My guess is that less advance is possible with 85 before knock detection backs off the timing as needed? I don't know exactly the inner workings of the PCM'* logic, I just try to figure it out based on my observations.
The MAF'* findings are an aspect of the PCM'* decisions on fueling, I get that, but engine loading and timing advance etc. are variable outside of just MAF readings. I think of it the same as a lot of performance engines today that advertise X horsepower on premium but they run fine on lower grades of fuel, just with less power.
My timing advance observations were at the same altitude, car, speed, RPM, time of day, and almost the same temperature with the difference being the fuel. Not just 91 then 85 but 91 then 85 then 91 again then 85 again. I couldn't think of a better way to confirm it (except to cycle a few more times but it snowed about then, changing speed, time of day, and temperature).