Hawk37 - when Black Baron talks about a 20% reduction in thrust he is correct. What he has done in his "reduced climb" is derated his engines through the FMC. This allows us to lengthen engine life, reduce wear and tear etc etc. I dont have the figures for the 767 but I do for the aircraft I am familar with, 737-400, from doing my ATPL exams.
The 737-400 comes usually with engines that are flat rated at 23,500lbs static thrust. However we can de-rate the engines to 22k or 20k through the N1 page on the FMC. We do this when we can still safely achieve our performance requirements with the use of less thrust, in order to reduce wear and tear on engine, extend engine life and reduce noise. At London Stansted (where I dispatch) over the last few days daytime temp has not got much over 5 deg celcius. So its pretty much ISA - 10 deg which is great for performance (the fog however isnt great for flying!!!). With a long runway and light pax load we dont need 23.5k, so you can de-rate to 22k. If we went for 20k you can see that we have reduced thrust by approx 15%.
On a cold day the air is more dense so the engine produces more thrust. The wing produces more lift because the air is more dense, so the climb gradient is higher.
The other thing you need to consider is that performance charts for GRADIENT of climb are only really applicable at low altitude when you want to clear an obstacle in the take off flight patch - ie as at London City, all those big buildings nearby.
With a jet airliner what you are really interested in is getting to cruise altitude in the shortest TIME as that make the aircraft more economical. Hence you aim to fly a jet at best rate of climb speed for all but the initial part of the climb.
Remeber max angle of climb (GRADIENT) - max altitude gain over shortest distance over the ground. (Vx)
Max rate of climb - largest altitude gain for shortest TIME in the air (Vy).
On the 737 FMC you can select from the climb page max angle or max rate climb.
I think the problem, Hawk37 , is that you are trying to make a hypothetical "text book/labatory" situation fit the real world. Interesting discussion though and useful performance refresher as it was some months ago I sat (and passed) the exam.