Came across something intriguing called Grey System Theory - anyone here heard of it? Apparently even with tiny data sets (as few as 4-6 data points), it's capable of outperforming many bread-and-butter time series algorithms with 10-100x larger data sets.

You can read a few papers about it at Scholar Google

Here's the paper introducing the theory.

I'm baffled that something so promising could be so obscure (there's no article about it @ Wikipedia). Did the articles oversell Grey System?

I haven't had a chance to read the introductory article yet, and only skimmed through the other articles to get a gist. I'm not really qualified to judge the merit of a predictive algorithm even if I were to read through everything anyway. Hopefully somebody is capable of enlightening me as well as the future interested parties.