The Brewmaster Damage-Tankiness Continuum

[Jocias]

http://www.wowhead.com/spell=126060

Pretty sure EH still has no CD and is spammable below 35% health.

[The Inevitable]

Pretty sure EH still has no CD and is spammable below 35% health.

Jocias is correct. While below 35% HP, EH can be spammed as long as you have energy due to Desperate Measures. And with the glyph, EH's energy cost is reduced by 5 while below 35% HP (above 35% HP its still 40 energy).

Because Desperate Measures has no CD on its effect, I was wondering if SimulationCraft's TMI value (and other stats) takes it in account since it significantly improves our survivability and self-healing capability.

[Nestar]

Jocias is correct. While below 35% HP, EH can be spammed as long as you have energy due to Desperate Measures. And with the glyph, EH's energy cost is reduced by 5 while below 35% HP (above 35% HP its still 40 energy).

Because Desperate Measures has no CD on its effect, I was wondering if SimulationCraft's TMI value (and other stats) takes it in account since it significantly improves our survivability and self-healing capability.

TMI absolutely takes that into effect. It looks at the worst 6 seconds over the entire duration of the sim. Whether Desperate Measures is implemented, I'm unsure.

The conditionals I put on both the expel harm line and the jab line ensure it won't use either one if it will cause a delay to the cast of keg smash (since keg smash generates significantly more chi per energy).

[Totaltotemic]

I have limited capabilities here at Blizzcon, but the static TMI nimber SimC presents is only the single worst TMI spike right? And the graph is the distribution of the next highest potential TMIs right? If that's the case, do scale factors merely represent the difference in that largest spike, or does it take into account the entire distribution?

[The Inevitable]

TMI absolutely takes that into effect. It looks at the worst 6 seconds over the entire duration of the sim. Whether Desperate Measures is implemented, I'm unsure.

Ok, fair enough. Do you happen to know where Desperate Measures would be implemented if it was implemented? Is it some kind of script in the program options somewhere or is it in the source code?

The conditionals I put on both the expel harm line and the jab line ensure it won't use either one if it will cause a delay to the cast of keg smash (since keg smash generates significantly more chi per energy).

I would argue that prioritizing survival by using Expel Harm when in the danger zone of sub-35% HP likely should take precedence over Keg Smash. Basically, the choice comes down to whether 1 Chi + effective healing for 35 energy would be better/worse than 2 Chi for 40 energy (basically its 1 extra Chi for purifying) in a low HP situation. Also, its entirely possible that, during the time you're below 35% HP, Keg Smash will be on CD and there won't be a choice at all. What do you think?

[Nestar]

Ok, fair enough. Do you happen to know where Desperate Measures would be implemented if it was implemented? Is it some kind of script in the program options somewhere or is it in the source code?



I would argue that prioritizing survival by using Expel Harm when in the danger zone of sub-35% HP likely should take precedence over Keg Smash. Basically, the choice comes down to whether 1 Chi + effective healing for 35 energy would be better/worse than 2 Chi for 40 energy (basically its 1 extra Chi for purifying) in a low HP situation. Also, its entirely possible that, during the time you're below 35% HP, Keg Smash will be on CD and there won't be a choice at all. What do you think?

Last I remember seeing the cooldown reduction in code was as part of the glyph and that's why I was under the impression it was gone (was being removed from there at the time). It would be a reference under expel harm to being under brewmaster spec.

As far as its value vs keg smash, adding that cooldown in is where it gets a bit wonky. Because you are constantly under 35%, if you leave the cooldown off. The sim never uses jab (because expel harm is better in absolutely every way). And honestly that's never a situation you would be facing in game. If it was to be implemented there would have to be a way to keep it somewhat realistic.

[The Inevitable]

I found the following starting on line 2189 in class_modules\sc_monk.cpp:

virtual double cost() const
  {
    double c = monk_melee_attack_t::cost();
    if ( player -> health_percentage() < 35 && p() -> glyph.expel_harm -> ok() )
      c += p() -> glyph.expel_harm -> effectN( 1 ).base_value();

    return c;
  }

but it seems only to deal the the cost of Expel Harm and not the CD.

As far as its value vs keg smash, adding that cooldown in is where it gets a bit wonky. Because you are constantly under 35%, if you leave the cooldown off. The sim never uses jab (because expel harm is better in absolutely every way). And honestly that's never a situation you would be facing in game. If it was to be implemented there would have to be a way to keep it somewhat realistic.

I see how that poses a problem. However, since TMI is looking at the worst 6 seconds, then it could be presumed that Expel Harm actually would effectively replace Jab during that period. And because its only 6 seconds out of the whole fight length, for purposes of calculating TMI, maybe SimulationCraft could just be set to where Expel Harm never triggers a CD. That is, no data outside of the worst 6 seconds would matter with regard to getting a more accurate representation of TMI, so it wouldn't matter how that change to Expel Harm affects the rest of the data (I hope that makes sense..).

[Nestar]

I found the following starting on line 2189 in class_modules\sc_monk.cpp:

virtual double cost() const
  {
    double c = monk_melee_attack_t::cost();
    if ( player -> health_percentage() < 35 && p() -> glyph.expel_harm -> ok() )
      c += p() -> glyph.expel_harm -> effectN( 1 ).base_value();

    return c;
  }

but it seems only to deal the the cost of Expel Harm and not the CD.



I see how that poses a problem. However, since TMI is looking at the worst 6 seconds, then it could be presumed that Expel Harm actually would effectively replace Jab during that period. And because its only 6 seconds out of the whole fight length, for purposes of calculating TMI, maybe SimulationCraft could just be set to where Expel Harm never triggers a CD. That is, no data outside of the worst 6 seconds would matter with regard to getting a more accurate representation of TMI, so it wouldn't matter how that change to Expel Harm affects the rest of the data (I hope that makes sense..).

It's the effective health change though, so being able to heal every single time you would use jab previously, will affect TMI. It won't affect dtps however.

I think the best way to handle it is going to be having it available, but not on by default. Some sort of APL option that allows it to be used off the cd it can be tested while the default apl and ranking lists remain unaffected. I took a look at the previous code for it and enough changes have been made it will no longer work, Perhaps Hinalover can give better insight into that.

[Deleted]

What is your opinion leveling as Brewmaster instead of Windwalker due to Ox Statue being kinda good for bigger pulls, worth it or not?

[tachycardias]

@above: I'm planning on it, but thats pretty of topic.

On topic- I don't understand how Agi could be high in our stat priority, as it is listed above in Hinalover's graph. It really only gives dodge, and gives even less than it did in MoP.
My priority I think will be something like (assuming Serenity, and I have no plans to use any other lvl100):
Mastery
Crit
MS/vers based on the fight
Agi
Haste

Versatility obviously increases in value a bit in magic heavy fights (though its scalar remains poor). MS increases in fights where you can remain in the same area for a while. Also, MS will be kinda nice in that you can stand on orbs and have them auto-heal (and not consume more orbs to over heal) when you take damage. This somewhat offsets the frequent downside of healing, which is the split second gap required between damage taken and healing received. Also, Tiger Strikes throws a funny twist to MS, amplifying its power.

I think Crit is really undervalued. I'm not sure how accurately Theck evaluates Crit. I know we won't be able to keep it up as much as we are in 6.0.2, but it has been tuned to be more valuable despite this. Also, the value differs considerably depending on whether you're taunt switching, or whether there are periods of high burst where you can make sure to use EB.

I hate to say TMI is less valuable just because I don't fully understand the logic behind the equation, but I think basic logic can win out in certian circumstances and one would be wise not to ignore this logic in favor of some equation that is touted as the be all-end all of Tank stat evaluation. I know there are several caveats listed by Theck, but unfortunately these caveats often get ignored by people casually searching the internetz and looking for "OK WHICH STATS SHOULD I DO MMOCHAMP?!?".

[Tech614]

What is your opinion leveling as Brewmaster instead of Windwalker due to Ox Statue being kinda good for bigger pulls, worth it or not?

levling as a brewmaster is fine, you pretty much global mobs for the first 5 levels anyways if you are SoO geared as a brewmaster. No point in needing any more dps. Maybe for the tail end of leveling wind walker may edge out, as the mobs start to get a bit beefier as your stats get worse.

[Totaltotemic]

On topic- I don't understand how Agi could be high in our stat priority, as it is listed above in Hinalover's graph. It really only gives dodge, and gives even less than it did in MoP.

Agility still gives AP, which still increases self-healing. In fact, it gives even more AP than Bonus Armor does. The problem in MoP was just that the only source of Agi decision making was in gems and you'd only get half as much as any other stat. In WoD, the only place for decision making for Agility is in trinkets, but then it's competing with Stamina which gets a lot more complicated.

As for all of the other stats, everyone needs to try to keep in mind that BrM is not a DPS. TMI is a more useful metric on its own than DTPS - HPS on its own in terms of how likely you are to die, but there is no simulation that's a silver bullet to give you tank stat weights. You have to take the boss the simulation is using into account, which in the TMI Standard Boss's case is a simple AA with a magic DoT that does not stack that is tanked by the same tank for the entire fight. Much like the Patchwerk style DPS simulation, this doesn't actually exist in-game anywhere. The difference between DPS sims and tank sims though is that DPS only cares about overall performance (not second to second) and there are a ton of situations in real raids where you stand still and do DPS for prolonged periods of time.

That's why we have to do theorycrafting to really understand tank stats; a simulation can't give you all of the answers no matter how it's set up. That's not to say that it's useless, it's actually very useful, it's just not the end.

DPS gear by:
1) Sim yourself (provided the sim isn't bugged and the APL is optimized)
2) Plug stat weights into calculator, do what it tells you.

Tanks gear by:
1) Sim yourself (provided the sim isn't bugged and the APL is optimized, which is a lot harder to do than it is for DPS)
2) Examine the encounters that are in your progression future and determine their damage patterns
3) Look at the boss's attacks and determine in which situations the sim is even relevant
4) Utilize a combination of TMI, DTPS, and HPS while taking into account that the sim does not actually reflect a real boss in the best way that you can to determine which stats actually make you the tankiest
5) Use DPS stat weights to determine which stats are also good for doing damage
6) Figure out what level of compromise between damage and tankiness you are willing to do

What sims can tell us that we wouldn't otherwise know is that Bonus Armor is good in all normal circumstances and Haste is generally bad in all normal circumstances.

We know this because Bonus Armor is so much better at reducing damage taken and increasing both HPS and DPS than any other stat that something abnormal would have to incredibly devalue it to bring it down below other stats. Those would only include armor reductions (which annoyingly take away the AP as well) or extremely heavy magic damage.

We also know haste is generally bad because it sims incredibly badly in every category (even with Chi Explosion) and there are no extraordinary circumstances that would make every single other stat worse while leaving Haste untouched or make Haste drastically increase in value.

But that's why those are the only "right answers" BrM has, the other stats all interact in different ways on different fights and have different strengths and weaknesses. This is why there are no numbers at all in the OP; it's about the philosophy of tanking stats themselves rather than how numerically good they are. Whether, say, Versatility or Multistrike is overall "better" isn't really a statement you can make. Even if Versatility totally outclassed Multistrike in both HPS and DTPS MS would still provide more damage. Even if MS totally outclassed Versatility in HPS and DPS it still falls behind Versatility in DTPS.

Sims are just like the old raidbots website (when will they ever update that to pull from WCL instead of WoL?), a useful tool if you know how to use it, but disastrous if you misuse it.

[Gethe88]

Sims are just like the old raidbots website (when will they ever update that to pull from WCL instead of WoL?), a useful tool if you know how to use it, but disastrous if you misuse it.

WCL already has a raidbots-like feature. http://www.warcraftlogs.com/statistics/5

[Totaltotemic]

WCL already has a raidbots-like feature. http://www.warcraftlogs.com/statistics/5

Oh that's nice, awesome.

[Hinalover]

I found the following starting on line 2189 in class_modules\sc_monk.cpp:

virtual double cost() const
  {
    double c = monk_melee_attack_t::cost();
    if ( player -> health_percentage() < 35 && p() -> glyph.expel_harm -> ok() )
      c += p() -> glyph.expel_harm -> effectN( 1 ).base_value();

    return c;
  }

but it seems only to deal the the cost of Expel Harm and not the CD.

I don't understand what you mean by not taking into account of the CD. The glyph does not change the cooldown of spell, other than lower the value of haste.