Talk:MTR Psi Testing

Perhaps the problem is...overflow?

The C standard requires unsigned integer math to be modulo arithmetic. This would show (when varying Psi Skill) as the "cutoff" from 100% to 0% MC success.

For a linear equation, the overflow point would be somewhere near 200 (from the problem with Skill 50 vs. Skill 255)

-- Zaimoni, 10:32AM Sept. 3 2006 CDT

I'd be fairly surprised if that were the case since I'm pretty certain the attack formula does involve Str*Skill, and Mike had Str set at 0. But his results definitely suggest different formulas for X-COM vs. alien than for alien vs. X-COM. If I take up testing again, the first thing I'd want to do is find the "border" cases -- 100% success and 0% success -- and change variables from there to see what they do, one by one.--Ethereal Cereal 11:19, 3 September 2006 (PDT)