I am pretty sure a person's calories burned are determined entirely by distance travelled.
I am pretty sure they are not XD.
Total spent energy = travel (movement) work + dissipative (friction) work
The first, travel work (your weight * distance, assuming round trip there is no need of accounting for elevation changes) is independent of your speed.
The friction, however, is (on a first aproximation) dependent on the square of speed. allright, the time you spend biking is less if you go fast, but that goes down linearly with speed, while the "wasted effort" goes up as the second power.
Going fast is more tyring.
Ah, about the fist post calculations, some people might eat the same (taking the bike or not) and then go to gym to burn the extra, in that case commuting saves you more that 100% of gas price!