Agile Monk

[Deleted]

Spreadsheets were merely a setback, a stepping stone to a much larger plan !

Hi !

I used to do some Monk theorycrafting since the MoP beta and i made a few spreadsheets, and i thought i could do something more for the release of the 5.4 patch. I've developped a little program to help players figure out stat weights, and the optimization of choices like Xuen/RJW.

Here it is :

http://www.filedropper.com/agilemonk_1
(click on the link "Download this file")
(updated link, see next message)
(you may face problems when you use this version depending on the country/language settings of your computer, try this version for better results : http://www.filedropper.com/agilemonk_3)

It only works for Windows. And Windows might not want to run it, it'll consider it as an unsafe application. My own computer didn't want to run it after i downloaded the one i uploaded on FileDropper. When Windows blocks the application or warns you that it might be dangerous, you should have an option to run it nonetheless. On Windows 8, you should click on "Additional Information", or something like that. Anyway, don't worry, there's nothing harmful in the "Agile Monk" file.

For now, only part of the Windwalker module is implemented. I will soon begin implementing the Brewmaster one. Eventually, there will be a stat weight calculator, a simulator, a "base calculation" mode where you can do things like comparing talents or dps priority between a few spells, and plenty of other stuff like accurate scaling from a combatlog, for both specs (and for both dps and damage reduction for BM).

This application is therefore in development, and it will remain so for a while, until i've added everything i can about Monk theorycrafting. I'm planning on using it as the base of a big theorycrafting application for the release of the 6.0 patch. So, even though the Windwalker Calculator is almost "finished", i still need to add quite a few things.

I hope i made the application clear enough for people to use it without much explanation. If it's not the case, please tell me and i'll arrange that. Basically, you just need to enter your stats, trinkets, etc... on the first page, then customize your targets on the second, choose your spells on the third, and click "Calculate".

And, well, it would be interesting to discuss the results given by the application. For instance, we could talk about the new Mastery. Or about the Xuen/RJW choice. You can try it yourself : enter your stats, and make on calculation with Xuen, and one with RJW. Then add one target in the "Fight" tab, and run another calculation.

You can repeat that until RJW outdamages Xuen. But it should happen pretty rapidly (by which i mean : two targets are enough for RJW to beat Xuen, even if in this case RJW won't generate chi). It might be surprising that an aoe spell can be effective against two targets, but RJW is quite a powerful spell. The damage of most Monk attacks, including RJW, Jab, Blackout Kick and Tiger Palm, is calculated by multiplicating a base formula (which we can call "Weapon DPS") by a coefficient corresponding to the ability. For example, the coefficients of these spells are :

Jab : 1.5
Tiger Palm : 3
Blackout Kick : 7.12 * 1.2 (the 1.2 comes from the 20% increase due to the DoT)
RJW : 12.59564659 if you add its 9 hits - RJW does nearly as much damage as a Rising Sun Kick !

If you spend 40 Energy on RJW on a single target, you will deal damage equal to, on average, base * 12.59564659.
If you spend 40 Energy on Jab, you will deal damage equal to, on average, base * 1.5, plus the Blackout Kick offered by the two Chi you generate, plus the Combo Breaker procs, for a total of base * (1.5 + 7.12*1.2 + 0.12*(3 + 7.12 + 0.2*8)) = base * 11.4504
(why 7.12 + 0.2*8 for the CBBK proc ? because the DoT of the BoK proc deals 20% of the damage of a coefficient 8, not 7.12 ; yes, a Blackout Kick from a Combo Breaker proc deals more damage than a normal Blackout Kick, it's a bug that appeared during the MoP beta and that has never been fixed)
One could argue that the CBTP proc also gives an opportunity to not manually refresh Tiger Power, thus giving damage equal to half a Blackout Kick minus one Tiger Palm. It's possibe to include this in the calculation, but let's make it simple and trust me, the result would be the similar in the end.

So 12.59564659 > 11.4504, RJW deals more damage than Jab + the rest ! So if you ran a calculation in Agile Monk with one target and RJW, you should get a bigger DPS than without RJW. But if you do it, you'll get a smaller DPS with RJW. Why ? Because of Tigereye Brew of course ! As RJW doens't generate Chi on a single target, TEB generation is slower, and the resulting DPS is inferior. But if you run the application with two targets or more, the massive damage of RJW makes it profitable against two targets, even if it doesn't generate Chi.

What's interesting is to try and figure out what is the best between Xuen and RJW. Well, in most cases, the application will tell that RJW is better than Xuen if you have two targets or more. And yes, that lowers the stat weight of Mastery, since you don't generate a lot of Chi.

But remember that this calculation considers that those two targets are present during the whole fight ! If you have some cleaving to do only every 3 minutes, Xuen might be ahead against more targets. It's possible to represent such situations by manipulating the "Use" values in the "Calculation" tab, but instead of explaining how, i've begun working on an addition to this tab to make it simple to tell the application when you've got to do some AOE.

There are more topics to debate on with this application, and i would be glad to hear what you'd like me to add, or whether you think i've made a mistake somewhere.

[Deleted]

Hi, i'm back with a little update on the application :

http://www.filedropper.com/agilemonk_1

It is now easier to compare the efficiency of your AOE spells, thanks to the addition of cells indicating the percentage of time when you do some AOE. For example, if the additional enemies you enter in the "Fight" tab are present 15 seconds every minute, then you should enter (15 / 60) => 25% in the AOE % cell. Similarly, if they are present for 30 seconds every 200 seconds, then if you want to use Xuen, enter "200" in its "Delay" cell, and 66 in its "%" cell which represent the percentage of time when its cleave hits targets while Xuen is out (30 seconds out of its 45 seconds duration is 66%).

For example, with the default gear settings of the application, if you add one lvl 92 target with 20k armor and the Physical Damage debuff, if we suppose that this target is present 30 seconds every 3 minutes, you can launch those two calculations :
- with Xuen : 180 sec "Delay", 66% cleave efficiency => 247k dps
- with RJW (don't forget to disable Xuen by entering 0 in its "Use" cell) : 30 sec out of 180 is 16.7% => 243k dps
Xuen wins.
Then we add a target :
- Xuen : 258k
- RJW : 267k
And if you keep adding targets, RJW becomes clearly better. But the difference with 2-3 targets would be more in favor of Xuen if the targets have the 5% Magical debuff instead of the 4% Physical.

I also added a checkbox indicating whether the damage of FoF is split among the targets of the "Fight" tab, or if they only hit your main target (by default, in the application the checkbox is not checked, which means FoF only hits your main target). You can use this checkbox to suppress the damage done to additional targets by your AOE spells. You can do this quite simply : have at least two more targets in the "Fight" tab, so that RJW generates Chi, and give to those targets a gigantic Armor value, so that the damage of RJW on these targets is negligible. And uncheck the "split" checkbox of FoF so that its damage is not reduced by the Armor of the additional targets.
You thus get the dps of RJW on your main target. It can be useful for fights where cleaving is not useful (the 9 klaxxi for example) ; remember that as long as RJW generates 1 Chi, it's a dps increase on your main target compared to Jab + 2 Chi. The problem is to know whether this dps increase is superior to Xuen's dps, and the Agile Monk application can help you answer this question...

[Crevan]

There's something wrong with the application. After entering my stats and hitting "calculate", I get a 30-digit dps number, with weights for 2-3 stats of the same order of magnitude, others being NaN. Unless, of course, the calculator isn't supposed to work yet in this version.

One more thing I noticed is the Cooldown / Delay = Use display for abilities is a bit wrong (or implemented in a funky way). Seems that the formula shows a correct "use" value only for integer values of "delay". Using a number with a decimal part makes the "use" field display Cooldown / (Delay * 10) instead. Selecting AoC trinket also does not affect the (visible) cooldowns of FoF and EB.

[Deleted]

TherUnless, of course, the calculator isn't supposed to work yet in this version

Nope, the calculator should work (and it does on my computer, and on that of my guildmates). Only the cooldown reduction of AoC shouldnt (and it's the only trinket special effect i haven't implemented).

I get a 30-digit dps number
[...]
Using a number with a decimal part makes the "use" field display Cooldown / (Delay * 10) instead

Those problems might be linked to your operating system, and particularly its language ; what os are you using (i don't think the application works with anything other than Windows, but which version do you have ?), and what is its country/language setting ? It might be a problem of writing 1.5 (one point five) istead of 1,5 (one comma five), although i'm quite sure i coded the application so that both would work. And both do work on my computer, and on the computer of people in my country. Could you post a printscreen of what you get in the "Use" cells ? And are you sure the "Use" value is indeed divided by ten ? Normally it should vary between 0 and 1 depending on the delay ; it can exceed 1 if the delay is shorter than the cooldown.

Concerning the 30-digit dps, it could also be a country/language problem. Try this and tell me if it solves it :
http://www.filedropper.com/agilemonk_2
And if it doesn't work : tell me what you get in the "Formula" tab when you check the "All" checkbox at the bottom of the page.

Additionnally : what do you get if you don't modify anything in the "Character", "Fight' and "DPS Calculation" tabs, and simply click on "Calculate" ?

[Saltycracker]

I was seeing the same issue when I tried this spreadsheet. It might be regional differences, or how you have your row-column notation setup.

[Hinalover]

I think you are overvaluing haste. You currently have haste; at least on my character; as a 4.8 scaling value. What I did was I ran my toon through SimC and adjusted the DPS Calculation tab's delay to match that of SimC. First up SimC is reporting a dps of 258k dps while your program is providing a 247,778.1 dps Secondly the haste is showing as 4.84375 scale value while simC is providing a 2.84 scale value. Finally considering I've pretty much hit a resource cap of sort, the haste's value has diminished and thuse lowers in value slowly



As you can see, between now (8230 haste rating) and +2000 (10230) the value of haste is near a constant 2.7 value. So I'm not sure what your factoring in. Maybe you still have RPPM being affected by haste (most are not anymore)

Also I'm not sure if your factoring in the Blackout Glyph's damage or not. Did not see anything regarding the glyph in the options.

[Crevan]

I'm using Windows 7 64 bit, with latest updates. It's in English, and system locale is set to English (UK) for non-unicode programs, though I don't think that would matter much here.

Screenshots of my results: character, scaling, formula, and spells tabs.



Decimal point issue:


8/8.1 ~= 0.987654320... (repeating, nice number). The displayed value is shifted one decimal point to 0.0987654...

Note the highlighted Agility cell in the Formula tab - that's quite a bit more than I entered in the character screen.

Running the program with default values gives smaller, but still wrong results, including negative dps and scale factors for everything but haste. And it still has the same Tiger Palm anomaly present in one of my screenshots. Will try the second version later and see if it fixes anything.

[Deleted]

I think you are overvaluing haste.

Oh, yeah, sorry i didn't explain this part. The calculation doesn't take into account gcd issues. So the value of haste keeps going up even if you're energy capping, since the application will still be able to cast all the spells that its energy generation offers. The stat weight for haste is fine when you're under the "haste cap", but if you are close to or beyond the point at which you're energy capping in the game, the calculation for haste becomes less relevant. Given the way the calculation is implemented, gcd capping can hardly be introduced by reasonable means. If you want to know what happens to haste close to 8k - 10k (and beyond), a simulation is a better tool than a calculation. What the application is good at is saying "haste is very good, it has this stat weight until you cap your energy too often", then we've got so say when the energy begins capping. The high stat weight of haste is just a confirmation that haste is the number one stat until you can't fit more spells into your rotation.

I am going to implement a simulator in Agile Monk, but it's not my priority, since there's SimCraft. My priority is to develop the Brewmaster mode, and to implement its simulator (since there's no Brewmaster simulator available, whether it be SimCraft or something else).

Remember one thing though : investigation is still needed on the subject of RJW in single target, which deals more damage than Jab + 2 Chi, but which doesn't generate TEB stacks ; my application suggests that this stack loss is big enough for RJW not to be viable in single target, but since RJW raises the haste cap (by not generating Chi), what would happen with RJW and some more crit/mastery transferred into haste ? Haste being a very good stat, i haven't found a good answer to the question of whether artificially increasing its cap thanks to RJW would be a dps increase. A simulator might give answers. I haven't yet tried on Simcraft to use RJW in single target and increase my haste.

Also I'm not sure if your factoring in the Blackout Glyph's damage or not. Did not see anything regarding the glyph in the options.

The application considers that the Monk is always in the back of the target, so the glyph shouldn't matter much. Unless there's a bug with the glyph that makes Blackout Kick deal more damage ? I didn't notice such a bug though.

@ Saltycracker and Crevan :

I implemented a more "handmade" conversion of the imput numbers, it should solve the comma/point and country/language issues... i hope. Try this one and tell me if it works :

http://www.filedropper.com/agilemonk_3