Wish he had posted log from the last video he put up to compare. I always kind of wondered if it was the jarring motions of a clutch car launching jiggling the blade and causing it to see errors in predicted TPS vs Actual. A 1.5260 foot in a stick car is much more violent than even say a 1.36 60 ft in an auto car. If I was him I would have tried to get hands on a couple low mileage stock TBs and maybe a new GM one, maybe internal spring or something is healthier.
I have customer than never had problems start to develop them, and got another LS2 but off a low mileage car and its back to normal?