Thread:Cookie-san/@comment-26247760-20150908220735/@comment-26348726-20150913204302

Haha yes, but he likes his (original) waifu Akizuki too much to bear not including her in this experiment :P He uses Akizuki as FS for basically everything that he can. I don't think it really matters whether he used Akizuki or not, since there's absolutely no way that the fully-loaded Kongou classes would have been able to deal just DD-level damage to those lightly-armored I-class and Ro-class DDs, even on T-red. Only those two DD classes were counted as targets for the purposes of the experiment (they conveniently have the same evasion), obviously hits on the boss were not included.

As for a 'formula' for the effect of improvement on accuracy, the most often quoted one has been '2*sqrt(star)', but that is clearly overestimating the actual effect. There's still not much organized effort going into the isolation of improvement accuracy bonuses, so it's really hard to come up with a better formula, and it is still unknown whether these accuracy bonuses stack linearly with multiple improved guns equipped. So for now, it's better to rely on ballpark estimates for +9/+max accuracy bonuses. An estimate that is most representative of the consensus right now is +3.5% +/- 1.0% per +9 equipment, and the types of equipment demonstrated to exhibit this accuracy bonus so far are all of the improvable guns and shells. Radars and sonars should logically exhibit this behavior too, but very few data are available...and as far as I know, I don't think AA guns and torps have been ruled out from this effect either. Torp improvement has been hypothesized to improve just the accuracy of the torp phase (and not the shelling), but who knows...