Is there no correlation between points and number of comments? That would be interesting to have some analysis on this, drilling further down and raking the quality of the comments. Or maybe this is too much? ;)
The question I would like to see investigated is the hypothesis that the best posts are those with a large value of (#points)/(1+#comments). I've used that for years as my measure of when an article that would otherwise have been missed should be read.
Yeah, that's also another heuristic that works (and one I use myself when browsing HackerNews in real-time). The only reason I required multiple submissions was to filter out posts that were of fleeting interest (product launches, or news articles that didn't manage to sustain HN's interest over a long period of time).
That would be great to do! Unfortunately I'm non-technical and tried to limit the scope of the analysis to the absolute bare-bones, haha. If you'd like to do it yourself, I'd be fascinated by the results :)