Regarding critical damage with the revamp

I'd also like to point out that I mis-applied a correction factor in my earlier calculations (namely, subtracting 0.5 from damage values to account for damage range rounding down; except that I accidentally subtracted 0.5 from the slope of the regression as well).

Amended:

Looking at iAmFear's data:

Max damage is 939

+50% mastery is 610
+30% mastery is 423
+20% mastery is 329
+0% mastery is 141

Linear regression on the four points yields y = 140.73 + 9.38x, where y is minimum damage and x is mastery percentage. Using this equation, y = 939 occurs at x = 85.10, or +85% mastery. This means that base mastery is likely to be 15% (so that +85% totals to 100%).

At x = -15 (0% mastery total), y = 0.03, nearly 0. This means that all of the damage is dependent on mastery.

Looks like Cyanne's approach was right in the first place.