As if this wasn't confusing enough...
Piston speed isn't constant, right? The piston stops at top and bottom dead centers, therefore accelerates and decelerate in between, or every 180 degrees of crankshat rotation. So using 'average piston speed' is kinda like using average speed while calculating fuel economy: Great if you only vary a few mph, but not if your commute invovles stop-n-go traffic.
The variable of rod ratio governs max piston speed for a given crankshaft rotational position, so wouldn't average speed be the same for every rod ratio if the same rpm is used?
Acceleration and deceleration rates, dwell time where the piston is at it's slowest, burn rate and bore size. Race engine builders ofter get to a place where the piston outruns the flamefront. Phew... so much going on I get dizzy. But wait, there's more, there's cylinder filling efficiency... what is optimal for the power cycle may hurt the intake or two other cycles to the point you have 1+1 steps forward, then 2 steps backward.
I think if you fix certain parameters (bore, stroke, port and head design) every engine has it's own 'sweet spot' where it is most fuel efficient, with fifteen (a number taken out of thin air) variables that influence it. I think you can make tweaks and changes that improve on FE, but question weather simply decreasing the final drive ratio until the engine no longer accelerates is the magical answer.
On the high performance side of things, I spent hours playing with "Engine Analyser" software varying only camshaft and exhaust headers selection. Then I put the thing on the dyno
I concluded the whole thing is one giant compromise.