Quantcast
View Post
zygote said:

I like the idea of searching for a better algorithm using more sets of data.  I think we should each come up with an algorithm based on these numbers and find one that we think will work, then the closest one to the results can be tested going forward to see how it predicts.  I'll try to come up with one that I think will be closer to the result.  Will update when I do.

That's the spirit ;)

I will wait for march NPD results before I start twisting the data but I like your idea.