Reply to thread

Message: [QUOTE="Orangezoo, post: 889054, member: 237"] Interesting model. Basically runs into the same issue discussed in the ennis thread this weekend in that there are still too many things at play to consider it viable. Its better used to gauge who has been tested vs who hasn't in hindsight than to predict anything. Fun analysis though. With that said if you combine circumstantial factors it could be used to assist with who might win one game between two highly rated opponents. Multiple games in a row the fact its a small sample size will reduce its effectiveness as a predictor. [/QUOTE]