That isn't surprising because the graph is in percent of fuel energy. That's why BSFC improves closer to full load, and why I prefer low RPM with open throttle.
Also, reciprocating loss is missing, but then the graphs don't deal in RPM.
this is from early 90's as the tittle "Automotive Fuel Economy: How Far Can We Go?" i have the pdf. i dont have a host for file sharing though. it has some good stuff in it.
Ok, I think I get it now. So if we burn 1 cup of gas in an engine at idle, it will run for quite a long time and it will turn so many rotations that the relatively small frictional losses per crank rotation will be additive and will be very significant.
Whereas if we run a cup of gas through an engine on a dyno at near full load, it will not run very long (fewer revolutions) so the frictional losses add up to less even though they are greater per revolution.
i have a bunch of statistics, CAFE stuff, general.... one could say boring. ill look for the studies and effects of etc etc, i know i have them in my back up hd. in the meantime i give CAFE 2008