Moonkin Talents/Specs Discussion - Page 3

• Adoriele · 11/20/08, 3:10 PM

Originally Posted by Khalanis

so people are going to have ~2340 int while raiding at 80???

Um. What? The spirit regen formula is well known. Using 637 Spirit and 888 int, adding in raid buffs, you get $.1*(.001+845.9*\sqrt{1100}*.009327)*5 = 78$ MP5 per point of Intensity. I honestly have no idea where you're getting your number.

Khalanis · 11/20/08, 3:35 PM

Originally Posted by Adoriele

Um. What? The spirit regen formula is well known. Using 637 Spirit and 888 int, adding in raid buffs, you get $.1*(.001+845.9*\sqrt{1100}*.009327)*5 = 78$ MP5 per point of Intensity. I honestly have no idea where you're getting your number.

you replied to that in less than 30seconds(the time it took to realize you said intensity not dream state and edit my post), nice =D

Erdluf · 11/20/08, 4:41 PM

Paper-napkin Splendor math

MF+SF rotation, MF usually runs out in the middle of an SF cast. With SF glyph you'd refresh MF about every 22s (no Splendor) or every 25s (Splendor). Over ten minutes, at 30% Haste

Total cooldowns = 600/1.5 * 1.3 = 520
MF casts (No Splendor) = 600/22 = 27.3
MF casts (Splendor) = 600/25 = 24

You save 3.3 GC's out of 520, for 0.6% increase to your SF damage. You also get a 0.6% boost to your MF damage (less time wasted with no dot ticking).

If your rotation has IS, Splendor means you refresh every 15s instead of every 13s.
IS casts (No Splendor) = 600/13 = 46.2
IS casts (Splendor) = 600/15 = 40

You save 6.2 GC's out of 520 for a 1.2% boost to your SF damage. Less IS down-time means about a 1.1% increase to IS damage.

In both cases your DPM increases (more damage per cast means more DPM). Less haste favors Splendor, more haste hurts it.

If you have IS in your rotation (for DPS or for debuff), Splendor looks like a good talent (better than 1% haste). If you don't use IS, (or only use it once every 30s, before casting Wrath for an Eclipse proc), Splendor looks weak.

This is for continuous casting. For fights where you have to stop casting, but your dots continue ticking, (First boss in Mana Tombs has spell reflection, ...) Splendor starts looking much better.

Edit: Standing in MK form, 42 haste from gear (2.30% @72), Hurricane tooltip reads 9.21s. 3/3 CF, 3/3 IMkF
10/1.03/1.03/1.023 = 9.214
10/(1+.03+.03+.023) = 9.234

Grawlen · 11/21/08, 3:52 AM

Has it been determined conclusively that iFF's crit bonus does in fact work with other people's faerie fires, including faerie fire (feral), or was that debunked?

If the crit bonus does work in that scenario, is the talent worth taking when you can count on another faerie fire being up?

klüger · 11/21/08, 6:51 AM

Originally Posted by Grawlen

Has it been determined conclusively that iFF's crit bonus does in fact work with other people's faerie fires, including faerie fire (feral), or was that debunked?

If the crit bonus does work in that scenario, is the talent worth taking when you can count on another faerie fire being up?

I think some scenarios to facilitate testing were set up (X resillience, 2x druids (one with imp ff one w/o) low critrate etc) but the actual tests were never performed the minimum amount of times to make anything conclusive. However I'd love to be wrong on this one.

• Adoriele · 11/21/08, 12:28 PM

Originally Posted by Erdluf

Edit: Standing in MK form, 42 haste from gear (2.30% @72), Hurricane tooltip reads 9.21s. 3/3 CF, 3/3 IMkF
10/1.03/1.03/1.023 = 9.214
10/(1+.03+.03+.023) = 9.234

Wow, that's fantastic! I'll do a double-check when I get a chance to confirm, and then update all of the modeling I have access to.

Lord BEEF · 11/21/08, 12:30 PM

That means that the talents are multiplicative with haste from gear, is that right?

Kuruk · 11/21/08, 12:39 PM

Originally Posted by klüger

I think some scenarios to facilitate testing were set up (X resillience, 2x druids (one with imp ff one w/o) low critrate etc) but the actual tests were never performed the minimum amount of times to make anything conclusive. However I'd love to be wrong on this one.

That is correct. Some testing has been done but the sample was far too small to be conclusive. Therefore, we are still in the dark about this one.

Thanks a lot of the Nature's Splendor calculations Erdluf.

So if we can agree that Celestial Focus is in fact multiplicative, does it make it a stronger talent than IIS and NS (assuming you are keeping both IS and MF up)?

• Adoriele · 11/21/08, 12:43 PM

Originally Posted by Lord BEEF

That means that the talents are multiplicative with haste from gear, is that right?

Looks like it, yes. It's a large benefit, and makes me wonder if Wrath of Air is the same way.

Originally Posted by Kuruk

That is correct. Some testing has been done but the sample was far too small to be conclusive. Therefore, we are still in the dark about this one.

Thanks a lot of the Nature's Splendor calculations Erdluf.

So if we can agree that Celestial Focus is in fact multiplicative, does it make it a stronger talent than IIS and NS (assuming you are keeping both IS and MF up).

It's possible. I'd need to poke Dedmon or download the SimCraft source and rebuild it with the changes, or Dope with Rawr. I'm not at the point where I could easily write in the necessary changes to WrathCalcs to show the difference yet. But it's a bit of a boost to both talents.

erragal · 11/21/08, 12:50 PM

Originally Posted by Adoriele

Looks like it, yes. It's a large benefit, and makes me wonder if Wrath of Air is the same way.

It's possible. I'd need to poke Dedmon or download the SimCraft source and rebuild it with the changes, or Dope with Rawr. I'm not at the point where I could easily write in the necessary changes to WrathCalcs to show the difference yet. But it's a bit of a boost to both talents.

Isn't the more significant change with this information our valuation of haste on gear? If it's multiplicative with our existing haste rating this really widens the gap between crit and haste as gear statistics (And changes the break-even point of Spell Power/haste).

Khalanis · 11/21/08, 12:53 PM

Originally Posted by Adoriele

3496 Base Mana. Moonfire costs 734 base, 23 less per point in Moonglow. Casting it once per 15s (I'll assume you have either 2T6 or Splendor) grants you a whopping ~8 MP5. If you've got the SF Glyph, it's about 5 MP5 (23 mana per 24s). Assuming no Eclipse (MF, SFxN), you'll cast ~9.5 Starfires over those 24s, at 17 mana per point, for ~33.5 MP5, a total of less than 40 MP5 per point. By contrast, Intensity is worth (raid buffed) about 78 MP5 per point at T7-25 gear levels. OoC is worth ~230, and DS is worth ~44 for the first point, less for the rest. So yes, Moonglow is definitely the lowest talent on the totem pole.

And no, Moonglow does not scale with haste. People have made the same claim about OoC, and it's not true. If you cast the exact same sequence with Haste and without, ignoring passive regen, you will cast the exact same number of times before running out of mana. Omen and Moonglow do not scale with haste. The other mana talents just scale negatively.

My mistake for not realizing that 80s would be getting 1100 int raid buffed. Factoring in that info I can see that moonglow isn't as good as dreamstate at those numbers.
I would like to refute your point about non-scaling. Given a set amount of time you will gain more mana from moonglow the more haste you have. you will spend more mana yes, but it doesn't change the fact that moonglow will provide a larger # bonus to your total mana available during that time. During the same amount of time the other talents will give the same regen regardless of haste. They don't negatively scale, they just don't scale.

Originally Posted by Adoriele

Um. What? The spirit regen formula is well known. Using 637 Spirit and 888 int, adding in raid buffs, you get $.1*(.001+845.9*\sqrt{1100}*.009327)*5 = 78$ MP5 per point of Intensity. I honestly have no idea where you're getting your number.

Even though this was in response to my error I feel I should inform you that your numbers are out of date.(unless these changes have been reverted)

WotLK makes yet another sweeping overhaul to the regeneration model. In 2.4, we all became used to the regen formula of:

5 * 0.0093271 * Spirit * Square_root ( Intellect )

which made spirit very valuable, and intellect slightly so. Not only was the coefficient lowered as we leveled to 80 (as expected), but Blizzard made the decision to lower the coefficient yet further in an attempt to remove our ability to 'regen infinitely'. Accordingly, the formula at 80 is:

5 * 0.005575 * Spirit * Square_root ( Intellect )

regen calculation thread where I quote from

Base Mana / Crit Rate
And this was the base mana #s I was using, which could themselves also be outdated(even though it's only about a month old >_>)

dedmonwakeen · 11/21/08, 1:09 PM

Originally Posted by Adoriele

Looks like it, yes. It's a large benefit, and makes me wonder if Wrath of Air is the same way.

It's possible. I'd need to poke Dedmon or download the SimCraft source and rebuild it with the changes, or Dope with Rawr. I'm not at the point where I could easily write in the necessary changes to WrathCalcs to show the difference yet. But it's a bit of a boost to both talents.

The code currently treats multiple forms of haste multiplicatively:

double druid_spell_t::haste()
{
  druid_t* p = player -> cast_druid();
  double h = spell_t::haste();
  if( p -> talents.celestial_focus ) h *= 1.0 / ( 1.0 + p -> talents.celestial_focus * 0.01 );
  return h;
}

double spell_t::haste()
{
  player_t* p = player;
  double h = p -> haste;

  if( p -> type != PLAYER_GUARDIAN )
  {
    if(      p -> buffs.bloodlust      ) h *= 1.0 / ( 1.0 + 0.30 );
    else if( p -> buffs.power_infusion ) h *= 1.0 / ( 1.0 + 0.20 );

    if( p -> buffs.swift_retribution     ||
	p -> buffs.improved_moonkin_aura ) 
    {
      h *= 1.0 / ( 1.0 + 0.03 );
    }

    if( p -> buffs.wrath_of_air ) 
    {
      h *= 1.0 / ( 1.0 + 0.05 );
    }
  }

  return h;
}

EDIT: While looking at this code I noticed that I had the IMKA at only 2%. Not a huge deal, since in all my runs it was being overridden by the Pally aura...... but if you use the tar.gz to look at simulationcraft code, it won't look exactly like what is written above.

EDIT2: The unwritten implication is that the SVN code repository is generally the best place to get the code since it is always up to date.

thedopefishlives · 11/21/08, 1:20 PM

Originally Posted by Adoriele

Looks like it, yes. It's a large benefit, and makes me wonder if Wrath of Air is the same way.

It's possible. I'd need to poke Dedmon or download the SimCraft source and rebuild it with the changes, or Dope with Rawr. I'm not at the point where I could easily write in the necessary changes to WrathCalcs to show the difference yet. But it's a bit of a boost to both talents.

The change should be easy enough to effect, I hope. Nevertheless, that's huge.

• Adoriele · 11/21/08, 1:35 PM

Originally Posted by Khalanis

My mistake for not realizing that 80s would be getting 1100 int raid buffed. Factoring in that info I can see that moonglow isn't as good as dreamstate at those numbers.
I would like to refute your point about non-scaling. Given a set amount of time you will gain more mana from moonglow the more haste you have. you will spend more mana yes, but it doesn't change the fact that moonglow will provide a larger # bonus to your total mana available during that time. During the same amount of time the other talents will give the same regen regardless of haste. They don't negatively scale, they just don't scale.

Even though this was in response to my error I feel I should inform you that your numbers are out of date.(unless these changes have been reverted)

regen calculation thread where I quote from

Base Mana / Crit Rate
And this was the base mana #s I was using, which could themselves also be outdated(even though it's only about a month old >_>)

For the formula, that's an error on my part, I grabbed the level 70 coefficient, but the 78MP5 number uses the 80 coefficient (I support both in my sheet, it was set to 80 but I grabbed the wrong coefficient when writing the formula). If you calc out the formula I wrote, you end up with ~130 MP5, which is what you would get at level 70 with those int and spi numbers.

Base mana numbers I grabbed in-game, though I'm not sure exactly what build, but they were externally confirmed on beta. I'll do a double-check on them once I hit 80, but I'm comfortable with them for now.

As for scaling, the only thing that matters in the end is the total damage you do. DPS is nice to worry about, but given that fights are not all of the same length, it's only secondary to the main goal of killing bosses. If you ignore passive regen, you will do the exact same amount of damage to a boss no matter what amount of haste you have. Similarly, you get the same amount of mana back from Omen. Where haste gains importance is when you add time-based measures to things. Bosses with enrage timers, soloing (haste is immensely useful here as an increase in XP/hr), etc. If we never had to worry about how long a boss fight took (HI2U Molten Core!), haste would be a useless, e-peen stat (not only is your 'DPS' higher, but you can fit in more casts before a boss dies, giving you a higher % damage done).

Let's take two example raids. We've got a boss with no enrage timer (Or, like Supremus, a very easy one), so each raid can take as long as they need. Raid A has 0 haste on its DPS's gear. None. They don't throw down Wrath of Air or Windfury, nobody specs into any haste talents, etc. They kill the boss in 3 minutes on the dot. Raid B comes in with 30% haste total on every single person's gear. Exactly that much, because they've been gearing up for weeks for this test case. They come in and kill the boss in 2'18.5". Exactly 30% faster. But every single person in the raid did the exact same amount of damage as their counterpart in raid A. Raid B's Moonkin got exactly the same number of OoC procs as raid A's Moonkin. On the other hand, though, Raid A's Moonkin got more mana back from Intensity and Dreamstate. This is why I say that OoC has neutral scaling with Haste, and passive regen effects have positive scaling. Because you will always get the same results (ignoring our RNG overlords) regardless of your haste.

Now, I'm aware that part of this argument is disingenuous. DoTs don't scale with haste, so Raid B will actually kill the Boss slightly after 2'18.5" due to the DoT portion not being affected, which means more nuke spells will be cast, which may grant a couple of extra OoC procs at the tail end of the fight. But if Moonkin A was OoM at the time the Boss died then, again ignoring passive regen, so would Moonkin B, except for the DoT effect.

Khalanis · 11/21/08, 3:27 PM

Yes, the logic is solid given your examples. There is a difference however, being that the raid haste isn't going up, just yours. The only way to increase the amount of spells cast during a boss fight is to make it last longer or increase your haste, thus increasing your personal % contribution to raid dmg. Because as a single unit out of 25+ your % dmg on a raid boss won't be 100% so you're not dmg capped as you would be if you were the only one doing damage. Note that this isn't necessarily an overall mana efficiency increase, just that the % of mana regen from moonglow vs other sources increases. At this point I suppose it doesn't matter seeing as we're only arguing semantics about the least valuable mana talent anyways =P(now that I've seen the insane int values we'll be getting @_@)