We were talking about going from 0 to 120rpm, on Pro sprint gearing that's from a standing start to about 36mph. Nobody can do that in 2s. If they could, they would need to make about 5 times as much power in our example, so chain friction loss would be nearly 5 times as high (nearly, because only roller friction loss goes up with power, side plate friction is constant). Inertial effects remain the same in terms of energy, so reducing the time by a factor of 5 just increases the power needed by the same factor. So, assuming you can find somebody with the 5kW* burst needed (and the long chainstays required to get the required grip off the line and keep the front wheel somewhere hear the ground), the big ring drive needs an extra 2.5W to accelerate but loses about 8W less in friction compared wit the small-ring drive.
*This is simple scaling from my previous example, which turns out to be wrong by a factor of more than 2. The kinetic energy of a normal sized track sprinter (100kg including bike and clothing) at 36mph (16m/s) is 26kJ (ignoring the small additional inertia of the rotating parts), so he needs 2.6kW for 10s just to accelerate, plus a rising amount to overcome drag which will be of the order of 700W at full speed. We can disregard drag if we just deal with the first rotation of the crank, but that still places an upper bound on acceleration rate on 100" gearing of 12rpm/s, unless you can find a rider who can exceed the 26W/kg starting effort described.
We were talking about going from 0 to 120rpm, on Pro sprint gearing that's from a standing start to about 36mph. Nobody can do that in 2s. If they could, they would need to make about 5 times as much power in our example, so chain friction loss would be nearly 5 times as high (nearly, because only roller friction loss goes up with power, side plate friction is constant). Inertial effects remain the same in terms of energy, so reducing the time by a factor of 5 just increases the power needed by the same factor. So, assuming you can find somebody with the 5kW* burst needed (and the long chainstays required to get the required grip off the line and keep the front wheel somewhere hear the ground), the big ring drive needs an extra 2.5W to accelerate but loses about 8W less in friction compared wit the small-ring drive.
*This is simple scaling from my previous example, which turns out to be wrong by a factor of more than 2. The kinetic energy of a normal sized track sprinter (100kg including bike and clothing) at 36mph (16m/s) is 26kJ (ignoring the small additional inertia of the rotating parts), so he needs 2.6kW for 10s just to accelerate, plus a rising amount to overcome drag which will be of the order of 700W at full speed. We can disregard drag if we just deal with the first rotation of the crank, but that still places an upper bound on acceleration rate on 100" gearing of 12rpm/s, unless you can find a rider who can exceed the 26W/kg starting effort described.