Lookee see, it's Billy G!

Hey HE. Not to go into gory detail, but the algorithm for the DPPI is a “predictive” model. That means the obvious. It is intended to be primarily used to predict more accurately future outcomes. Some models are “retrodictive” meaning they are more designed to give credit for past performance with less emphasis on being predictive. For predictive models, the teams have to be meaningfully “connected” statistically. For this connection to be statistically “robust” would take every team playing ten games or so randomly distributed (I.e. not limited to conference opponents only). The way my algorithm it requires a minimum of three games for every team to run properly. That has to do with the fact that the algorithm calculates the average scoring margin of a teams opponents opponents opponents. So short answer long, when every team has played three games I will be up and running. I haven’t looked at the schedule for every team but if there are a handful of teams that take a while (say into December) to get three games, I will fabricate those handful of games for this few teams so I can get it up and running. Hopefully by next weekend.