Jump to content


Photo

Cataclysm Balance Theorycraft and Spreadsheet Thread


  • This topic is locked This topic is locked
231 replies to this topic

#221 beromar

beromar

    Von Kaiser

  • Members
  • 29 posts

Posted 05 December 2010 - 02:18 AM

As of WrathCalc build 101201 I'm getting better results by swapping to the (no reforge), and reforging the Mastery on the to Spirit. It's an increase of 43 haste and 4 hit (putting you slightly over the hit cap instead of 1.6 under), at the cost of 48 crit for a net gain of ~31 dps.

WrathCalc is also still showing ~76 dps ahead of . Is this still accurate? I couldn't find anything on the actual ICDs, though it looks like they're being given an ICD of 75 seconds and 100 seconds respectively (20% uptime). Are these just conservative numbers until live testing can be done?

#222 Hamlet

Hamlet

    Mike Tyson

  • • Guide Author
  • 11,540 posts

Posted 05 December 2010 - 02:31 AM

Yeah, the gear that's in there isn't meant to be a BIS list. It's just a bunch of good gear that Tecton put in there when we were testing the sheet. It makes perfect sense that the other relic is better, since it avoids the waste of reforging to crit on another item. Just an example of how a little reshuffling can squeeze out better stats in ways that are really hard to completely summarize in succinct rules.

Like I noted in the guide, it's saying Witching Hourglass is best, but I need to look into that more. The difference in value between haste and mastery shouldn't justify dropping that much Int.

e: Oh of course, the averaged-out haste proc is hitting a new DoT breakpoint. That's definitely a problem. For now I'll just make a note about it, but it will be something to try to find a clever solution for next time I make a big update.

e2: and to answer the other question, it's currently assuming a 1/5 uptime for all procs (this was usually the pattern, and it leads to sensible average values here as well). One there's been more of an opportunity to test them all, I might add it minor corrections for expected proc time after the ICD.

#223 Rahucun

Rahucun

    Glass Joe

  • Members
  • 9 posts

Posted 05 December 2010 - 06:14 AM

I am not sure if people mentioned it already i kinda lost track on those posts but i would like to know something.
Lets say i use 250 mastery rating to get my solar eclipse and i get 45% dmg from eclipse. If i change to my regular gear which has 0 mastery am i keeping the 45% because the tooltip seems to be saying that way? Or am I actually dps under a 37% basic eclipse?

Thanks for your time.

#224 Guest_aceofsween_*

Guest_aceofsween_*
  • Guests

Posted 05 December 2010 - 08:14 AM

On the topic of haste procs, isn't it possible that the effect of say a trinket that procs 10% haste is greater than an increase in 10% damage for the DoTs if the duration of the proc is long enough and therefore greater than the average value of the proc?

For example, if there were a trinket whose effect duration was longer than the duration of our DoTs, you could squeeze two casts in that duration. Wouldn't this net a gain that is greater than the average value of the haste effect? I'm specifically thinking about effects in the 20 second range, which if given would allow you to cast at least one DoT (Moonfire/Sunfire or Insect Swarm) off before the effect of the trinket wears off.

I'm still thinking the effects of this out as I write this, but it should be something worth examining. The biggest problem (as far as WrathCalcs is concerned) is that this is going to be an incredibly difficult thing to model. In order for it to actually be a net-gain there would have to be several factors to consider. First of all, there's a good chance you would have to clip your DoT in order to meet the time requirement. Therefore the damage gained from the haste benefit would have to be greater than the remaining damage of your clipped DoT. That is complicated further by the effects of Eclipse.

At first, I also thought there might be a problem with delaying Nature's Grace as well, but no matter when trigger Nature's Grace, you will always get the full 15 sec. of the effect unless your half-rotation (from one Eclipse to the next) is shorter than 15 seconds.

In the end, it's probably a knot that's too big in scope for WrathCalcs to attempt to model, but it is something that we should probably consider for optimal performance during a fight.

#225 Hamlet

Hamlet

    Mike Tyson

  • • Guide Author
  • 11,540 posts

Posted 05 December 2010 - 08:27 AM

The big issue would be having to clip to take advantage of the haste. It's generally never been worth it to clip for trinket procs, except for edge cases like clipping exactly one tick, which seem too thin to worry about. That shouldn't be too different here--these trinkets are around as strong as past trinkets. In a bit more detail: a ~2000 haste proc will give at most 2 extra ticks. So same result as before--it's only going to be worth clipping if the clip cuts off exactly one tick (i.e. you refresh between the third-last and second-last tick). Too rare and minor a benefit to realistically think about.

It's a slightly different case when we talk about use: effects. We have , but the principle is no different for spellpower procs (and don't forget Synapse Springs). One some other thread I briefly got into this lately--it seems like it might be optimal to wait to overlap those things with Eclipse. Slightly different issue though.

#226 Guest_aceofsween_*

Guest_aceofsween_*
  • Guests

Posted 05 December 2010 - 08:41 AM

It's a slightly different case when we talk about use: effects. We have , but the principle is no different for spellpower procs (and don't forget Synapse Springs). One some other thread I briefly got into this lately--it seems like it might be optimal to wait to overlap those things with Eclipse. Slightly different issue though.


Generally speaking, wouldn't this issue only crop up for DoTs and in that case it only is important that the effect of an item like is active as you proc Eclipse and cast your next round of DoTs (therefore getting the effect of the item proc over the full duration of your DoT).

So ultimately, it would benefit only you to delay the use of the item up until the point where it's effect lasts until you reapply your DoTs, correct? Again it's something of a fringe effect that is going to be hard to judge, especially considering that all it would take is a few procs from Euphoria to throw off your rotation's timing.

#227 Hamlet

Hamlet

    Mike Tyson

  • • Guide Author
  • 11,540 posts

Posted 05 December 2010 - 08:49 AM

Mmm, right, that is true for haste effects. Timing them with DoT's and/or NG is good, but Eclipse doesn't matter. So maybe you'd want to macro the HoI use to IS.

Spellpower effects though, you want to overlap with Eclipse still.

#228 Guest_aceofsween_*

Guest_aceofsween_*
  • Guests

Posted 05 December 2010 - 09:01 AM

Right, because Haste only directly increases damage of DoTs. For everything else, it only increases the rate at which you do damage. This is different from Spellpower/Int, which increases directly increases the damage of all your spells, and therefore is subject to the fluctuations of percentage based buffs like Eclipse.

On the topic of specifically, here is what I had come up with:
The added damage from delaying the proc is going to have to outweigh the damage "lost" by not proccing the effect (which is a product of the reduction in uptime of the buff). A full stack would net you 1815 Haste, which is currently valued at right around 3085 DPS. The stacking Spellpower effect would total up to 435 spellpower, which nets you right around 1085. So in order for you to have a net-gain, you would have to pick up an extra 2000 damage for every second you wait to use the proc. And again, this damage is only going to come from your DoTs.

From what I've figured using WrathCalcs, the difference in damage between Insect Swarm and Sunfire (assuming a Solar Eclipse) with the proc and without it is ~26500 damage which would mean you could conceivably delay the proc by 13 seconds before it would be better off simply using the proc.

That's an incredibly large window honestly and it sparked another thought... Isn't the same also true for delaying an application of a DoT when the cooldown for the use effect is close to expiring?

Editted Note: Obviously, I'm using the DPS values for Haste and Spellpower from WrathCalcs, which are going to fluctuate slightly depending on your exact gear make up. The same is true for the damage output of Insect Swarm and Sunfire. Lower gear levels will make that "13 second" window smaller, because the DoTs will do less damage.

#229 copialinex

copialinex

    Von Kaiser

  • Members
  • 59 posts

Posted 05 December 2010 - 04:33 PM

Does anyone know if Euphoria works like the attack table or if it's merely a % proc when gaining energy? I know that when hitcapped this isn't a issue, since both the table and the proc would behave samely, but in the case you're not hitcapped, it could lead to slightly different results:
If you are 1% under hitcap:
If it behaves like a table, you have 1% chances of missing, 24% chances of proccing double energy and 75% chances proccing normal energy when casting a spell.
If it's a proc when gaining energy, you have 1% chances of missing, 23.76% chances of proccing double energy and 75.24% chances of proccing normal energy when casting a spell.

Yeah, I know the differences are trivial, but maybe at higher tiers (if haste > hit) this would be useful, so I wanted to share my thoughts.

#230 beromar

beromar

    Von Kaiser

  • Members
  • 29 posts

Posted 06 December 2010 - 04:29 AM

nvm delete

#231 MatsT

MatsT

    Don Flamenco

  • Members
  • 485 posts

Posted 06 December 2010 - 12:18 PM

Most if not all level 80 trinkets with proccs have a 45 second internal cooldown. This includes trinkets like Dislodged Foreign Object that has the same 20 second duration and 10% procc chance. If we assume anything at all it should be 45 seconds.

#232 Tecton

Tecton

    Soda Popinski

  •  Patrons
  • 3,946 posts

Posted 06 December 2010 - 12:24 PM

Rogues, at least, have found their level 85 trinkets to have a 75 second ICD, so I don't think 45 seconds is a safe assumption.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users