Dragoons: A Rotation Reborn

[EasymodeX]

It's not overly complicated. Accuracy is worth whatever you trade it for. In this model we try our best to trade acc for skill speed. In some cases, acc for crit, and in the rarest cases acc for dtr.

Picking crit as an average, HQ Eggs caps at 24 crt, 10 acc (crt). This is 34*.184 = 6.26.

Buttons caps at 16 dtr and 10 crt, which is 7.01. Although I think with normal base stats, buttons will only be 9 crt, and eggs would be lower at like 22-23 crt. Also egg acc would be ~9. Shrug.


If I use a dynamic calculation then I have to run a GRG Nonlinear solve which takes about a billion times longer than a simplex linear solve, and isn't even necessarily accurate (but should be because solving for this isn't exactly difficult). If I use a hardcoded value for food then the error is at least localized to the food.

So I just manually override food with buttons and run a linear solver from there.

Edit:

And as Ayvar said, I run it once with each food as override and compare the final #s. They are accurate within the "hardcoded value food error" which is like ... 1 acc + 1 crt or something at most.

[AugustusSB]

^^ I'm an engineer and just brute forced the hell out of this with a VBA script in excel. It runs through every single possible armor combination there is in my database and picks the global maximum of DPS weight while equal to or above the acc requirement i enter. But I do like your approach, would of saved me some time if I thought about it from that perspective initially. Script usually takes about 30 seconds to a minute on this machine to go through it all. However with every new piece of gear I add I double the amount of combinations by two and depending on a whole lot of things could multiply the calculation time to who knows hahah.

Edit. In between every gear set possibility I apply the food buffs to the current set and do my rounding (flooring). It should be pretty accurate provided the game doesn't have any weird order of operation/rounding/truncation rules