The following predictions are based on the ELO ranking system. This system was first used in chess to calculate the skill level of players and predict the most likely winner. It’s a good system, but soccer is not chess because soccer can end in a draw. I’m using a modified formula that tries to predict the chance that a team will come away with points. This includes not only wins, but also draws.
The system starts each new team at a score of 1500 and then adjusts their rating as the season goes on. My current team scores include data from 1996-2016 (with some adjustment to account for the changing MLS win/draw rules). The following are not included: Domestic tournaments, preseason friendlies, and international games. These scores are used to rank teams for my Power Rankings table and the values are now included in the chart.
Interesting stuff! Will you provide verification statistics as the season goes on? Have you used it in the past and have verf stats for the past?
I added a tracker to my table for this season that will, with hope, help me track the accuracy as the season progresses. I’d have to go back and manually make the calculations for the past three years though.