The 'issue' with statistical analyses is that the results generally leads to more questions. You may like to look into several weeks instead of 1 week to see if the same users end up high int he curation game. I also think you have to look into the auto voters, since I expect same accounts to vote for same authors over and over again, and these may come high on the curation charts. Maybe other Steemians analysed auto voting of high SP holders and auto vote for the same authors just prior to these. Since HF19 I have a feel the number of high SP holder auto voting increased over manual voting. I never was good in curation, and generally vote for those posts I like, at the time I read the post, regardless of the 15/30 minutes thing. But writing this comment, I'm getting interested to see if I can define some interesting vote rules to increase curation rewards. I may look into this one day.
Thanks fro the reply @edje
Yes i agree a longer time frame would be better - i'll need the data from steemsql to do that though.
Manual curation will generally do better than auto voting - but you can do ok with auto votes for no effort at all once it's set up.
I've added you to this weeks Curation data :)
Ok, interesting! Am a bad curator, I vote for what I like when I see/read it, so I also give most of my curation away when voting for posts or comments that are just a couple of minutes in the system (like your comment). Also I may vote on posts of days old. But am curious what the results will be. Thanks for including me.