Following up from this discussion, it would be desirable to display the percentage of a user's edits that are non-automated at the main edit count page, rather than only in a subpage. This is because it is an important data point that can help complete the portrait of a user's editing activity, and because having it more visible may reduce the incentive for editors to run up their edit count through making lots of low-value AWB edits.

The existing AutoEdits query is quite expensive, presenting a hurdle. The proposed solution to this is to only query for tagged edits, which might miss some niche tools or old edits that predated the tagging system, but will be much less expensive and should give a good enough metric in most cases. A small, non-prominent disclaimer could explain the discrepancy between the count and the count on the subpage for anyone curious.