Originally Posted by ebbv

I'm pretty sure haste doesn't reduce or increase latency, ergo why I continued to use the 92ms figure even for full length casts. Please direct your derogatory tone elsewhere, it's not at all helpful.

Originally Posted by Vannik

If your spell cast is 3 seconds long and there is 92 ms of lag. (3.092-3 = .092) which is what your numbers were quoting, but you had haste on . . . if your spell casts are 2.87 seconds long (because of the haste you had) 3.092 - 2.87 =! .092.

Originally Posted by Hate Monkey

Vannik are you not understand a few of ebbv's posts or something?

With straight Fireball spam, he had an average of 92ms delay between all the Fireball tests.
With a 3 Fireball/Fireblast rotation, he had an average of 333ms delay because of the GCD incurred by the GCD.

Now when you factor in the bug of Ignites not stacking when Fireblast and Fireball crits at the same time, and the fact that at a certain level of gear, a Fireblast crit will do about 1/3 that of a Fireball. Combine those two flaws of that cast rotation, and you'll see that it is in fact, not worth using a Fireblast rotation past about a T4 content level. But then again, this varies depending upon effective crit rate too.


On a completely different note, has anyone crunched the numbers yet with a full Haste set of gear, 10-15% or more, at the T6 gear level, with a 35%+ crit rate raid buffed, where does the dps end up if you stager Bloodlust, IV, Skull, rather than popping them all at once to get the rolling bugged ignites going? Should have like 2.2ish second cast fireballs at this point.

Originally Posted by Vannik

Your first parsings have ignore 20% of your long casts. Let's neglect completely how you may have had a different ingame latency between days and times of testing things.

You said "Ignoring the first set, the last 4 still each took an average of 635ms longer than we'd expect to complete the rotation. That is to say that our 7.24sec rotation becomes 7.875sec rotation in reality, and that means that 8% of your cast time is being eaten up by latency."

you can't have a 7.24 second rotation, it's capped at 7.5. I think people arn't worried about the haste, they are worried that you're adding increased lag because of the cooldown. if you go buy your first set of data, which i would prefer since it was all taken at the same time, 7.875-7.5 = .375 / 7.5 = 4.7%%. fireball spam gave you 3.1%. Now if you include your new data WHILE STILL hitting the GCD (which is why people said a 3/1 rotation would be better: 333ms => 4.2%

Originally Posted by Rounced

I don't get why people keep saying this. One aspect of the "fix" to the casting mechanics is that it should turn 1.5s casts into true 1.5s casts. Thats why the GCD occurs whenever you request a spell.

I think the real problem is that we are all relatively proficient at using Quartz and /stopcasting so that we can aim for "button up" right when the casting bar hits the red but since Quartz doesn't run with an instant cast spell we lose that frame of reference and our "button up" occurs with a slight delay.

Just did some testing on the PTR. Cast 20 Fireballs and 28 for each of the other spell sets.

Casting with Quartz and aiming for "button up" at the red
Fireball - average 3.046
Scorch - 1.602


Spamming cast button as fast as I could
Scorch - 1.56
Ice Lance - 1.593


With all the sets I had several casts under the 1.5s mark which leads to an interesting question; is the lag due to me or the mechanics of the new system?

Fastest Casts (no haste gear)

Fireball - 2.719
Scorch - 0.969

Scorch Spam - 1.312
Ice Lance Spam - 1.375

Being able to cast under the supposed minimum casting time also seems like it should be worth exploring to see if it is real or just due to the data being recorded on my client and the latency between my client and the server causing the discrepancies.

So does anyone want to pull out a G15 and go after Dr Boom with Ice Lance and Scorch spam again (Manly)?

Originally Posted by Etherealz

I can assure you that instant cast spells will give you latency issues - and I am using the G15 spam cast methodology. My average fireball cast is very close to the true 3s fireball - obviously not counting the initial latency during the start of chain casting.

Just to be sure - I went on live realms and tested this - using a g15 with a very small delay - it took 5 minutes for me to get an entire rotation behind. I have a 180 ping on live at the moment. And coincidentally 7.5s over 39 fireblast rotation sets is 192ms on each fireblast.

Similarly I did a test with just fireball spam - and in that same 5 minutes I was about 1.5s behind. Take 200ms off for the start of the chain and that a 13ms delay per fireball - which matches pretty well since my keystroke delay is about double that.

Edit: I also made sure to take off all haste gear during the test - And also, I think maybe you have instants and 1.5s channeled spells a bit confused .

Originally Posted by Rounced

I'm a bit confused as to what if anything you are establishing there.

5 minutes to do what? 1.5s behind what? 200ms and 13ms over what number of casts and with what keystroke delay??? What the heck is "G15 spam cast methodology"?

are you using a G15 to spam cast Fireball?!?!? Have you read any of the thread previous to the release of 2.3??? Where they showed that spam casting using a G15 on a spell with casting time should destroy your casting time since you will be introducing a GCD with every "key press" and have the spell request only sent to the server once the server responds back that the spell completed, which completely removes any advantages from the new casting system. If you were still getting true 3 second casts using a G15 to spam the server then everything that so many people proved during the testing for 2.3 has just been completely invalidated and I guess everyone better run out and buy a G15 ASAP and set it for a very small delay and just spam away.

(if that isn't what you are doing then I apologize for my sarcastic response but I really can't follow what you are saying you did from your post)

My spam casting was only to test with instant cast and 1.5s spells and the results were clearly within the realm of uncertainty (1.56s average for Scorch and 1.593s average for Ice Lance) to show that the instant cast mechanics are detrimental to its use in a rotation. There seems to be an average delay per spell cast request and that may influence whether the rotation is a dps boost or not but nothing that I have seen leads me to conclude that incorporating an instant cast spell gives any additional latency over the simple additional lag involved in requesting any spell.o

Please take a little time to make your post legible and be sure that it actually shows what you are saying when you "prove" something. Otherwise you are doing nothing more then saying "it looks like to me that there is a delay and [b]I know[b] that it is fact because I can see it" when there has been no proof of any kind so far that conclusively shows any of it. You might as well have just jumped into the thread and said "Cars go fast" for all the relevance your testing had in establishing anything as proof.

Until someone does some testing with a couple thousand fireball-ice lance rotations, and compares it to straight fireball chaining and straight ice lance chaining, I will continue to have serious doubts that there really is additional lag involved in that particular mechanic. Your "5 minutes for me to get an entire rotation" with a g15 definitely hasn't done much to change that particular perception.


One last thing, you say you tested this on the live realms, there are supposedly some additional refinements coming to the casting mechanics in 2.3.2 and that was why I did my testing on the PTR. Since 2.3.2 is literally right around the corner maybe you should move your testing there so that it will be relevant once the patch hits instead of being completely outdated and more useless then it already seemed to me.

Originally Posted by Roywyn

Mana vs. Damage

With the recent changes - talents, gems, flame caps (3m cooldown) - mana management has suddenly become more interesting in a min-maxing scenario.

"But there is no mana issue!" - Well, there is, but in a more subtle way. Can you spec out of all mana talents, use flame caps, destruction potions, molten armour, no evocation and still last through a fight? Maybe with a very good shadow priest, but what if you don't get one on Sharaz/Gurtogg?

There are quite a few means of recovering mana, but which one to chose in which situation?
Let's take the following scenario:

A deep fire spec as base, Fireball spam. T6 level gear, 350 spirit raid buffed (100mana/tick or 250mp5 when not casting), 12k mana (for Evocation, a Z'A staff can push that to 13k).

I take a 6 minute boss fight, uninterrupted casting. That way, you can use your 2m CDs 3 times, or 3m CDs 2 times. That fight length is a multiple of these cooldowns, so nothing can get a benefit from cooldown coming up early.
The actual length of the fight shouldn't matter too much to the model, but I think that one is a pretty decent model.


How much is spell hit worth? At T6 level gear, people generally grab their spell hit from gear and gem for damage, which is generally accepted a s the best choice.
Seeing that T6 gloves are usually socketed with 12 damage instead of 6dam/5hit+2dam bonus, 5 hit is not as good as 12-8=4 damage. I'll just set the value of 1 spell hit to 4/5=0.8 damage.


Regarding talents and talent choices - Arcane Meditations means you lose Elemental Precision, Arcane Concentration means you lose Icy Veins. A fire spec needs 42 points to maximise PWF, Critical Mass, Fire Power, Pyromaniac, Combustion, Molten Fury, Empowered Fireball.
To get Clearcasting and Icy Veins, one would lose more in fire talents than the Tri-Spec gains.



Talents

Elemental Precision
3% hit translates to 37.8 hit rating or 0.8*37.8 = 30 damage. 3% cost reduction is 12.75 mana less per cast, or +21mp5.

Icy Veins
20% haste for 20s every 180s => 2.2% haste averaged. Makes it ~35 haste rating averaged, or 39 damage and -12mp5.

Arcane Concentration (Clearcasting)
Saves ~40 mana per cast, or +66.5 mp5.

Arcane Meditation
Returns 30%*250mp5 = +75mp5 while casting.

Master of Elements
Assuming 35% crit chance, it saves 425*30%*35% = 45 mana per cast, or +74 mp5.


Armour/Evocation

Mage Armour returns 30%*250mp5 = +75mp5 while casting.

Molten Armour gives 3% crit ~= 68 crit rating ~ 51 damage and +6mp5 from Master of Elements.

Evocation with 2/5 T6 gives you 75%*12k = 9k mana at the expense of 10s cast time. That means a 9000mana/6min = +125 mp5 at the cost of 10s/6min = -2.77% haste = -43.6% haste rating = -49 damage, +15mp5.
So, using Evocation is a +140 mp5, -49 damage change.

Using Evocation in the last moment of Heroism makes it -38dmg +11mp5 from the time lost casting, or +136mp5 at the cost of -38 damage.
Then again, you'd want Heroism when the boss hits 20%, which is too late to use Evocation.


Potions

Mana Potions return +100 mp5, or +140mp5 with the Alchemist's Stone (few mages are alchemists though).

Destruction Potions give 120 damage and 2% crit for 15 seconds on a 120 second cooldown. That's +19 damage and +0.5mp5 from Master of Elements.


Gems/Flame Caps

Mana Gems return +100 mp5, or +125 mp5 with the Serpent-Coil Braid.

Flame Caps give 80 damage for 1 minute on a 3 minute cooldown, or +26.6 fire damage average.


Trinkets

Icon of the Silver Crescent averages at +69 damage, Hex-Shrunken Head at +88 damage, an ideal DM:Crusade is +80 damage.
Thus, I'll just assume a static +80 damage trinket to compare the other two mana trinkets to.

Serpent-Coil Braid gives 30 crit rating, 12 hit rating, 25mp5 and 225 damage for 15 seconds every 2 minutes. I take 12 hit as 12*0.8 = 9.6 damage (see above).
This adds up as +60 damage and +28 mp5 together with Master of Elements.

The Alchemist's Stone gives +15 to all stats and 40mp5. The 15 stats yield about +5 damage (crit and IDS), 248 raw mana, or 450 mana with evocation, and +0.5mp5 from crit.
So, a total of +5 damage and +44mp5 to +47mp5 averaged over the fight.


Blessings

Blessing of Wisdom gives you +41mp5 or +49.2mp5 if talented.

Blessing of Kings gives you ~35 spirit and ~60 intellect. This is ~19 damage, +1.5mp5 from crit, 900 mana, or 1615 mana with Evocation (and some spirit ticks), +24mp5 total.


Summary/Comparison

This is a list of choices you can make to maintain your mana.
It's ordered by efficiency, by the amount of mp5 you gain per point of sacrificed damage.

1) 5.27 mp5/dmg - Chaining Mana Potions over Destruction Potions gives you +99.5 mp5 at the cost of -19 damage.

2) 3.70 mp5/dmg - Chaining Mana Gems over Flame Caps gives you +100 mp5 at the cost of -27 damage

3) 2.86 mp5/dmg - Evocation gives you +140 mp5 at the cost of -49 damage.

4) 2.03 mp5/dmg - Arcane Concentration instead of Icy Veins yields +79mp5 at the cost -39 damage.
5) 1.80 mp5/dmg - Arcane Meditation instead of Elemental Precision yields +54 mp5 at the cost of -30 damage.

6) 1.40 mp5/dmg - Serpent-Coil Braid gives you +28 mp5 at the cost of about -20 damage.
7) 1.35 mp5/dmg - Mage Armour gives you +69 mp5 at the cost of -51 damage.
8) 1.36 mp5/dmg - Talented Blessing of Wisdom over Blessing of Kings gives you +25mp5 at the cost of -19 damage.

9) 0.89 mp5/dmg - Untalented Blessing of Wisdom over Blessing of Kings gives you +17mp5 at the cost of -19 damage.

10) 0.63 mp5/dmg - The Alchemist's Stone gives you +47 mp5 at the cost of -75 damage.


This means if you're min-maxing to the extreme, you'd chain mana potions as first priority, and mana gems as second priority.
Using Evocation is a clear third priority, and respeccing to 10 or 18 arcane are choic 4 and 5, both with comparable returns when you can get the hit from somewhere else.

After those, using Serpent-Coil Braid, Mage Armour and Improved Blessing of Wisdom instead of Kings (if you lack paladins for both) are the next choice.

A normal Blessing of Wisdom is a weaker choice after that, and the Alchemist's Stone significantly weaker.

Originally Posted by Etherealz

I'll agree with most of the estimations used - except for the hit one.
While there is a lot of hit on tier6 level gear it should still be taken as face dps value.

Originally Posted by Roywyn

You used a 6dmg/5crit in your gloves, so you value 5 hit < 5 crit in your current setup

Originally Posted by Etherealz

I was using the stats of 1600 spell, 280 crit rating and 31 haste - I did mean 1 pt of +spell dmg vs 0.54 hit rating vs 1.2 crit rating vs 0.75 haste rating.

Using both my own spreadsheets and Lhivera's calculator - I get these comparisons. I haven't found much err in Lhivera yet.

Edit: I was also considering icy veins in the haste calculations.

Originally Posted by Roywyn

Obviously, it all depends on your current gear. If you don't have proper trinkets, SCBraid may be your best trinket, and gives you a chunk of mana back in addition.

Illidari Cowl, rest T6, Belt of Blasting, Tempest/Chronicle, Skull, Solarian Wand gets you +14.4% hit.
So without Elemental Precision, you have to get some +hit from elsewhere, so +hit has a larger benefit.


The whole idea was to give a general scheme of the mana regen mechanics and their hidden damage cost.
Of course it's going to depend on your stats, and more even on the gear you have and the gear you can switch in.

You used a 6dmg/5crit in your gloves, so you value 5 hit < 5 crit in your current setup

Originally Posted by Muphrid

1 +dmg = .54 haste is simply low compared to typical values. Same for haste, in my opinion.

Originally Posted by Etherealz

For somebody raiding ssc calculating for say 1200 + dmg

it would be more like

1 dmg = 0.71 hit = 1.55 crit = 0.96 haste.

But since we're calculating for raid dmg, I thought I'd use my fully buffed stats to chose my raid gear :P.

Originally Posted by Searix

Edit: Added in PTR testing data

AVG: 2.887, expected fireball cast time (Yes i used haste): 2.84

DPS lost to latency: 1.65%


Trial 2:"Spammed it so hard my arm felt it"


AVG: 2.889

DPS lost to latency: 1.69%


Trial 3:"Mouse wheel casting"(Bound mousewheel up and down to fireball, went constantly up and down in a rubbing motion, it took me a couple minutes to adjust to doing it)


AVG: 2.851

DPS lost to latency: 0.38%

I took a high school AP stats class, i know how stats work in a decent degree, and i can attest that i didn't doctor or subvert the data, but by that note i'd like test results outside of my own on this.

I leave the diagnosis to others

Originally Posted by Muphrid

I don't doubt it. It's just interesting how quickly the equivalences shift, especially since we usually emphasize that these equivalences don't change very much for simple gear changes or upgrades. Then again, it's an order of magnitude difference in terms of the changes we're talking about, so it's not wholly surprising, either.

Originally Posted by andastra

I've actually seen the equivalences of spell haste shift quite a bit. I don't flask or use much in the way of consumables on the early fights on MH/BT anymore. I've noticed that using flask + oil shifts the equivalent of spell haste to me from roughly 1.05 damage to 1.2.

Going from roughly T5 to T6 or getting the 4pc T6 bonus could cause quite a shift on spell haste equivalency.