On revertrisk score fetch in ORES extension:
- increment a counter for each revision score that is greater than the minimal threshold (0.50) for a revision to be considered as "likely to be reverted".
- If we plot these in Grafana, we'll be able to see spikes in attempts to flood a wiki with likely-to-be-reverted content.
On revert (In WikimediaEvents, most likely):
- increment a counter when an edit is reverted, and the "likely to be reverted" score in ORES was > 0.50 for the revision being reverted
- ... increment a counter if the "likely to be reverted" score was < 0.50.
That would help us visualize the accuracy of the results