You can use the API to find some of these answers.
I have to assume the karma ends up falling into a pareto distribution, so averages will be heavily skewed and kind of useless. Something like quantiles may be more useful.
Until you know the distribution, you don't know whether an average is informative or misleading.
My impression is that there are a lot of high-karma accounts that post quite infrequently. I think a measue should take that into account.
giantg2•6mo ago
PaulHoule•6mo ago
Any sample of users who actually post is going to have higher karma than the universe of HN users, many of whom haven't logged in for years.
Somewhere between the time I had karma 35,000 and 60,000 I had a script that I ran sort-of-daily which captured the top 100 users including the ones who don't have a number posted on that list. I was planning to make a chart of how top user's karma was growing over time but I guess I was never competitive enough to actually make the chart.
EPendragon•6mo ago