K. First of all, praise Elo for blutric engines, they made this easy. Due to the steady-state nature of thermopiles and the blutric system as a whole I think this is actually pretty definitive; there IS equal loss.
Test method: set up, run it to a redstone cell via blutric engine, turn on, let stabilize.
Results: Voltage was essentially the same cabled or uncabled which was what we were actually testing for here. Amperage (and thus resultant wattage) was higher in the cabled setup due to 4 sides having water as opposed to 3 in the uncabled tower setup.
No pics: Bleh, didn't realize you can't upload directly here and have to link urls and I can't be bothered; you'll just have to take my word for it. 20 thermopiles in each, the less efficient tower setup was 60.8v, 7.75a (471.7w). Efficient cabled setup was 61.1v, 10a (611.5w)
p.s. as I mouse over to hit post reply I see "upload a file"... sigh.