Post by BeachbytheBay on Feb 12, 2024 2:39:45 GMT -5
for INstance Long Beach played and beat King twice. Massey only shows the one match. It almost makes me think there is some other entry that gets fouled up on the import of schedules before matches are played, cause that one would not seem to be a time schedule quirk
I really don't think the issue (double counting of a match, or single count of a two matches with the same results) creates all that significant of an error, simply because the ordering that results from beating (or losing) to a team doesn't alter the rating % impact of a result all that much.
If the issue was incorporating incorrect results from matches, then yeah the impact would be significant.
true though that it's not precise until corrected, and the omission or double counts of matches until they are corrected is annoying. I'm just glad there's well established algorithm provided aside from RPI. Massey ratings cover a lot of sports. And he actually has two rating indexes. His standard, and then his Power rating, where I believe the Power rating incorporates sets and/or points as well as the simple rating based on W-L. Match predictions are based on the Power and not the standard
I guess I don't consider it actually garbage. the reason is double results or leaving out a single result is not introducing incorrect match results, if that makes sense. Garbage to me would be getting the actual match scores wrong. Trying to think of a better word to use, I think incomplete is more appropriate. And yes, until complete, there's a an error to the ratings or predictions, which I doubt is all that significant the more a season goes along and more data is completed. Every rating is pretty wild and somewhat 'garbagy' towards the start of a season until enough matches have been played and enough data is in the rating system.
