Ok, I gotcha. If the whole setup is moving back and forth a lot during the ride then there is definitely energy being expended flexing the bands -- similar to working against the suspension of a mountain bike.
As far as PT accuracy goes, I think you're asking whether the PT is still truly capturing your physical output correctly. IOW, whether the extra energy expenditure into the rubber bands is being counted by the PT or bypassing it? If that's truly the question, I think the answer depends on how you're riding, and is probably a combination of the two. Roughness in the pedal stroke itself which causes surges in wheel speed is probably handled by the PT. Bouncing due to movement of your body (if any) could represent effort that is not being captured by the PT. For example, if you were to sit on the bike without pedalling and just rock the bike seat forward and backward by contracting the abs/hips then there would be some work there that's not transmitting through the PT. My gut feel is that any rocking due to body motion is probably the smaller effect during seated motion (much larger during sprinting or pushing out of the saddle), but that's just a guess.