It is anecdotal insofar as the anecdotes are from my limited cohort ( few dozen subredditors who have signed up with weightgreat.com ). Also, you are right abo

It is anecdotal insofar as the anecdotes are from my limited cohort ( few dozen subredditors who have signed up with weightgreat.com ). Also, you are right about the multicollinearity issues ( supper time predictor entangled with other predictors). But the overall effect is pretty significant. There's some preliminary material here on the models used. 1. https://news.ycombinator.com/item?id=13630203 https://news.ycombinator.com/item?id=13630203 2. http://weightgreat.com/forecast.html http://weightgreat.com/forecast.html