SMN target swap - how many GCDs to break even?

[Havenchild]

I think they were referring to the Hourglasses as, since there are 4, which are all separate, it would require 8 GCD in order to dot them all with B+M before having to return to the first one to dot again. B + M are your highest potency/s compared to Bio 2 so I guess it makes sense that you would only rotate those two over 4 separate adds. (changes in a good WoD party where you wouldn't have the option to return to the last one, since it's dead so I see where you are coming from in that regard)


T8 is my favorite example of a SMN trying to be bursty and losing a ton of resources, by Baning/Fester. Where as even Doting it up when it spawns (Facetarget) and Ruin is pretty efficient, and saves 2 Aetherflow for the boss, or 2 for the Dread if your party dps is slow.

-----------------------------------------------------

Every different pug I join for random FCOB farming I have to change my rotation. As annoying as it seems, it's probably my favorite part of SMN, with how versatile it is in so many different situations.

No, I gave a great example.
The hourglasses do last over 9 seconds and that's why multiple miasma/bio rotation works. The 9 seconds as indicated by Xero's graph (Amazing BTW, that really clarifies a lot!) is for single target where you have absolutely nothing better to do after you've set up your dots. In the situation of the hourglasses this is NOT the case, because you have three other hourglasses to dot up. If you put three dots on each hourglass, bio would be falling off the first one by the time you hit bio on just the third one. The idea there is you constantly rotate your most potent skills and drop the least potent one from the rotation. By the time you've finished miasma/bio on the last hourglass and return to the first, bio will be just falling off, so there is a moment where you had a single tick of 8 dots for 300 potency, and since you then refresh that rotation a further time, you maintain 300 potency worth of ticks until one of the three dies.

I feel like I better understood what you were trying to say after reading this tidbit. In a good party where the first add will die in 8 GCDs, the M / B setup is likely better.

However wouldn't this change if they are alive past 8 GCDs as well?
For example if you only dotted the 3 hourglasses not focused by the horde with all 3 DoTs (9 GCDs), you would be doing 330 potency universally in one tick by the time the final bio lands on the third add.

The first Bio would be at about 2 seconds or so. From here on you could technically ignore Bio II falling off on the first add (about 10 seconds or so left here, because your assuming it will die in those 10 seconds) and just refresh Miasma and Bio on the hourglasses, and overall that would give you more potency universally no? For about an additional 10 seconds or (3 to 4 server ticks depending on where you land on the server timer) you are doing 330 potency. Unless my math is completely wrong here.

Yep, but generally you have a good idea of how fast things are dropping.

You start casting Bio II or Miasma and by the time you finish, its already at 75% hp? Drop the second dot, immediately bio then Ruin II
You finish 1 dot and its close to 50%? Ruin + Ruin II.

This is more or less the way I make those calls as well. Although assuming we keep the same DPS as well, in repeat attempts yes, it would be Miasma over Bio II for that little bit of potency.

[Whiston]

You gave a bad example.


Which also leads to my next point of the number crunching in this thread in regards to using what. All of this changes in real time, depending on your fellow DPS no?

No, I gave a great example.
The hourglasses do last over 9 seconds and that's why multiple miasma/bio rotation works. The 9 seconds as indicated by Xero's graph (Amazing BTW, that really clarifies a lot!) is for single target where you have absolutely nothing better to do after you've set up your dots. In the situation of the hourglasses this is NOT the case, because you have three other hourglasses to dot up. If you put three dots on each hourglass, bio would be falling off the first one by the time you hit bio on just the third one. The idea there is you constantly rotate your most potent skills and drop the least potent one from the rotation. By the time you've finished miasma/bio on the last hourglass and return to the first, bio will be just falling off, so there is a moment where you had a single tick of 8 dots for 300 potency, and since you then refresh that rotation a further time, you maintain 300 potency worth of ticks until one of the three dies.

But yes, undoubtedly SMN is the class that really is affected by the rest of the party.
Honestly in most situations where an add exists for a brief amount of time (Bennu's for instance), it's not really worth a summoners time unless the main target is already dotted and, again, you really have nothing better to do but throw out a few ruins.

[Sleigh]

To anyone doing math, remember that the server tick can happen any time from the second you apply (it will even take the timer down a couple seconds on the initial duration to compensate!) to the full 3s later. This means something like Bio can be worth 80 potency within almost exactly 3s rather than a static 6s from the time you cast it, if lucky enough.

Probably best to assume a 1.5s average first tick from application when calculating any numbers.

[pandabearcat]

You just prorate the tick, so if its 0.5 seconds till GCD from the application you just say its 0.5/3*potency.

I'm a bit confused by what Xero put down but I understand the idea of it.

Ideally bane saves you 3 GCDs that are used to apply dots.

Applying dots during those 3 GCDs gives you what is approximately 35 (1 Bio2 tick) + 20 (miasma application) + 110*0.5 (roughly all dots applied at 6 sec, we stop at 7.5 sec) = 110 potency

3 Ruins instead (because of bane application) gives you 240 pot, and you're saying 410 > 240 for single target. Mana costs and ramp aside, this makes sense and for double target bane is never worth it.

[Judge_Xero]

I'm a bit confused by what Xero put down but I understand the idea of it.

Ideally bane saves you 3 GCDs that are used to apply dots.

Basically, I think what it comes down to is casting DOTs is stronger than casting Ruin. (80 Potency)
Bio - 240 Potency (@18s)
Miasma - 300 Potency (@24s)
Bio 2 - 350 Potency (@30s)

So while you are saving having to cast your Dots by using Bane, that only allows you to cast extra Ruins (3 x Ruin is only 240 Potency VS Fester 300 Potency). Overall the DOT windup is not that bad since Bio is instant Cast and becomes part of the dot ticks before the GCD is over. Even at the full 2.5s GCD you can have all your DOTs running before the second Dot tick (obviously dependant on when the server tick happens).

Hitting 2+ Mobs with Bane is obviously better.

I'm mainly talking about the situations where a single add spawns. You will often see SMN's that will Bane to that mob and then blast it with Fester. Sure it looks cool, but you just lost like 600 Potency worth of damage, to do 300 Potency.

(300+ from Dots ticking, 300 Potency for using Bane which could have been a fester)

You could have easily cast Bio and Ruin II spammed while auto attacking, for more damage in only a few extra seconds + had 2 Aetherflow stacks to Fester something.

You just prorate the tick, so if its 0.5 seconds till GCD from the application you just say its 0.5/3*potency.

This I'm confused about. You get the full potency of a dot tick whether you cast it 2.5s or 0.5s before the server tick happens. There is no prorating.

[SunnyHirose]

This I'm confused about. You get the full potency of a dot tick whether you cast it 2.5s or 0.5s before the server tick happens. There is no prorating.

In theorycrafting a DoT with some sort of time cutoff (e.g. clipping the DoT or considering a burst period), it is often a useful abstraction to consider its damage probabilistically, as if the timing of your cast and subsequent cutoff is so unrelated to the tick timing that you have an equal chance of landing it anywhere inside the 3s window.

[Judge_Xero]

In theorycrafting a DoT with some sort of time cutoff (e.g. clipping the DoT or considering a burst period), it is often a useful abstraction to consider its damage probabilistically, as if the timing of your cast and subsequent cutoff is so unrelated to the tick timing that you have an equal chance of landing it anywhere inside the 3s window.

Oh interesting. But how do you chose this value? Does it matter just as long as it's a value between 0.1-3? Since there is no way of actually knowing where the tick will land between casts?

How does that relate to instant cast dots vs others?

Probably best to assume a 1.5s average first tick from application when calculating any numbers

Is that what Sleigh meant here? Best to just cut it right down the middle?

Xero you forgot that with bane you get the dot ticks going while you ruin.

Fester - 3 GCD (B2, M, B (Instant) and the mob can't die prior to 15s (475 Potency(Dots) + 260 (Ruin X 3 + Miasma Cast) + Fester) (1035 Potency)

Bane - (550 Dots @ 5 ticks for 15s) (480 Ruin @ 6 casts) (1030 Potency)

-------------------------------------------------------------

I did consider the dot ticks going while you Ruin. Since you can have all 3 Dots hard cast before 6 seconds of that second Dot tick, theoretically you would only lose out on the 75 Potency from not having Miasma (35) and Bio (40) up for that initial tick.

Also consider that you do not get the 20 Potency from hard casting Miasma when you Bane.

So here is what you are left with in the end.

3 x more ruin + 75 Potency from that initial tick (315 Potency) VS 20 Potency from a hard cast Miasma + Fester (320 Potency)

Also, if the mob lives past the life of Bio2 and Miasma then you are losing the extra damage it took you to set up the dots you baned from.

~2.5 GCD (slight wait to make sure Bio registers before you Bane) ~ 3 seconds or 35 Potency from Bio 2.

----------------------------------------------------------------------

Edited: So considering probability......

Should the first Bio2 Cast be 1.5/3*Potency (17.5) - where each other Dot cast after that has 100% probability to fall within the server tick timing?

I would like to know how to accurately graph the optimal spell usages to use within each estimated time frame for mob life. I did a graph a while back, but it was based off the server tick always starting @ 1 (and it was a huge visual mess by the time I was done)

[pandabearcat]

I actually redid the math, and with 3 GCDs with bane up you get 2.5 ticks, which is 275 potency.

Adding it to the triple ruin damage, bane will beat out fester even at 1 target, at 515 vs 410 potency. However this ONLY works if you bane a full rack of DoTs.

Xero you forgot that with bane you get the dot ticks going while you ruin.

[SunnyHirose]

You can break timings for spells down to a few factors:

- Cast time: instant or percentage of GCD
- Activation of the skill: For instant skills, the moment it goes on cooldown. For casts, the moment the cast ends.
- Animation delay: The time after activation for the effect to actually happen. For many skills this is just plain damage. For DoTs, buffs, and debuffs, this is also when it is displayed on the status bars (ignore the timing of any floating text, that's just aesthetic). I never took that close a look at ACN, but on MNK, for example, some animations have noticeably different animation times.
- Effect duration: Best I can tell, when there is an effect, you get the full listed duration of an effect from the moment the animation ends and the effect begins.

Therefore, if you let Bio tick all the way out, you get its full potency. If you let it tick most of the way out but you clip it at 0.03 seconds before the effect ends, then there is a very small chance you will lose a tick... 99% chance you won't, 1% chance you will. Over a bunch of attempts, that 17.97s Bio is 240 - (.01*40) = 239.6 potency (or 240*.99 + 200*.01). If you were to somehow clip it at 3.03 seconds beforehand (d'oh!), you could say it's 199.6 potency because 1 tick is always lost, and 1 tick has a 1% chance of being lost.

Instant or casted doesn't matter if you only consider time till the next use; the animation is supposed to be the same (I believe, and latency does affect this a little), so the timings of the effects will be that number of GCDs apart. If you're running it against a timer of some sort, however, I think you'd need to factor in the animations somehow. Bio is an instant cast, for example, but the animation is pretty slow. Like, it wouldn't be accurate to end a test with Bio + Ruin and conclude that you got 2 GCDs of Bio ticks of that last one; it was more like 1.4 GCDs or so.

[pandabearcat]

I think we were talking about two different scenarios.

If you have 2 long term targets that will live for the entire duration and longer of a full rack of dots, then fester is technically a dps gain because the dots from 2 additional ticks is not as much as miasma cast + fester.

What I was talking about is if you get an add to bane to that you need to kill - at what lifetime of the add will bane be higher dps, and thats the 9/6/4 seconds that I came up with.

However, for long term two target, I would still recommend bane, because it saves you hundreds of mana which is pretty significant. If mana is not an issue, or your dots aren't lining up for reapplication (as is often the issue in midfight banes) then it is worthwhile to instead use fester.