Originally Posted by
deklund
It makes sense that in the best case, the frame would return energy that can transfer to the crank, but I honestly can't see how that will happen. The frame is flexing laterally on the chain side down stroke and then I presume springing back as soon as the crank is at the bottom. My sense is that this will translate to heat in the frame (though not quiet the melting point of titanium) and is purely lost energy.
Do this test:
1. Lean a bike next to a wall with the handbar on the wall
2. hold the brake lever down to stop it from rolling forwards
3. step on the pedal near the crankarm and push it down and sideways, notice how the frame flexes 2-3cm sideways?
4. gently let up on the pedal, can you feel the frame pushing back against your feet?
When you're pedaling the bike, that frame unflexing pushes against your foot and spins the crank. Time after time, tests are done on this frame-flex phenomenon (it appears to be done every couple decades). The results over and over again indicate that frame-flex robs little or no energy from the total energy output. No one has been able to measure quantitatively and to a significant amount any speed differences between a flexy versus super-stiff frame.