Hey all -- in case any of you are interested, I've been charting the output from the script when it re-calculates the rankings (now doing it every 3 hours). Currently, it's still using all comments from all time. I wanted to get a little more data before tweaking that, so I have a baseline to compare to.
Each column is a different execution, ranked 1 (top) to 25 (bottom), in descending order of ranking. They are not completely even intervals (the first two are a day apart, the next few are about 12 hrs apart, and the last 6 are in a 36 hr timeperiod). The color coding is:
- Bright Yellow: Song is new to the top 25
- Bright Orange: Song has appeared in the recorded history but just re-appeared
- Black: Song drops off the top 25 after this execution
- Light Green: Song has risen in ranking
- Light Red: Song has fallen in ranking
- Dark Green: Song has risen in ranking, but drops off after this execution
- Dark Red: Song has fallen in ranking, and drops off after this execution
- Brown: Song was new to the top 25, but drops off after this execution
Things of note:
Lots of yellow and black, which means that there is a fair amount of rotation in content. I can't say that the yellow songs were introduced by new-to-the-site users (I honestly don't know -- it would take a lot of time to analyze all those songs individually), but it's good to at least see new content appear there. The songs dropping off are from as high up as #6 and are actually weighted most heavily around the center.
The rankings in the first-25 range from ~1.5 to ~10, with the majority being towards the lower end of the spectrum (which means getting into the first-25 is perfectly accessible if you post regularly on other people's threads).
Next week (probably starting this weekend), I will adjust the script to only use posts made within the last 30 days and see how that affects the distribution of rankings.