2013-06-02, 01:37 AM
I tested it out a bit and looks like attack gives a huge rounding error for anyone with %attack potential combined with low values of delta A. Got a discrepancy of over 100 when changing stuff around for some reason. I'm guessing MS rounds attack to integer values before being used in the range formula, so that invalidates the assumption I made about how much error is present when calculating attack.

