Reply to thread

Message

[QUOTE="Orangezoo, post: 884829, member: 237"]
Statistical models are far from an exact science. Considering in basketball there are very few constants every game efficiency models are a fun way to compare ratings but often are over valued due to the fact they correlate well with something as simple as wins and losses. If you were to take the top 100 teams in terms of something as simple as their record you could fairly easily predict the correct 80 pct who will go to the tournament. Maybe even 90 pct. For this shear fact you are almost guaranteed to find that efficiency ratings are a decent predictor. Wins and efficiency are closely related.

In many ways its like saying the humidity levels are high so that means it is going to rain. You can have some accuracy of predicting the weather just based on that but its not some fancy mathematical way of analyzing it. It is a hobbyist thing and nothing more.

The reason that efficiency models or any other models are not the sole way to analyze teams is due to matchups , type of defense etc etc. Two good def 3pt fg defenses could play polar opposite defenses. Those things matter when you look at matchups.

I am a much bigger believer in the eye test than efficiency ratings or models. And I have a degree in mathematics and deal with modeling and forecasting every day at work. Where efficiency and predictors help are in bubble team discussions and seeding weaker conference teams where the eye test against good competition is limited. Its a secondary tool and nothing more.
[/QUOTE]

Verification

What is a Syracuse fan's favorite color?