Anyways, as a totally unscientific study, I took out my aluminium work boat (18' Thunderjet with an Evinrude 135 HO) and did a quick comparison. Here are the numbers:
TRIM "UP":
- no gauge, but likely ~3/4 of the way "up", relative to the max the engine will let me while under speed (it will not let you start "trailering" the motor at speed).
-Exactly 5000 rpm - close to WOT, but not quite (it will do just over 5500 flat out, at which point I trim fully up as well)
-GPS measured 69 km/h (43 mph) fairly consistently
TRIM "DOWN":
-without touching the throttle or steering wheel, I just held down the trim switch until the engine bottomed out.
-Just under 4500 RPM - at the same throttle!
-GPS measured 58 km/h (36 mph) fairly consistently.
Fuel consumption is a bit trickier as I don't have a guage to measure that, but bear with me - the most the fuel flow can change is the percentage rpm change, as the throttle was the same - the least it can change is not at all, since the throttle was the same. If it didn't change at all than fuel economy is directly related to speed changes - in this case 17%. If it is related to RPM, than the 10% decrease in RPM somewhat counters the 17% loss in speed, for an overall change of 7%
SUMMARY:
By trimming up, my engine gains a full 500 rpm - good for a 17% (!!!) increase in speed, and a 7-17% increase in fuel economy / range. YMMV, no pun intended

Denon