User talk:Sean.hoyland

Khirbet 'Ein Karzaliyah (Arabic: خربة عين كرزلية), Jordan Valley: December 2013 - January 2014

Id'eis (Arabic: ادعيس), Jordan Valley: May 2014

This user talk page might be watched by friendly talk page stalkers which means that someone other than me might reply to your query. Their input is welcome and their help with messages that I cannot reply to quickly is appreciated.

Archives

This page has archives. Sections older than 5 days may be automatically archived by .

Graph of edits by socks

Is it possible to compare that with total edits to the same pages over the same time periods in order to determine what portion of edits are by socks? Levivich (talk) 19:38, 22 September 2024 (UTC)[reply]

Yes, it's possible, but I've been trying to avoid getting hate mail from the wikimedia cloud people. It's on my to-do list. I can probably do it in bite sized chunks that aren't too annoying for the servers. Sean.hoyland (talk) 06:59, 23 September 2024 (UTC)[reply]

I should add a couple of things

Presenting the ban evading data separately was a deliberate choice because, for me,
- the absolute numbers matter regardless of the relative numbers - just one ban evading actor on a talk page or edit warring is often enough to start a fire as they have nothing to lose by being blocked
- the total edit counts will obviously include lots of unblocked hard working ban evading actors
I'm pulling data at a per actor per month resolution and staging it in a local DB so there's quite a lot of data.
The ban evasion data used for the plot (including sockmaster info from the woefully incomplete sock category graph) is available here. Sean.hoyland (talk) 07:25, 23 September 2024 (UTC)[reply]

Thank you for putting that together! Another question: Is it easy to run the analysis for a small subset? I wonder what's it look like just for one article, or "top" articles, like Israel-Hamas war, Israel, Israeli-Palestinian conflict? BTW, I agree with you that absolute numbers matter, and also that edit count is a pretty poor metric anyway in terms of influence or disruption. (One well-placed revert or RFC vote can influence NPOV way more than 1,000 typo fixes, as I'm sure you already well realize.) But this is interesting data anyway, at least to check my own assumptions about how "widespread" socking is. (Less than I thought!) Levivich (talk) 17:08, 24 September 2024 (UTC)[reply]

Yes, it should be easy to run it at the article level. I'll have a look. It's hard to say how widespread socking is. Socks tend to only make a relatively small percentage of their edits in the topic area compared to outside. Whatever it is, it's presumably much less than the honest folk, I would hope anyway. It's not encouraging when you read the literature and see ban evasion detection rates in other systems as low as 10%. Sean.hoyland (talk) 17:44, 24 September 2024 (UTC)[reply]

I think it was Georgia Tech that noticed that ban evading accounts are statistically much less likely to swear than normal editors. So, it might be worth suspending the WP:NPA policy in PIA temporarily to help identify all the suspiciously polite ban evaders. Sean.hoyland (talk) 17:53, 24 September 2024 (UTC)[reply]

Great idea. Maybe we could trial run it for a day or a week or something. At the very least, it would be cathartic for all of us. Levivich (talk) 17:58, 24 September 2024 (UTC)[reply]

@Sean.hoyland can you cite me the research that socks are less likely to swear than non-socks? VR _{(Please ping on reply)} 19:09, 25 September 2024 (UTC)[reply]

Vice regent, I think it was https://doi.org/10.1145/3485447.3512133 from a couple of years ago. Sean.hoyland (talk) 04:07, 26 September 2024 (UTC)[reply]