1

Closed

[Tree] mana numbers and ET

description

Basic stats page (v2.2.12) is showing
 
Int 1008
Spirit 817
MP5 531: Mouseover shows 772 O5SR.
 
Gaining 772-531 = 241 when going O5SR implies 2241=482 Spirit regen (factor of two because Intensity already gives me 50% spirit regen).
 
I believe the formula for spirit regen is 5
(.001+sqrt(int)spirit.003345) = 434.
 
So I'm seeing a discrepancy of about 10% on O5SR regen?
 
HT Heal: Shows 12229 (SP is 2644). I think that would be consistent with ET being multiplicative, but http://elitistjerks.com/f73/t37038-restoration_glyphs/p8/#post1240879 showed ET as adding .4 to the HT coefficient, not multiplying by 1.4.
 
Nourish: I don't have test numbers for these (I'll try to get them):
 
1) I'd assume that if ET is additive for HT, it would also be additive for Nourish. Rawr is using multiplicative.
2) I think I've heard the Glyph of Nourish bonus (and the tier bonus) are both multiplicative, so with three HoTs and both bonuses you'd have xxx1.181.15. Rawr has these as additive.
3) I'd assumed that the 20% bonus for having any HoT present was multiplicative. Rawr has it additive with GoN. I don't recall any seeing any tests one way or the other.

file attachments

Closed Sep 20, 2009 at 2:16 AM by Jothay

comments

wildebees wrote Aug 7, 2009 at 9:51 PM

Thanks for the feedback. I'll look at this over the next day or so.
Some feedback so long:
O5SR: 5(.001+sqrt(int)spirit*.003345) is being used in the code. (In code common to numerous models). The reason why the numbers aren't mathcing up is that replenishment is being added to both the MP5 and O5SR numbers, if I read the quick code check correctly, but I'll play with the numbers and try to see if I can show the more clearly. (Also still using the old innervate, so that tooltip needs to change as well).

ET: That part of the code comes from before I started working on it. No complaints yet, therefor I haven't messed with it. I copied that logic to Nourish for the 3.2 mod. I'll go check the EJ post. Thanks

Nourish: I have 4 T7, so will go try testing that. Not using Nourish glyph atm, but will see.

Erdluf wrote Aug 8, 2009 at 7:45 PM

I looked at numbers from a Black Knight attempt last night. I didn't record my SP, but Rawr says my toon should have had 2213 (IMotW, SP food, Dalaran Intellect, no pally/shami/priest buffs). If I back-calculate from Rejuv ticks (1913,1914) SP should have been 2210.

Using that I saw Regrowth ticks (1043,1044) about 1.2% less than rawr formulas predict. Regrowth direct was in predicted range.

Lifebloom ticks were in range, but blooms (2586-2587, or 7759 for a three-stack) were about 0.5% less than predicted.

Tranquility seems to have ticked (5503,5504) for about 2% more than I would have predicted from the wowwiki coef of 53.8%/tick (I assumed ER, GoN and Genesis all applied to Tranq).

Nourish: ET needs to be additive to base coefficient. 20% bonus for hots needs to be multiplicative.
6/12/18/24% bonus from Glyph cannot be additive with the 20% for HoT. It might, or might not be additive with GoN (results weren't conclusive). The most telling result was a Nourish that landed for 6903 with three Hots present. That is over 300 above what Rawr predicts.

I really just need to run some tests on myself, ungrouped.

Erdluf wrote Aug 12, 2009 at 5:24 PM

Ignore my numbers from the Black Knight post. Looking in Armory, I seem to have taken Starlight Wrath instead of Genesis (old Moonkin habit, I guess). My computations assumed Genesis, so they will be wrong.

wildebees wrote Aug 15, 2009 at 5:04 PM

Sorry been a busy week and only got to changes now. Just checked in patch 36098:
-Changed Innervates for 3.2 mode (3 min cd, half size). Number of Innervates calculated based on fight duration.
-Changed Empowered Touch Talent to be additive to both HT and Nourish
-Changed MP5 breakdown on main stats display to give a clearer breakdown

I didn't change anything for the different HoT bonusses for nourish yet. Any further testing?

Current formula is:
Nourish_noHoT = (Base + Sp * 0.6611 + ET) * (1+GoN) * (1 + MS) * (1 + ToL)
Nourish_HoT = Nourish_noHoT * (1.2 + (4T7 +Glyph) * #HoTs)

Should this be
Nourish_HoT = Nourish_noHoT * 1.2 * (1 + 4T7 +Glyph) * #HoTs) ?

Let me know when you get more testing data.

wildebees wrote Aug 22, 2009 at 9:43 PM

Did some testing just now.
Spellpower was 1919, was out of tree form, don't have ET, have GoN and 4T7, don't have nourish glyph.
Did some tests using different number of hots on myself and recorded lowest and highest numbers.
With regrowth, rejiv and lifebloom, I got a 4762 and a 7134 crit.

If we assumed Nourish_HoT = (Base+Sp0.6611) (1+GoN) * 1.2 * (1 + (4T7 * #HoTs) ), these 2 numbers fall just outside the lower bound (4784 and 7176). Thus this version of the formula doesn't fit the data.

All the numbers fell within the ranges of the current Rawr formula, but too few tests to be conclusive.

Erdluf wrote Sep 15, 2009 at 3:46 AM

I finally got around to testing. Results are attached. Unfortunately these are hand-logged, and two of the entries are suspect.

128 casts, split between balance+0, balance+1, resto+0, resto+3(glyphed).

I get a very good match for:
ET is additive.
It appears glyph is additive with the 20% bonus for having a hot. Their sum is multiplicative with the various talents.

Glyphed + 3 Nourishes uses: GoNMSSToL*(1 + .2 (nourish with hot) + .18 (glyph))

Erdluf wrote Sep 15, 2009 at 12:12 PM

Note: My Nourish test numbers from last night look like they match the current Rawr code (Glyph is additive with 20% for HoT). No need for any further change.

Erdluf wrote Sep 15, 2009 at 1:08 PM

WTB: A way to edit comments here.

Nourish coef in Rawr:
coefDH = 0.6611f; // Spreadsheet says .69, wowwiki says .6611, 1.5/3.5 = .43, confused!

Wowwiki now says .673
My attached spreadsheet puts lower and upper bounds between .6705 and .6726

I think a likely value is
1.5/3.5 (cast time) * 1.88 (BC +heal to LK +SP conversion) / 1.2 (bonus when HoT is on target) = .671429

I plan to update wowwiki right after I do this post.

wildebees wrote Sep 19, 2009 at 9:21 PM

Changed coefDH to .671429 in patch 36874.