I noticed an interesting thing with Hot Streak on EU-Beta-Servers. I wanted to test, wheter a third critical Scorch or Fireball in a row is wasted or not if you do not cast your instant pyro asap after the second crit. Therefore I casted some Scorches on a Target Dummy, always waiting a little between casts, so they can't overlap. And this is what happened:

....... Scorch (critical)
....... Scorch (critical)
.. ++ Hot Streak
....... Scorch (critical)
... -- Hot Streak
....... Pyro (instant)
....... Scorch (critical)
.. ++ Hot Streak

That means you can delay your instant Pyro for one cast without loosing any damage. However, if you cast 4 Sorch/FiBa criticals in a row without using a Pyro in between, you would waste one Hot Streak-Proc (you can't get two stacks of Hot streak).

Next thing I will test is, if it is possible to cast 4 Fireball Crits in a Row followed by two instant Pyros. Schould be possible, when you cast your pyro before the last Fireball hits the target, though it would be stupid to do so dps wise.

Screenshot of the combatlog:



EDIT:

It is indeed possible to cast 4 critical Fireballs in a row followed by two instant Pyros:

....... Scorch (critical)
....... Fireball (critical)
.. ++ Hot Streak
....... Fireball (critical)
....... start casting Fireball
....... shoot instant Pyro (Hot Streak fades)
....... Fireball lands (critical)
.. ++ Hot Streak
....... shoot second Instant Pyro (Hot Streak fades)

Screenshot:

I was pretty sure this was wrong, because I have often used Invis to drop threat in the middle of a boss fight while my Water Elemental was attacking. So I went and tested it.

I summoned the water elemental, told it to attack a wolf, and hit invis. The elemental got two shots off before invis kicked in, without interrupting the fade.

When invis kicked in, the water elemental switched to passive mode and stopped attacking. This confuses me a bit as I'm certain I've never seen such behavior on a boss fight.

The Water Elemental took several hits and about a thousand points of damage while I remained invisible.

The wolf then cast its AOE howl, which broke me out of invis.

So the only thing here that would be a problem is the Water Elemental switching to passive mode. (I didn't try manually switching him back to defensive or aggressive.) However, non-controllable pets don't have aggressiveness settings, so this may not be an issue for the images.

The "Combat Ratings at level 80" thread has a very exhaustive test made to confirm that your spell resist rates are unchanged.
If it wasn't tested, it wouldn't have been in the first post
Link to Althor's test: http://elitistjerks.com/861956-post131.html

You also get your +3% from Misery/Faerie Fire instead of ToW.
So, 11% hit needed for for Horde. Alliance only needs 10% if the Draenei racial survives the next patch.
They're going to take a deep look at racials, with Stun resist changing I can't see 1% party wide +hit not get changed.

That means 262.3 spell hit to cap 10% hit, 288.6 spell hit to cap 11% hit.


[Edit]: The 9% spell hit cap "urban legend" came from Malan when he did some testing when Enhancement specced.
Nobody knew at that point that his melee hit talent (Dual-Wield Specialisation) also gave him 6% spell hit.

Your pet only see what you can see, in arena whenever I cast invis, my pet stop seeing the other players (unless they have see invis , in this case the pet will be able to stay on them). But as your pet doesn't go invis at the same time, it's still targetable by the players/boss and can be killed, basically he's just blinded by your invisibility, and as he cannot see anything he will just usually start following you (and will probably show the opposing team which pillar your are hiding behind). That's not really a passive mode.

Having your pet alive will prevent you from getting out of combat, but not from reseting your agro to 0. In pve or pvp if you really want to go out of combat (in case of wipe in a 5 man instance,won't work for boss fight of course) or you just want to drink/eat in arena the only way is to dismiss your pet before the last invis tick.

Concerning pve hit rating , many people on random forums are claiming that the new hit cap is 9%. Is that true or just an urban legend?

The 9% hit cap was either only true for a couple of patches over 2 months ago or never true. Either way, it has definitely not been true for quite a while now, and I wish that misinformation would stop being spread.

Crit > Haste for Fire Mages

Of course that's not universally true. But I liked starting with a bold claim. So, here are my assumptions:
1) High gear level like in the previous tests. That's 3k spell power fully buffed, 15% haste on gear, 20% crit before talents/(de)buffs.
2) Living Bomb doesn't get MoE procs due to how the spell is designed. No Burnout mana loss either,

The max-DPS Fireball spam rotation with Living Bombs and instant Pyroblasts then looked as follows:
Base: 5050 DPS, 259 MPS
With +10% crit: 5509 DPS, 255 MPS
With +14% haste: 5648 DPS, 290 MPS
With +529 damage: 5713 DPS, 259 MPS
All those increases have the same rating/gem/budget costs.

So, +damage gems are the best you can get, but we knew that already. The question is Potent vs. Reckless orange gems for yelllow socket bonuses.

Gemming for haste over crit in this setting yields 139 DPS for 35 MPS more mana spent.
Even if Living Bomb crits cause burnout mana loss, that increases the +10% crit value by less than 1 MPS.
That is worse DPM and a worse mana conversion than using Living Bomb.

The other variable are Armours.
You can gem/gear for haste over crit and use mage armour instead of molten.
Which one is better depends largely on your intellect/spirit on your default gear.
Always switch to mage armour if it's needed to sustain Living Bombs.

The DPM conversions are as follows:
5.6 DPM for weaving in Living Bombs.
4.2 DPM for using Molten Armour over Mage Armour at 700 int / 500 spi total.
4.0 DPM for gemming Haste over Crit
2.5 DPM for using Molten Armour over Mage Armour at 1k int / 700 spi total.


That means that haste gems are only good if you can weave Living Bombs through a whole fight and still have mana left.
Use Mage Armour if it's necessary to sustain Living Bomb weaving.

On a sidenote, it takes around 3.5k spell power until Haste gems overtake Spell Power gems in an infinite mana setting.

Mirror image looks pretty good.

1) Even if a mod instantly identifies which mage is real, your switching between different images is going to mess up their targetting, casting, and positioning.

2) yes it interrupts casting, but if that is removeable it is no problem. Might mess up mages in pvp but I am pretty sure they would either eat the risk, stick to instant casts, or learn the switch timing to compensate. The ability to mess up targetting and positioning is way too powerful to whine about the impact on casting bars. Its quite ingenious, I think. A mage has a choice between mirror images with no target switching and MI with switching, but the latter pays a dps price if chain casting, no price if spamming instants in pvp. Seems fair.

3) health will be important, but situational. Pop mirror image when a warrior is on you and you are surrounded by the images, and yeah, a WW may dust them all. But pop when you have him sheeped, and the warrior is in trouble. If he tries walking toward you in defensive stance w/spell reflect, he is dead from all the damage. If he interecepts, you blink and if uses WW on the images he has the main mage nuking him safely while getting damage bonuses from zerker stance. if he pops intim shout, the images are scattered and when fear ends they won't be clustered for a WW kill. Its win/win for the mage. Similar with rogues. pop MI and blink (30 yards now with glyph) and the rogue has to choose between killing the images and dealing with the main mage. Either way he chooses I am considerably better off than I am today.

4) This directly addresses one of the mage's major problems in arena, vulnerability and reduced dps when focused. Mage now has iceblock plus a MI that will make him a targetting nightmare for 30 seconds.

It's the same 3k SP fully buffed, and 15% haste, 20% crit pre talents/(de)buffs.
It doesn't actually matter in this case because the crit-vs.-haste comparison is nearly completely independent of your stats.

As for the sets - there are a lot of cloth items, and off-hand and a wand with mp5 instead of spirit. These item types are generally seen as intended for spirit classes though.
A lot of the gear makes so little sense at all that it has to be placeholders. That's what I hope at least


As for fresh Naxx-10 gear, I'll have a look once we know what kind of stat levels we'll see.
The guideline from what we've seen so far is:
Stack spelldamage. Prepare to get a shitload of resists.
dmg = 1, hit = 1, haste = 0.8 (infinite mana); haste = crit = 0.6 (LB limited by mana) from what it looks like.
That's for fire spec. For frost and arcane, crit is worth less.

Not really.

Both sets have identical spell power, and almost identical spell hit. The mage set has 30 more intellect and 10 more spirit, whilst the warlock set has 103 extra stamina. The main difference is crit versus haste. The mage set has 190 crit and 94 haste, whilst the warlock set has 81 crit and 204 spell haste.

There's no way I can understand those numbers - is there a scenario possible where the warlock set isn't simply better? Especially given the different ratios of haste versus crit. It's perplexing also that their set has less spirit than the mage set, whilst actually warlocks want spirit a lot more than we do.

What is even more annoying is that you also change facing when these shifts happens. The number of shifts also seems random - sometimes it happens 1 time pr Mirror Image use - other times it happens 3 times. Not sure if any of your talents affects the images - but they definately cannot apply winter's chill. If you blink the mirror images doesn't blink with you. If you invis the images keep shooting. If you change target the images doesn't (unless the old target died)

It's the same 3k SP fully buffed, and 15% haste, 20% crit pre talents/(de)buffs.
It doesn't actually matter in this case because the crit-vs.-haste comparison is nearly completely independent of your stats.

As for the sets - there are a lot of cloth items, and off-hand and a wand with mp5 instead of spirit. These item types are generally seen as intended for spirit classes though.
A lot of the gear makes so little sense at all that it has to be placeholders. That's what I hope at least


As for fresh Naxx-10 gear, I'll have a look once we know what kind of stat levels we'll see.
The guideline from what we've seen so far is:
Stack spelldamage. Prepare to get a shitload of resists.
dmg = 1, hit = 1, haste = 0.8 (infinite mana); haste = crit = 0.6 (LB limited by mana) from what it looks like.
That's for fire spec. For frost and arcane, crit is worth less.

In Blizzard's world, crit rating = haste rating. From an itemization standpoint, they consider them the same. And locks, even with affliction's new talent, have some builds which get a reduced effect from crit, while mages in their usual raiding build get a lot of procs and benefits based on crit, so I can at least see why from their viewpoint mages would prefer more crit and locks would prefer more haste. It's not true very often, especially while crit rating has such a terrible conversion rate compared to haste rating, but I can see where they'd make the mistake, at least.

What boggles me from a pure item point standpoint is how 103 stam = 30 int + 10 spirit. Stam was multiplied by 1.5x in TBC, but that'd be 60 points, not 100, for 40 "other stuff". Just seems strange. Losing 100 stam should buy us 60 int, not 30, no? (I know the iValue formula is a RMS-type, but this still doesn't "feel" like a good trade to me, even as lowly as I value stam. In fact, given RMS penalizes heavy focus on one stat, splitting stam into int and spirit should result in an improved return, not a reduced one.)

The biggest issue I have with it is the lack of hit. Roy is probably correct (I haven't done the math, myself) that stacking *any* rating before you reach 2.2k spellpower is inefficient, but hit is still going to be better, point for point, than crit *or* haste. And everyone needs a lot of it now, with talents being 3% for every spellcasting class and build, rather than 10% for some. So I'm sort of surprised T7 isn't hit-heavier and crit- and haste-lighter. It'd allow for later tiers to slowly shift from hit to haste and crit as non-set slots gain enough hit to cap, and there'd be much fewer complaints, since every spellcaster set could be equally hit-heavy. (At the very least it'd postpone the complaints until T9 gives locks more haste and mages more crit...and give us more time to convince Blizzard out of their "mages like crit" mentality.)

Oddly I just did the same exercise and only got the following: 1423 dmg, 309 crit, 271 hit, 269 haste, 678 stam, 679 int, 354 spirit. Choosing the best equipment I saw on the Naxx10 list for each slot. No trinkets yet. Spell/hit gems in every yellow slot, full spell in every red slot. No enchants.

Also, you call 328 hit 13.04% somehow. By my calculations it'd be around 12.5%. And the 271 I saw on all-best-hit-rating equipment would be under the 11% cap with a spriest or boomkin.

On a sort of tangential note, what's it going to take to convince Blizzard to make 2 separate mage sets, one with spirit and one without -- an "arcane" set and a "Frost/fire" set. I know spirit's much improved from its previous incarnations, but it's still useless to any non-arcane mage, even more so now that Imp DS wont' stack with other buffs.

At least prior to WoTLK we always had the gear itemized for Warlocks to fall back on -- but now we won't even have that given how Warlocks will actually want Spirit. I feel like we're going to be in the same situation were 75% of Cloth DPS gear will be so undesirable that it'll pass to Moonkins simply because of the over-abundance of spirit on it.

In fact, they don't even need 2 mage sets, just 2 caster sets: "caster -- Spirit" and "caster -- no spirit". That way warlocks and arcane mages would gravitate to one set and frost/fire mages to the other.

You don't get the threat dump from Invis if you don't tell your water Elemental to stop attacking before you invis (it keeps you in combat). My guess is this is meant to fix that issue -- albeit a sloppy fix.

Edit: Sorry for two replies in a row. I should have combined them into one post.

Maledict, the mage set is possibly biased in favor of fire mages, who actually might prefer crit to haste now. (It is at least debatable.)

The lock set is biased in favor of destro, since all that haste is of limited utility for demo or affliction builds.

Arcane and frost mages will prefer the lock set, and affliction and demo locks will envy the mage one.

I'm not sure if this is intended or not, they could just both be placeholders. The lack of set bonuses is itself telling, and leans in favor of the placeholder explanation.

Anyhow, the picture is sufficiently complex that I don't see this as a clear cut case where one set is better than the other and am willing to treat the two as roughly (very roughly) equal for beta testing purposes.

I'm more worried about all the cloth and wands with MP5, which is very bad for the item points cost compared to spirit for every class that can use them.

But what does it matter really ? We already have a 3min invis, which honestly I think is simply far too good. Not to mention, its only 3s 'cast time'. I don't really think you'd need far more than that. If theres a fight where aggro matters, tranquil air will be put down, and be considerably more effective than it is now.

The way I view it, the major advantage of MI (threat-wise) is that it stalls short your aggro for a solid 30s, and also makes it impossible to pull for that duration, without causing any dps loss. Thats the upside of it. However, even if it didn't play with threat at all, invis is still a ridiculously strong spell as it is.

Doesn't tranquil air totem not exist anymore?

To go back to the point about gemming for hit or gemming for damage: if we assume that gemming damage is better than gemming hit, then we (usually) get down to the following inequality...



Where i_h/i_d represents the ratio of hit and spell power's item weights (depending on the specific gems we're talking about, this could be close to 1 or as high as 1.2; it's usually around 1/.85 = 1.18 or so, and the differences are artifacts of integer rounding).

m is the base damage, r the coefficient, d spellpower, g_h the rating conversion for hit (~2623 per 1), h hit chance.

As is typical, I'll solve for spell power.



So spell power must be less than this value (which increases as hit increases, so it's a sliding value, not a threshold). For maximum generalization, we can just let h=1, take i_h/i_d = 1.2 due to rounding (this is for worst-case performance of hit), and just look at the dependence on m/r.

What we get is this:



And as I've previously posted, most of our main nukes have an m/r around 900 or 950, making the sliding threshold around 2200 spell power or so. Arcane Blast is about m/r = 1380, so the threshold is roughly 1770 +spell, and so on and so forth.

Since most of our gear decisions revolve around small changes at the hit cap, this is fairly reflective of how real gear decisions will work. Obviously the model is highly simplified--symbolic math quickly gets hairy for more than one spell in rotation, but for the most part, the spell power threshold should be a kind of weighted average of the thresholds of the individual spells, weighted by the proportion of overall damage dealt by the spell.

Considering how little spell damage it took for hit to become worth gemming in TBC, this does strike me as a fairly significant shift. It seems the power of 1 rating has deflated much faster than the power of 1 spell power. I must admit, when I set out to write this post, I did not expect the results to match Roywyn's remarks, but the algebra is quite clear.

So I tried for fun to check the images mana sustenancy since they seemed to have an uncanny tendancy to go melee up mobs at the end. Sure enough they do manage to run oom.

8814 max mana
828 mana per fireblast (9.39% mana)
621 mana per frostbolt (7.05% mana)

Using no external mana regen, sometimes they go oom before the 30s is over. On average, they seem to finish at 10%- mana. Using gear or no gear has no iumpact on their shown mana pool or damage.

Now I need to test to see if they get the group/raid mana regen, and if they are affected by totems/bloodlust.

How does a Priest's Fade spell work?

If I remember correctly, it can give you negative threat temporarily.
The important part is that your threat doesn't matter when you cast Fade, so you can cast it proactively to prevent getting high threat.

Mirror Image is listed as a 90 million threat reduction buff just like fade.
What I'm thinking is that it may make you immune to threat spikes at the start of a fight:

0:00 - Cast Mirror Images, you're at -90.000.000 threat.
0:01 - Scorch, blow Trinket/Combustion/IV/whatever cooldowns.
0:29 - You're at -89.900.000 threat now. Then Mirror Image fades.
0:30 - You're now at 100.000 threat.

The only way you could pull aggro now is if your tank together with Hunter and Rogue Misdirects and couldn't produce more threat than you in 30 seconds, which is quite unlikely unless their threat is generally bad.
If things are close, you can also just use invisibility at 0:27.


Can anyone test if it works like that? It should, unless I'm remembering Fade wrong. It's been changed pretty often now.

Assuming the mirrors do gain from raid and party regen then i would assume they would also gain mana from Imp WE when soloing and pvping? I suppose that is only relevant in pvp assuming frost spec, as there would be enough mana regen in raids to support it anyway without the WE.