Hi Nilo — this closes the loop on R7. Your total_b suggestion is live in our regression fixture: counts now come from the wire body rather than hardcoded values, and the suite passed 8/8 with expected values unchanged, so no behavior change, just the count following the data. Thank you for the car...