In theory, yes - in practice, the measurements are often far from exact (and I do know some staging drivers which apply 'cosmetics' to the values for certain devices…) and not very reliable. There's only one way to find out, test it.
In theory, yes ? Let's say that the measurements are correct. The SNR is the same coef. I just don't get the part if we have bad noise and strong signal, what will be the effect on the throughput. It may cause more jitter or who knows, but jitter isn't the throughput measurement after all so I want to know if I should focus on strong noised signal or lower signal with less noise for more throughput if I am forced to choose between these two options.
Let's say that the scenario is even more pronounced. -85 / 95 vs -60 / 70 , which one wins ?
Personally I think it would be -60 /70, but I'm not sure if that's just me.