My gripe with the "plane on a treadmill" Mythbusters episode was that testing it goes against the spirit of the problem, their awful grasp of science aside.
It's one of those theoretical physics questions that relies on a lot of assumptions that you can't actually re-create in real life.
When you're answering a physics exam paper there are always certain rules laid out before a question. Ice that is completely frictionless. The force of gravity is always exactly 9.8 ms-2. Air resistance is negligible. Energy lost as thermal energy is negligible. Friction is a constant force. Acceleration is constant and does not fluctuate.
There are a million and one of them and anyone with any qualifications in physics will be well aware of them. And in the case of the plane on the moving runway it relies on one of these assumptions, in this case a conveyor belt that can instantly and precisely match the aircrafts speed. We're not taking into account the effects of friction, wind resistance, pretty much anything. The only fact we are given is that the treadmill instantly matches the aircrafts speed. Of course, I'm sure that since it was released into the wild the question has now become bastardised and shortened, but in its original form that is how it would have been phrased.
Of course in real life we can't manufacture something like that, it is simply impossible, and this leads to two answers. In real life the aircraft will take off, the aircraft will be able to accelerate quicker than the conveyor belt can in the opposite direction. Obviously the net result is a take off.
But on paper it can't, due to these little physics assumptions that we make, and the correct answer is obviously no, it wouldn't take off. This is why it's against the spirit of the question trying to test it. It is a question designed to hone your analytical skills. Carrying out experiments is easy, but picturing obtuse and often illogical concepts and actions in your head can prove exceedingly difficult, and that is exactly what this question is designed to test.