Warning Signs in Experimental Design and Interpretation
norvig.com
norvig.com
http://norvig.com/experiment-design.html
is good stuff, very helpful for evaluating hundreds of other submissions to Hacker News, and it's written by a LISP hacker (now director of research at Google), so what's not to like?
Useful info (assuming one's utility function relates to acquiring more accurate beliefs), I had hoped other hackers might enjoy. Pity, as it seems to have missed the main page.
Many of us have probably made one or two similar arguments against "new paper proves xyz" overstated claims. But a large collection of these red flags, with historical examples to boot, in one document, is simply fantastic.
Thanks for sharing