This is true but the real question is what are these plates worth? Posi saw no gain. If they are only worth 5-6hp that is probably close to being within the range of error for a dyno. We are talking about 1-2% difference on a 500hp car.
If the deflection could be controlled then yes the welds would not fail but a thicker plate with possible ribbing would need to be used and also welding the plate somewhere midspan like the bypass valve housing
Edit: The welds will still be a structural component but will not have to deal with the inplane rotations caused by the oscillating plate. Think about breaking a wire or paper clip you bend it back and forth a couple of times and it snaps. That is what the plate is doing to the welds when it oscillates
I think we are actually on to some good ideas here - hopefully Steig is watching this thread, even if he doesn't want to participate actively. You are correct, though, it is really the "big question" as to the efficacy of these plates. I mean, I'm sure that Steigs thinking was too reduce turbulance, and make the air directed through the i/c, rather than swirl around in that big chamber, and hitting all of the baffles. What still remains unknown, though, is "how" the boosted air actually behaves at the outlet. Since air always seeks the path of least resistance (like water), it seems to me that it wouldn't spend enough time in that area for the turbulence to be a big factor - especially with the vacuum of the open intake valves drawing the charge as it is being pushed through the rotors. This would explain why most people say that they notice little, to no effect by removing it.