This is really great data, esp considering our baseline for 1 ns0 edit within 24 hours is ~25%.  This type of data helps push our understanding of how users come into the system.

When Dario gets back, we should take a look at these anon account creators to see what their  UMAPI stats are -- I'm expecting their edits in 24 hours would be very high, like maybe 90%.


On Fri, Jun 21, 2013 at 1:54 PM, Steven Walling <> wrote:

On Tue, Jun 18, 2013 at 6:51 PM, Steven Walling <> wrote:
Among our first test cases, my team would really like to know if significant numbers of people are signing up via the links English Wikipedia puts in system messages directed at those viewing semi-protected pages and anonymous editors. I've put up a proposal about this on-wiki at

These two campaigns have been enabled for a little over two days now. For the curious, I've gone ahead and shared preliminary results at

Obvious caveat here is that a week's worth of data would be most reliable, since it accounts for normal behavioral changes in a working week. But I don't expect there to be a huge change in the proportions overall. 

EE mailing list

Howie Fung
Director of Product Development
Wikimedia Foundation