I go along with the subjective results matching expectations camp, but it could be true. How about if someone (not me, I'm not messing with my mpg) did some objective tests with 87, 89, 91, etc. Do 0 - 60, 1/4 mile (if you have the facilities), 45 - 65, etc.

On the other hand if octane affected performance I would think that car mags (Motor Trend, Car & Driver) would do acceleration tests using different grades of gasoline and report on those differences too. The fact they don't may say something.