Ok, fair challenge Tony. So here goes....

Firstly, I have an issue with your premise that the jocky wheels on the rear derailleur "sap your exergy!" I'd like to see figures on their consumption. In the absence of figures, my assumption is that the increase power needed would almost be immeasurable, and certainly not noticeable to a rider. I find it hard to imagine Lance Armstrong powering up the Alps cursing his jocky wheels!

Secondly, in your suggested system the gear change would be almost unworkable. To keep the chain at the correct tension, it would have to move from one sprocket to the next instantly and simultaneously on the front and rear mechs. I'm not sure this is achievable, and if it were I think there would be power loss during the change.

Thirdly, in order to keep the chain at the correct tension you would only have certain matched pairs of front and rear sprockets available. This would limit your gear ratios. Keeping an efficient, and comfortable, cadence would become a problem.

Fourthly, the front mech has traditionally had a maximum of 3 sprockets. If we were to increase that to say 6 (to give 6 gears) the pedal separation would have to increase to gain the required extra clearance. You can only make the pedals so far apart before it becomes illegal for girls in skirts to ride them!

Fifthly, Na only joking, I've run out now. On a more positive note, I fairly sure that there was a car maker that looked at a similar idea. It used two cone shaped drums instead of sprockets, and linked them with a semi-flexible metal band. The band was moved up or down the cones to achieve similar results. Not sure how far the got with it.

Lastly, just looked at Wikipedia to check I'd not missed anything blatant. I think the car method I described is under the heading "continuously variable transmission", though I couldn't be bothered to read it all. An experience the current audience can probably sympathise with!

Last nail in the coffin....it seems that the rear tensioner is a bit of an unsung hero on more fronts that I had previously thought. Wikipedia points out that possible failure modes in chain driven systems are numerous, including thermal expansion, and foreign object entanglement. I reckon both these problems would knock your system for six, whereas my tensioner would eat them for breakfast!!

In the spirit of your blog, if your original post didn't put people to sleep, I'm sure my response has made it dangerous for them to operate machinery - chain driven or not!!

Bill