No, I'd definitely like to see a controlled test of 5-10 different TIM's all being used in a controlled setting on a card like this. With a GPU, simply replacing the factory paste (spackling putty) with 'anything' else will almost guarantee better results...as long as you apply it correctly.
My challenge: perform this exact test (I am pretty sure there is one for CPU/TIM tests somewhere...i've seen it...) but provide controlled applications, tests, benchmarks and results. I'd bet that at VERY MOST, any TIM will perform within no more than +/-5C from the rest (not including stock *gunk*).