I'm saying that we may have to break the AC into chunks based on levels, and approximate. I really doubt a duplicate of AC can be made without seeing the source.
Using only 2 variables, with one constant (AGI) and the second incrementing by small values, I really don't see how rate of change can move from 1 to 3 to 2. It almost makes me wonder if EQ uses an approximation of a formula :P
The way I see it is that 99% of eqemu'rs won't be able to tell the difference anyway. Let's get this close by getting a lot of data, and forcing a series of best-fit lines based on the variables we can approximate. We can first approximate one variable's effect, then approximate the next, then build up from there.
killspree is working a dark elf rogue up the ranks with dodge and parry, and I'm doing a gnome rogue without dodge or parry. If we could get data of a third class rogue with dodge but not parry, then we could probably BS an equation for those skills for rogues. If we test it on another parrying class, and we get something close, we keep those skill approximates. We lock them in, incorrect as they are, and keep approximating data effects until we have wrong, albeit close, data produced.
__________________
It's never too late to be something great.
|