An interesting point no doubt.

While I considered that it could be the downwash on the stab I'm starting to wonder. At lower speeds the wing is operating at a higher lift coefficient. Hence the downwash angle is steeper and will hit the top of the stab more strongly so the stab appears to be operating at a more negative angle of attack. As the speed rises the lift coefficient for level flight lowers and the wing moves to a lower angle of attack. This does two things as I see it. One is that the pitching moment has more airspeed to work with so the pitching torque increases. More nosedown.... At the same time the downwash off the trailing edge of the wing isn't angled down as much since the Cl is lower so the stabilizer appears to be operating at a lesser negative angle of attack. So less tail down force at the stabilizer to counter the nose down torque of the airfoil's pitching moment. All of this will sort of balance up to a point where the pitching moment increasing and the stab download decreasing suddenly crosses the line and it's LAWNDART TIME! ! ! ! Unless the pilot steps in with some up elevator to pry the nose up and slow the model down to below the critical speed range where the balancing act runs afoul.