Dragoons: A Rotation Reborn

[EasymodeX]

Does this take into account the head slots? I know the mythology one has a stupid amount of accuracy

Yeah, but switching to Myth (acc) from Allagan (no acc) also loses 21 crit. The weight cost per acc gained is -0.159 which is slightly worse than the middle pieces which are around -0.14 through -0.149.

E.g. switching from Myth Gloves (no acc) to Allagan (acc) loses 15 skill speed, which is -0.146 value per acc.

After fiddling around a bit I made a second set which is nearly identical in overall dps stat weight (1 less out of 618), but has 4 more acc:

A: boots, bracelet, earring, gloves, helm, leggings, ring1 (16acc)
M: belt, chest, neck, weapon, ring
[470 acc total]

As opposed to the one I listed earlier:

A: boots, bracelet, earring, helm, leggings, ring1, ring2
M: belt, chest, gloves, neck, weapon
[466 acc]

Edit:

I've ported the data to google docs. Download / export from there to manipulate and look at the gear combinations.

https://docs.google.com/spreadsheet/...LVWhiT2c#gid=0

[Ayvar]

As opposed to the one I listed earlier:

A: boots, bracelet, earring, helm, leggings, ring1, ring2
M: belt, chest, gloves, neck, weapon
[466 acc]

Ah, so the original groups had the helm from the Allagan set. It was just that your first post didn't have helm anywhere at all. Thanks!

[Revuhlooshun]

I've ported the data to google docs. Download / export from there to manipulate and look at the gear combinations.

https://docs.google.com/spreadsheet/...LVWhiT2c#gid=0

A minor correction: The Myth bracelets give 11 accuracy and 11 determination, not 11 accuracy and 11 skill speed.

[EasymodeX]

A minor correction: The Myth bracelets give 11 accuracy and 11 determination, not 11 accuracy and 11 skill speed.

Fixed, ty.

This can very easily be solved using binary integer programming, and is doable with excel's solver
[...]
Let me know if you want to see the modified sheet.

I have no idea what binary integer programming with excel's solver is; that sounds awesome. Please do send me a copy.

[Ferboten]

I have no idea what binary integer programming with excel's solver is; that sounds awesome. Please do send me a copy.

It's basically a type of optimization problem where the variables are binary valued, everything is linear, and there are certain constraints that need to be met (like high enough accuracy, and only 1 piece of equipment per slot).

Here's the spreadsheet: https://mega.co.nz/#!ecUDSDZB!LBH82A...bJ5BvC98sq6Z18

The solver can be accessed from Data -> Solver. From there you can see that our goal is to maximize the total weight by changing the selection cells for the items, subject to the above constraints, and as well constrained so that the variables are binary. After that, make sure the solving method is set to "Simplex LP", hit solve, and it will find the best solution.

[Zerosix]

It's basically a type of optimization problem where the variables are binary valued, everything is linear, and there are certain constraints that need to be met (like high enough accuracy, and only 1 piece of equipment per slot).

Here's the spreadsheet: https://mega.co.nz/#!ecUDSDZB!LBH82A...bJ5BvC98sq6Z18

The solver can be accessed from Data -> Solver. From there you can see that our goal is to maximize the total weight by changing the selection cells for the items, subject to the above constraints, and as well constrained so that the variables are binary. After that, make sure the solving method is set to "Simplex LP", hit solve, and it will find the best solution.

Yo Ferboten from rift? The one who made the spreads and calcs? Welcome on board to drg if that is you, my man.

[Ferboten]

Yo Ferboten from rift? The one who made the spreads and calcs? Welcome on board to drg if that is you, my man.

Yep, that's me

[Ferboten]

I've ported the data to google docs. Download / export from there to manipulate and look at the gear combinations.

It seems like there is a lot of work being done to find out the highest scoring set of gear that meets certain accuracy requirements, but there is a better way that is both faster, and can achieve the best possible result.

This can very easily be solved using binary integer programming, and is doable with excel's solver (but not google docs, unfortunately).
I quickly whipped up some solver parameters for your spreadsheet, and ended up with a set that gives 470 accuracy and total weight of 616.959 compared to your 616.958. I realize that is a minuscule difference, but we're all about minuscule differences here, aren't we?

If you swap to using both allagan rings and the 0 acc allagan necklace, you should get the same score.

Let me know if you want to see the modified sheet.