I did not miss the point. You called it a test of X but you actually did test Y.

However, its still valuable and interesting. You just should have done the test back to back with a thin pinion. And Roelofs point is very valid as well. Rcs are very small, so any difference any change, all details matter.

Though Junkie your test was interesting. It isolated just the gear friction aspect of power loss. So you cant call it a thin pinion test, because even an aluminum thin pinion is signifigantly lighter. Also as Roelof pointed out metal on metal is the worst case scenario for that. But you're just 1 test shy of having a full data set, maybe 2 tests if you can test a plastic spur vs metal pinon. Or a thin aluminum pinion and thin plastic pinion.

So you have your original test, the pinion moved for half engagement, and then a true thin plastic pinion for the best case scenario. But with a 7 KV difference between the first tests, it really is a small difference.