Relative to the above, I've become far less fond of the d20 as a resolution mechanic. Every number having the exact same chance of appearing as any other has really started to bother me. Yes, modifiers increase the chance of success, but ultimately what the die is saying is that extreme success is weighted just as evenly as extreme failure. I like dice pool mechanics, but they too don't seem to really reflect the fact that an average result is far more likely in any given situation than the extremes. Having more dice may mean more chances of success, but it doesn't really mean more average success.

I played around with using 2d10 instead of 1d20. I liked being able to call a "nat 1" "snake eyes". It did better with providing average levels of success which on the whole produced a smoother game, but I think my players preferred the "swingyness" of the d20. Being able to reliably do something, especially if you were good at it, seemed, in my testing, less fun than betting extreme success against extreme failure.