Help with Statistics

RockinX

Yo, I'm attending a Statistics class and we're seeing Logistic Regression right now. However, when talking about -2LL (log likelihood ratio), the professor talked about it twice, and contradicted himself both times. First he said that the closer -2LL was to zero, the better the model served as a predictor, and the following time he said the opposite.

So which one of these statements is the true one?

Inquisitor77

http://www.statsoft.com/textbook/stathome.html

RockinX

I don't care that I haven't found what I'm looking for yet, but this is just too awesome.

THANK YOU!