Talk:World 6/@comment-25733368-20160313160034/@comment-26154973-20160314003559

It's the result from solving the model (on ギミック tab) vs the data (データ tab), though a bit complicated, so you can try yourself by applying damage calculations directly to the data from データ (I tried for NB Type 3 data only, basically, 2.5 is too much, while 1.5 is not enough, e.g. many divergences for non-scratch looking damage values, while something like 1.9 is ok, 2.0 is also ok, except point #17, add DB data, equipment detection, scratch detection and inverse solvers for coefficients and that is probably what they're doing).