Shadowpriest FAQ for 4.0 - Page 8

Ninahagen · 10/27/10, 8:56 PM

hit/spirit (below 17% hit) > intell > haste > crit > mastery > hit/spirit (above 17% hit).

For Hade$, I tested the 50% haste level, where VT has 8 ticks, it's not 1269 but 1268 (shadowform, darkness).
I don't know if it could be 1267, but it's not 1269.

Niimue · 10/27/10, 10:04 PM

Originally Posted by Nysem

Vampiric Touch - 5 ticks base

Ticks	Haste	Rating
6	10%	57
7	30%	663
8	50%	1269
9	70%	1876

I can confirm that at exactly 1269 haste (38.70%) + Shadowform and 3/3 Darkness, 8 ticks of VT were present.

roug · 10/27/10, 11:24 PM

deleted

nadoo · 10/27/10, 11:26 PM

@Ninahagen
You've taken a portion of my post out of context and have not even address the same issue that you quoted me on. I was talking about dps from MB alone, as I stated. I think it would be more constructive to read the entire post and comment on the concepts I was trying to outline. I was making a point leading into the last two questions of my post which are the most important part and I believe outline the most logical progression on the debate between MB and MF. I will try to get some number and do some maths but I'm sort on time at the moment which is why I posted the questions (hoping someone else can help out).

Hade$ · 10/28/10, 3:40 AM

Originally Posted by Ninahagen

hit/spirit (below 17% hit) > intell > haste > crit > mastery > hit/spirit (above 17% hit).

For Hade$, I tested the 50% haste level, where VT has 8 ticks, it's not 1269 but 1268 (shadowform, darkness).
I don't know if it could be 1267, but it's not 1269.

Thank you it's noted.

1268 should lead to 49.97% total haste, not 50%.
It's the same thing with 819 haste rating and the 0/0/0 spec.
The game should round some calulations but I don't know which ones.

@Niimue : The fact that the 8th is present at 1269 is not a guarantee of it being the exact point that give us this 8th tic. That doesn't give us the exact threshold. All the bold numbers tested were tested with their (value - 1) haste rating on my stuff in order to know the exact threshold. Ninahagen told us that 1268 already give the 8th tic. So, we can assume that the 1269 given by Nysem and me is wrong as a threshold. The difference is not that big but it could make a difference in the stuff optimisation.

I'll continue to try improving my formula in order to make it correspond to the reality.

Ninahagen · 10/28/10, 6:13 AM

I tested it at that time.

For Nadoo, I'm sorry if I offended you, but still I think there's something very wrong with what you said.
I've just reread it with much care, and it still bugs me.

It's true you have to know the real dpct between both spell (flay and MB without orbs) accounting orbs generation and shadowfiel's cd reduction. But, and that's what makes me think your post is very misleading, if you conclude MB = MF in term of dps (without orbs), then both strategies are roughly equivalent, and you do gain dps from mind blast when spamming it systematically on CD (and losing dps from Flay, but it's equivalent).

I'm especially basing on

also taking into account casting MB before 3 orbs is a dps lose for the MB cast

Because that's untrue.

It's a DPC (or DPCT, same thing anyway in that case) loss, but not a dps loss (since you cast a lower DPCT Mind Blast more often).

But I agree the most important portion of your message resides in the two last questions, that I would reformulate :
- Is MF or MB the better DPCT spell, knowing MF procs orbs and reduces Shadowfiel's CD.
- If MF > MB, is it better to wait for 2 orbs, or 3 orbs (and risk useless orb's proc) ? AKA By how much is Mind flay better than Mind Blast.
(if MB > MF just spam MB).

Nysem · 10/28/10, 1:22 PM

Originally Posted by Hade$

Thank you it's noted.

1268 should lead to 49.97% total haste, not 50%.
It's the same thing with 819 haste rating and the 0/0/0 spec.
The game should round some calulations but I don't know which ones.

@Niimue : The fact that the 8th is present at 1269 is not a guarantee of it being the exact point that give us this 8th tic. That doesn't give us the exact threshold. All the bold numbers tested were tested with their (value - 1) haste rating on my stuff in order to know the exact threshold. Ninahagen told us that 1268 already give the 8th tic. So, we can assume that the 1269 given by Nysem and me is wrong as a threshold. The difference is not that big but it could make a difference in the stuff optimisation.

I'll continue to try improving my formula in order to make it correspond to the reality.

The formula that was used to obtain them has given accurate percentages for all spells so far, but I'm not 100% sure how the game rounds its numbers internally. That's why I made sure to point out that my ratings columns are approximate, meaning they could be off by 1 haste rating (Which seems to be the case here). I don't want to say there is an issue with the formula quite yet, I suspect the way the game is rounding numbers is to blame.

I'll try and verify my expected ratings more precisely ingame and add them to this post later.

Edit - I did find something strange when I verified the breakoff for VT's 7th tick. The expected breakoff to get the 7th tick with haste buffs is 663 haste rating. I didn't have 7 ticks at 662 haste rating, even though that would be 29.98% haste. The game didn't round like it did for you at 49.97% total haste. I had to have the expected 663 (30.01% total haste) rating for that next tick.

Edit2 - For VT's sixth tick, it happened at 56 haste rating (9.997% total haste) instead of 57.

Niimue · 10/28/10, 7:04 PM

Originally Posted by Hade$

Thank you it's noted.

1268 should lead to 49.97% total haste, not 50%.
It's the same thing with 819 haste rating and the 0/0/0 spec.
The game should round some calulations but I don't know which ones.

@Niimue : The fact that the 8th is present at 1269 is not a guarantee of it being the exact point that give us this 8th tic. That doesn't give us the exact threshold. All the bold numbers tested were tested with their (value - 1) haste rating on my stuff in order to know the exact threshold. Ninahagen told us that 1268 already give the 8th tic. So, we can assume that the 1269 given by Nysem and me is wrong as a threshold. The difference is not that big but it could make a difference in the stuff optimisation.

I'll continue to try improving my formula in order to make it correspond to the reality.

It was not clear to me what was tested and what was calculated.

The following are TESTED values for VT with 3/3 Darkness and 5% Shadowform Buff:
1267 = 7 ticks
1268 = 8 ticks
1269 = 8 ticks

EDIT: Actually, as I look through posts earlier on, I see that Naphomci found the same result. So consider this a confirmation of his findings!

Paganchika · 10/28/10, 9:23 PM

And as far as rng is concerned insofar as stacking orbs, casting MB on 2 orbs or on MB's cd I've noticed some happy accidents with getting a 3rd orb during MB's cast.

nadoo · 10/28/10, 11:29 PM

Originally Posted by Ninahagen

But I agree the most important portion of your message resides in the two last questions, that I would reformulate :
- Is MF or MB the better DPCT spell, knowing MF procs orbs and reduces Shadowfiel's CD.
- If MF > MB, is it better to wait for 2 orbs, or 3 orbs (and risk useless orb's proc) ? AKA By how much is Mind flay better than Mind Blast.
(if MB > MF just spam MB).

Your still missing the point. If MB with 2 orbs is > MF that does not necessarily mean you get max dps from casting at 2 orbs. If the difference between MF and MB 2 orbs is less that the difference between MB 2 orbs and MB 3 orbs then you should wait for 3 orbs. This is something I took into account with the wording of my previous questions.

Also why are you talking about damage per cast? All you've done is take away the time factor which prevents you accurately comparing abilities with different cast times. You certainly can not compare MF and MB in any meaningful way using dpc as you've suggested in the quote above.
Could you also mathimatically show how MB with 2 orbs with not less dps per MB cast than MB with 3 orbs? Basically I'm quite sure my original statement is correct because I can't find error in the maths so if you could disprove it I might understand what you mean.

I have done some maths which suggests that casting MB at 2 orbs is better for maximum dps but its way too crude to actually conclude that, which is why I won't post it yet (may just be beyond my maths abilities anyway).

EDIT:
MF vs MB

My SF dps is 2017 and it has a 15 sec duration so:
2017*15= 30255 damage per cast
30255/1.5 = 20170 dps
MF crits reduce the SF CD by 10 seconds per crit so 10/180*100= 5.5% decrease is SF CD per MF crit
20170*0.055 = 1109 dps per MF crit

Given 30% crit chance the chances of crits per full MF cast are:
Chances to crit once per full cast would be p=1(3*1)*(0.3*1)*(1-0.3)^(3-1)= ~44%
Twice would be 19%
Three times would be 3%
Converted to dps it would be:
0.44*1109= 488
0.19*(1109*2)= 421
0.03*(1109*3)=100
Total = 1009 dps per MF assuming SF always has more than 10seconds of CD left.

the probability that a 1 orb will proc during a MF should be:
p= 1(3*1)*(0.18*1)*(1-0.18) ^(3-1) = ~36%
the probability 2 orbs will proc is ~8%
and 3 orbs is 0.5%

My MB average hit is 4636 so the damage from an orb is ~1545 and damage increase to MF is:
From the chance 1 orb procs 0.36*1545= 556 damage
From the chance 2 orbs proc 0.08*(1545*2)= 247 damage
From the chance 3 orbs proc 0.005*(1545*3)= 23 damage
Total = 826 damage assuming 0 orbs have proced when MF is cast.
Note: If you have 1 orb already proced then only 1 or 2 orbs can proc from the cast, if you have 2 orbs proced already then only 1 orb can proc from the cast.

My MF average hit is 1982 and cast time is 2.2 secs so: 1982*3/2.2=2702dps
So MF dps including sin and punishment is 2702+1009= 3711
With 0 orbs proced at MF cast its 3711+826 = 4537
With 1 orb proced its 3711+803=4514
With 2 orbs proced its 3711+556=4267

My MB is a 1.4 sec cast and average hit is 4636 so dps is: 4636/1.4= 3311dps
With 1 orb = 4414dps
With 2 orbs = 5518dps
With 3 orbs = 6622dps

So the difference between MF and MB 2orbs is 5518-4267=1251 dps
The difference between MB 2orbs and MB 3 orbs is 6622-5518=1104 dps
MB with 2 orbs is the better cast by 1251-1104=147dps.

So if you have lower than 30% crit you are better off casting MB at 2 orbs. Crit scales better with MF though so at some point MF may become better than MB with 2 orbs. Infact I would full expect MF to be the better cast for well geared priests as my gear is pretty substandard.
This is still a little crude in some parts of the calculations though but I figure its a decent look at the relationship between the MF, SF and MB.

Notes: I have the 4t10 bonus.
The cast times were taken from a combatlog.txt (think I had a bit of lag at the time I did it)
The average damage was taken from a recount which I have screen saved if people want me to post it.

Ninahagen · 10/29/10, 5:57 AM

Originally Posted by nadoo

If the difference between MF and MB 2 orbs is less that the difference between MB 2 orbs and MB 3 orbs then you should wait for 3 orbs.
(...)
Also why are you talking about damage per cast? All you've done is take away the time factor which prevents you accurately comparing abilities with different cast times. You certainly can not compare MF and MB in any meaningful way using dpc as you've suggested in the quote above.

You may compare MF and MB using DPCT, in fact that's what you did (not very well), maybe thinking it was a dps based.

My SF dps is 2017 and it has a 15 sec duration so:
2017*15= 30255 damage per cast
30255/1.5 = 20170 dps

SF's dps cannot be both 2017 and 20170. 30 255 (damage) / 1.5 (casting time) = DPCT.
I advice you staying in a pure damage view, then in the end switching to dps ou DPCT if you want, but don't mix that quick.

You corrected somebody claiming he was not accounting time (which I don't believe), then not accounting time yourself.
I use your values, shorter :

SF's damages : 30 255 damage.
SF's base dps : 30 255 / 180, 168 dps.
SF's DPCT : 30 255 / 1.5, 20170 dpct.

One MF's tick that crit decrease SF's cd by 10 seconds. You must evaluate how much MF you need in order to have SF available (time passes while MFing). The difference between the 2 cooldowns gives you the DPS gain (not as simple as 30 255 / avg_Cooldown - 30 255 / 180, because you cast less MF when casting more SF, you must work with DPCT, search for TiFi's work in here).

About your work on crit, I think an easier way is to multiply by (1 + %crit x 1.09). Way easier than probs.

My MB is a 1.4 sec cast and average hit is 4636 so dps is: 4636/1.4= 3311dps
With 1 orb = 4414dps
With 2 orbs = 5518dps
With 3 orbs = 6622dps

It is not a dps difference. If you must wait in order to have 3 orbs, MB's dps is damage_MB / avg_time_between_2_MB.

There's no way Mind Blast does 6622 dps in your performances.

---
edit : Method

MF does mf_basedamage in 3 or 2.5 sec with 4T10 (no haste).
MB does mb_basedamage in 1.5 sec (no haste, nor orbs).

At this point, casting MB instead of MF results in a local difference of mb_basedamage - mf_basedamage x 1.5 / 2.5, in term of damages without accounting crit.
With crit it's (mb_basedamage - mf_basedamage x 1.5 / 2.5) x (1 + %crit x 1.09).
So, if you cast MB (without orbs) every avg_time_between_2_mb, the dps gain or loss is
(mb_basedamage - mf_basedamage x 1.5 / 2.5) x (1 + %crit x 1.09) / avg_time_between_2_mb.

But of course, what about orbs, and Shadowfiel's cd ?

If you cast a MB instead of a MF, you lose 1.5 / 2.5 x 3 (ticks) chances of proccing orbs.
The damage from an orb is (0.34 + 0.043 x (mastery - 8)) x mb_base_damage x (1 + %crit x 1.09).
The damage from two orbs is juste twice that and 3 orbs three time that.
For each tick, you have 18% chances proccing one, so the expected numbers of orbs you may have from this (non-integer) number of ticks is 1.5 / 2.5 x 3 x 18%.
And so, the average orb's damage lost from casting a MB and not a MF is
1.5 / 2.5 x 3 x 18% x (0.34 + 0.043 x (mastery - 8)) x mb_base_damage x (1 + %crit x 1.09).
And for dps lost, divide it by avg_time_between_2_mb.

Casting Shadowfiel instead of mind flay results in a local difference of sf_basedamage - mf_basedamage x 1.5 / 2.5.
With crit, if crit is shared between you and your pet, multiply by (1 + %crit x 1.09).
So the dps lost or gained is (sf_basedamage - mf_basedamage x 1.5 / 2.5) x (1 + %crit x 1.09) / time_between_2_shadowfiels.
Anyway, if you crit 18 times with MF, you earn a free shadowfiel, and so (sf_basedamage - mf_basedamage x 1.5 / 2.5) x (1 + %crit x 1.09) damages more.
So, with 1.5 / 2.5 x 3 ticks of MF, you lose in average (1.5 / 2.5 x 3 x %crit) crit ticks, which is a portion of what's needed to gain the damage bonus from an additionnal shadowfiel.
Thus you lose
1.5 / 2.5 x 3 x %crit / 18 x (sf_basedamage - mf_basedamage x 1.5 / 2.5) x (1 + %crit x 1.09).

And, rougthly, that's it.

By casting MB instead of MF, you gain and/or lose, in term of DAMAGE,
(mb_basedamage - mf_basedamage x 1.5 / 2.5) x x (1 + %crit x 1.09)
-
1.5 / 2.5 x 3 x 18% x (0.34 + 0.043 x (mastery - 8)) x mb_base_damage x (1 + %crit x 1.09)
-
1.5 / 2.5 x 3 x %crit / 18 x (sf_basedamage - mf_basedamage x 1.5 / 2.5) x (1 + %crit x 1.09)

If the total is positive, cast MB on each cd, and by this gameplay gain total / avg_time_between_2_mb dps (probably nothing).
If it's negative, cast MB the later you can, and try not to lose orbs (2 orbs strategy is less risky than 3).

Dps from orbs only depending on the mastery, and the frequency of your MF ticks, thus quite constant (little better with a 3 orbs strategy than with a on cooldown strategy, because you cast relatively more mind flay, but that has already been incorporated in MF calculation, so Orbs DPS is considered constant whatever the gameplay).

Hade$ · 10/29/10, 6:07 AM

Originally Posted by Niimue

It was not clear to me what was tested and what was calculated.

You're right, I was a little messy ^^
I did some tests and I wanted to predict the future ones in order to make them quicker.

The board contains both predicted numbers and tested numbers.
The tested numbers are Bold and/or Red/Green.

Here's my new results for S0/1 D1/3 :
Three results were 1 rating point lower and the last one is 1 rating point higher than expected :

Conditions	DP1	SWP1	VT1	DP2	SWP2	VT2	DP3	SWP3	DP4	VT3	DP5	SWP4	DP6	VT4
S0/1 D0/1	205	273	328	615	820	984	1025	1367	1435	1640	1845	1913	2255	2296
S0/1 D1/3	171	238	292	576	779	942	983	1321	1388	1591	1794	1862	2200	2240
S0/1 D2/3	137	204	258	539	740	900	941	1276	1342	1543	1744	1811	2146	2186
S0/1 D3/3	104	170	223	502	701	860	900	1231	1298	1497	1696	1762	2093	2133
S1/1 D1/3	7	71	122	393	586	741	780	1102	1166	1359	1552	1617	1939	1978
S1/1 D2/3	-26	38	89	357	548	701	740	1059	1122	1314	1505	1569	1888	1926
S1/1 D3/3	-57	6	56	322	511	663	701	1017	1080	1269	1459	1522	1838	1876

I removed the S1/1 D1/3 line because it's impossible to test ^^

Nimiks · 11/02/10, 2:55 PM

In case someone is wondering, WoL doesn't take in account Shadowy Apparitions for the moment, that's why you may see a 3/5% lower dps on the log than on recount.

edd · 11/02/10, 7:54 PM

... and no combat logger can. The necessary information isn't in the combat log - there's nothing to tie the shadowy apparitions to the player. Just so you know it isn't Miles and co at fault - Blizzard has to fix this.
I'm not particularly convinced Recount can do it right either, but I've not tested it.

Nimiks · 11/03/10, 2:25 PM

Shadowy apparition appears in recount, but i don't know if it recognize yours or not, we were 2 shadow priest in my last raid. And in fact the combat log doesn't mension the players name, sadly.