Moonkin patchnotes

[mercychan]

This is all very interesting but unless something changed and nothing was said the damage of a spell is calculated the moment you fire it off (when the cast is finished but before it lands) if this is still the case which im 99% certain it is then this whole experiment is flawed based on your belief that damage is calculated when a spell lands.

Might be when the cast finishes, but I ran this test on the target dummie's groin so the travel time of wrath was nonexistent. I'll edit that bit, thanks

- - - Updated - - -

Realize that point will move, depending on the amount of mastery you have.

I'm not sure how much that point would move, it might just be (and probably is) marginal. Come t18 when we have much different mastery values then sure this is worth revisiting. But for t17 progression it's fine to assume ~47-55% mastery to be standard practical values.

[Pippilongear]

I'm not sure how much that point would move, it might just be (and probably is) marginal. Come t18 when we have much different mastery values then sure this is worth revisiting. But for t17 progression it's fine to assume ~47% mastery to be a standard practical value.

I ran some calculations using all 665 ilvl gear (to approximate a mix of mostly Highmaul Heroic with a few crafted / CM drop / heroic dungeon type gear thrown in) and our raid buffed stats were roughly:
75% Mastery
14% Crit
12% Haste
7% Multistrike
4% Versatility

That changes the transition point a bit

[mythiix]

I ran some tests on the Beta to try to approximate at what points on the eclipse bar Wrath becomes stronger than Starfire and vice versa.

Stats
--------
11.66% Crit
9.43% Haste
47.08% Mastery
8.18% Multistrike (the data below excludes all multistrike damage)


I took the damage done by Starfire and Wrath and divided them by their cast times to find their effective DPS at several points on the eclipse bar. I recorded five values for each data point and threw the averages (in thousands) into excel and came up with this:



Which shows that:
1) Starfire remains stronger than Wrath until about 1/5th of the way into solar.
2) an empowered Starfire (with SotF) remains stronger than an unempowered Wrath 3/4th of the way into solar.
3) an empowered Wrath (with SotF) remains stronger than an unempowered Starfire 2/3s of the way into lunar.

And without SotF you can see these positions shift toward the center of the eclipse bar:


Tests were done without any of the proc enchants on my weapon for better precision. Keep in mind that damage is calculated when your spell cast finishes and not when you start casting it.

Not saying that you are implying this in your post but I know people will assume it... this doesn't represent everyone's transition point between wrath and starfire. A lot of it depends on the player's haste and mastery levels so people should take into account that this is only true at those percent ratings given the math behind it is correct.

Lappe linked a spreadsheet about a week ago which made this process a lot easier.

I don't know if it has been updated with this past weeks buffs but it is inevitable that it will be.

I asked him earlier how to use this spreadsheet in a PM and I will paste his response below.

"If you just want the transition point. Its easy, from the spreadsheet you set it to show when SP/Sec of Starfire and Wrath are equal (whichever way you're going)
Then you just check the Seconds since combat line. You'll have number there from which you can calculate the timer for your bar.
0seconds means the combat is starting right now.
10seconds is at lunar peak (the actual peak, the buff starts already at 8.03)
20seconds is at 0 point
30seconds is at solar peak (the actual peak, the buff starts already at 28.03)
40 is the same as 0, just reached a full cycle

Now lets say the transition point from Solar to Lunar is at -40 Lunar energy, this equals to 37.51 seconds in combat.
that means you've been 17.51 seconds on Solar side at that point
its 7.51 seconds after reaching the actual Solar peak
its 5.54 seconds after the eclipse bar starts moving back (in game visual)."

"Whole eclipse bar is done with sine wave, so you'd have to do the math with 105*SIN(Time*PI/Cycle-Timer) Time being time in combat, Cycle-Timer being the cycle time from 0 to lunar to 0 (20 w/o euphoria, 10 with it)
That'll give you the eclipse position at a certain time."

Hope this will be of use to others and thanks again to Lappe for his help.

[mercychan]

We're looking at the interaction between Starfire and Wrath, both of which scale appropriately with Mastery and Haste. You shouldn't expect Starfire to become dramatically stronger than Wrath as you gear up in t17, which is why I expect any shift to be marginal. When I get some time I'll relog on the beta and run a few more tests with different gearsets, but that'll take a few hours.

Edit: nevermind looks like beta ended today ;___;

[mythiix]

I completely agree when only looking at 1 tier but as you advance through WoD, there will be a larger difference between the transition point now vs. the transition point down the road 2 tiers later. The point of underlining it was to just state that it is not a fixed value and edge them towards doing their own calculations.

[mercychan]

I ran some calculations using all 665 ilvl gear (to approximate a mix of mostly Highmaul Heroic with a few crafted / CM drop / heroic dungeon type gear thrown in) and our raid buffed stats were roughly:
75% Mastery
14% Crit
12% Haste
7% Multistrike
4% Versatility

Could you show me the gear set that you used to get these stats?

[Pippilongear]

Could you show me the gear set that you used to get these stats?

I made a wowhead profile equipment set but I guess I wasn't logged in so it has expired :P I will see if I can recreate it in the next few days.

[huth]

The graphs are DPS(e) vs. Eclipse Power and this relationship is non-linear in practice. This is because when you cast a spell at an arbitrary point on the eclipse bar it can return a value (deal damage) within a range of values (the calculation of this value depends on your eclipse power, as well as other stats (which are held constant)).

No, it isn't. Eclipse Power to Damage is a linear function. A known linear function. You didn't actually have to do any field tests, you could just have mathed the whole thing out.

[mercychan]

No, it isn't. Eclipse Power to Damage is a linear function. A known linear function. You didn't actually have to do any field tests, you could just have mathed the whole thing out.

In practice, as in, on a boss fight it is non-linear. I'm not sure how to explain it better. Sit in front of a training dummy and cast a wrath out of combat and see how much damage it does at Eclipse Power 0. Then reset combat and do it again, and that second wrath will do a different amount of damage. Our FoN treants work the same way, they can deal damage within a set range of values. This randomness makes it non-linear, but if you were to attack infinite times at arbitrary points on the eclipse bar and calculate their DPS(e) vs. Eclipse Power, you can expect a linear graph, yes.

[Erdluf]

Testing is good. Sometimes the tooltips are just wrong (or the spell is wrong, either way they don't match). For instance, FoN tooltips are currently wrong by about a factor of three.

That is entirely different than saying "because I once saw SF as higher DPS than Wrath at some position on the bar, use SF at that position." That would be like saying "I cast Sunfire at 17 energy and it crit, so, you should always cast Sunfire at 17 energy for more damage". You should require a large sample before reaching that kind of conclusion.

[huth]

This randomness makes it non-linear, but if you were to attack infinite times at arbitrary points on the eclipse bar and calculate their DPS(e) vs. Eclipse Power, you can expect a linear graph, yes.

Which is exactly what you need to do for exactly this reason. The randomness is an outside factor that has nothing to do with what you're actually trying to calculate. There's a reason theorycrafters usually work with averages, not discrete values.

Your sample size is way to small. 5 is basically nothing with the amount of randomness involved. With your crit chance, chances are that one spell crit once and the other never did, completely messing up your results(effectively giving you 6 casts of spell a vs. 5 for spell b).
Not that crit chance even matters, since it affects both spells equally anyway. Only factors that do not apply to both spells, like empowerment, actually need to be tracked.

In fact, we actually only need the relative base DPET between Starfire and Wrath(IIRC, SF is 5% higher than Wrath) and the total Eclipse bonus to calculate when you should be switching.

Your switch points are d_SF * (1 + Eclipse_L(t)) * Emp_SF = d_Wrath * (1 + Eclipse_S(t)) * Emp_W, solve for t.

d_SF = Starfire base DPET
d_Wrath = Wrath DPET
Eclipse_L(t) = Lunar damage bonus at time t
Eclipse_S(t) = Solar damage bonus at time t
Emp_SF = 1.3 if you have Lunar Empowerment, 1 if you don't
Emp_W = as Emp_SF, except for Solar Empowerement. Caution: May vary due to Talents

Eclipse Power is a function of time.

[Owlkin]

How many targets we need in a worst case scenario (solar peak) that Starfall was better SS?

[Deleted]

depends on mastery/haste but as a general rule 3.

[Deleted]

I quite frequently have the "problem" that charges are generated faster than I can use them with Starfall, so I often need to use SS to burn off a charge (on fights like protectors etc)

[xtramuscle]

I quite frequently have the "problem" that charges are generated faster than I can use them with Starfall, so I often need to use SS to burn off a charge (on fights like protectors etc)

Yeah sucks to be you.. just apply the usual rule of trying to spend it towards the peak of eclipse and life is good.

[Deleted]

Less gear will solve this.