Sort:  

Yep. I would say that most of the accounts generated in the 2017/18 run were alts - which means that the retention numbers are nowhere near as bad as the numbers tell. A lot of the accounts these days are created to play games also.

Here's the thing, whether you assume most of the new users are alts or not - the flatness of the curve implies that even if most new active accounts are not new users - because the active author and curator numbers are not rising, either way most actual new users are not sticking around (or they are replacing older active users that are leaving).

or those that are leaving are alts that are falling into disuse for a number of reasons, including being caught for something like spam or plagiarism.

The chart shows daily active users. So such abandonments would have to be very recent (ie we'd be losing the same number of such accounts each day as we are getting new active users). Given no particular change, you'd either expect most such accounts to have left some time ago, or be cycling if alts/farming/plagiarism is still viable(approx self replacing).

To me it looks like "all active" has a 30 day trend from 16,000 up to 18,000 - with that 200 difference being through DApps - am I missing something?

Well over 6000 new accounts became active in that period, but daily active accounts only increased by 2000 (or ~1600 just looking at weekly peaks).

I'd say it is partly due to the power down period from the fork too. The next couple months will be interesting to watch.