Seems to me Gates belts have a center "groove" that keeps them in alignment and the pulleys are machined accordingly ... So we have a toothed belt and a self aligning system. Yeah, I suppose you could throw if off somehow. But if you are QR, I wonder how that would happen? Stuff comes apart and goes back the same. Unless you lost a spacer or something ...
My Harley made somewhere around 75 HP and would occasionally pull the front wheel with a stock belt drive. I had the rear wheel on.off quite a few times w/o issue. OTOH, my bud was always complaining that his Sporty rode strange .. So I took it for a ride - you, wandering, feeling skatey, etc.
So we decided to check the rear alignment. We used an old 4' fluorescent tube (they roll and are straight) from front wheel to rear. Yup off. So I grabbed his belt to see what it was doing? Rock hard, no give. Went to my bike, 1/2" to 3/4" deflection with my thumb and middling pressure. Went back to his and backed it off to about the same. 90% of the evil handling went away. The moral here is cogged belts do not need to be nearly as tight as say a V belt. They can live nicely with just enough tension to stop them from "climbing" out of the cogs in the pulley.
And with only human HP input, they could prolly be lightly tensioned by our current standard. I gotta see what Gates actually says about minimum tension ... Looks like 35# is the minimum tension for any of the road bike types. I don't think it would be too hard to get to 35# ... And even if I could not do that roadside, I'll bet I could get some tension into it and ride home w/o standing into the pedals.
Automotive belts run high tension, but they are either V belts or flat belts with tiny vee's. They depend on surface friction. Bicycle belts I have seen are more like chain drives than other types of belt drives
Here's the Gates Tech Manual:
Gates Tech Resources