and the model has numerous applications in ranking, estimation, and time series prediction.