idea: do a match and after get a "sensitivity" that maximazes your accuracy.

[Kirin]

and R_R no.. the sensitivity can be estimated only if the "error" you produce is almost the same all over the training dataset that it captures

for example when i tried the algorithm using a sensitivity which was too high for me, my aim was very jittery (is that even a word?) and at the end of the training the optimizer wasnt able to find the optimun because the samples i produced had a low signal to noise ratio (if i can call it this way in this issue)

edit.
i have to thank Ronald_Reagan for helping me test, and for opening my eyes on how foolish it is to try to estimate sensitivity by confronting player aim and ideal aim, this can never work right.

so today i completely rewrote the whole thing into a accuracy pilotated bisection over a interval of sensitivities, which seems to always converge to the right value plus it works with any weapon
Occam's Razor "theorem" is once again right! XD

source code for this hopefully final release will be coming tomorrow :)

one pitfall for this method is that it requires more time from the user than the one before. but it's worth the result imo.