Lots of testing was done over the years on a Triumph Bonneville using various fuels. It was done on the dyno using as scientific a procedure as possible considering the time and money available. The jetting was set to provide the richest mixture that made good power. The timing was set to make the most torque at the least advance. The fuel air curves were measured using the dyno operator's lambda sensor stuck up the tailpipe. Often there was another lambda sensor attached to a bung in the header pipe.
It was important to use multiple sensors. The header bung sensor was always problematic and comparison with the tailpipe sensor showed this. It would have been a disaster if the header sensor readings were used to adjust the fuel/air mixture in some cases.
All sorts of gasolines were compared, a heavily oxygenated ERC race gas developed for the Australian market, Sonoco Standard, Sunoco Supreme, a highly oxygenated Sunoco leaded gas, and Sunoco GT260 Plus which is used now. Some Sunoco Supreme-toluene-nitropropane mixes were also tried.
The last mixture showed no advantage on the dyno but it launched the bike to some insane speeds in actual use. This is something I cannot explain. Otherwise, none of the fuels gave any significant advantage after the mixture and timing were optimized for each one. This is something else I cannot explain. Many people see performance differences with varying fuels.
The latest version of the Dynomation computer based engine tuning program can look at the use of different fuels. One project for next winter is to model the performance using the various fuels I tested and to see if the virtual results match what we saw on the dyno. It will be a lot of work but the results should be interesting.