Interesting, CDR. I did a Z2 ride tonight and used only 53x19, 17, & 16, with cool down in the 19, and averaged 202w for 1hr 4min @ 89 avg cadence, 18.9mph avg, and never breaking a hub-indicated 23mph. It was a steady state ride, and I don't think I shifted more than 4 times throughout, so I'd expect the consistent power results to be pretty accurate.
Trainer Profile | Hot Z2 Action! | Times and Records | Strava
That's a striking difference from your results. My F2 is many years old, maybe 8, but I don't remember. Could the fluid properties change over the years? How else to account for that variance?