Please respond to the following questions, and provide as much detail as possible for each.
- Problem: The work done in T391964: [Epic] Recent Changes ORES Enabled Revert Risk Powered Filters Rollout Plan is beneficial to the 20 wikis we've calculated RevertRisk "high risk" thresholds for, and it would be great to see this expanded to all wikis (or at least all Wikipedia), excluding Commons and Wikidata due to scaling concerns.
- [Optional] Possible solutions:
- Re-running the threshold analysis in T391964 for all wikis (or all Wikipedias) excluding Commons and Wikidata, and updating InitialiaseSettings.php with the "min" score for the "high risk" threshold.
- nice-to-have: also calculating the "low risk" threshold identified in the research, so that it's possbile for software to interact with a risk level for an edit that is considered "low risk"
- Enabled projects:
- T391964: [Epic] Recent Changes ORES Enabled Revert Risk Powered Filters Rollout Plan in that it can be more widely used, especially on our highest traffic and impact wikis
- T364705: Provide AbuseFilter condition for revertrisk threshold, in allowing the "high risk" (and ideally also the "low risk") thresholds to be used in AbuseFilter filters when deciding how to mitigate incoming edits. This feature would likely be especially useful in combination with other AbuseFilter variables like T354599: [EPIC] WE4.2.14b Provide IP reputation variables in AbuseFilter
- Urgency and importance:
- The relevant features are already launched; this is about scaling the work done in T391964: [Epic] Recent Changes ORES Enabled Revert Risk Powered Filters Rollout Plan to have a wider impact
- [Optional] Notes: Is there anything else you'd like to share?
- Thanks for considering!