I guess a big part of my thinking comes from my experience with rally cars.
I've watched plenty of people spend lots of time and money trying to save weight, and have it make zero difference to their stage times - and often make the car fragile and difficult to live with.
Personally, I owned a car that was 1160kgs, and then a rule change meant it had to be ballasted up to 1280kgs - and I swapped from a 70kg to a 110kg navigator at the same time. The ONLY time the extra weight was an issue was braking in the wet. Acceleration was not changed by any noticable amount - in particular, there was one regularly used stage start where I never knew whether or not I should change into 4th just before the first significant corner - I'd run out of 3rd gear in exactly the same spot, regardless of what the car weighed.
The heavier version of the car sat better on the road, over the rough stuff - basic sprung:unsprung weight stuff. In the case of a heavier rider on a dirt bike, most of the rider's weight is 'double sprung' (ie: the bike's suspension and the rider's arms & legs), so its gotta be even more helpful, right?
I think this is a big part of the "manhandling" issue - a big/fat rider has less need to manhandle the bike, because the rider's weight automatically dampens out a lot of the bike's nervousness.
The ability to move your weight around is a big influence to how much traction a bike gets on dirt. Obviously, more weight is harder to accelerate but I reckon that in lots of circumstances, if that extra weight is put over the rear wheel, the increase in traction cancels out the extra total mass.
Ah, shit, I'm sounding like I'm saying that being a fatty is a good thing...
Really, the point I'm trying to make is that extra rider weight doesn't kill off engine performance all that much, and may improve handling and traction in some circumstances.
Going back to the original post, even on a drag car, losing 2.6% of non-rotating mass has LESS effect than increasing HP by 2.6%. Every part that rotates has both a static mass and a rotational mass - and to accelerate the bike/car, both of those masses must be overcome.
Imagine you had a 200kg bike/rider, and a 200kg lump of steel and you had to push both of them to 10kph.
If they were both on ice, it would require the same amount of energy for both.
But once you let the bike's wheels rotate, you'd find that the bike requires quite a bit more energy to get to 10kph because you have to get the wheels/chain/engine internals rotating as well.
For the record: 90kgs, down from a peak of 98. Aiming for 85.
