Is this a result of any changes to GCD behavior in beta? I can't recall ever trying such a macro on live (mostly since there has been no point to instantly interrupt a cast before Hot Streak) but I can't ever think of a time where I experienced any latency/GCD issues using a CS macro immediately following a start to a spell cast. I was under the impression that macros are sent as a string of commands essentially simultaneously and the stopcasting command in the macro would interrupt both the cast and the GCD on the server-side essentially instantly. There shouldn't be any latency related GCD conflict if I understand it correctly.

Again I'm still without a beta key so I'm not sure if you're referring to the GCD-related changes, but it seems like it would work fine on live.

I wouldn't consider that valid DPS testing, without a reliable means to drop below 35% and stay there (not receiving heals from any source including VE, chain heal, CoH, etc.).

I don't think we will ever truly know how that number was achieved, but you can get an idea that some programming must have been horrifically bugged for it to happen. The way I look at it is; consider FFB with a 300% crit bonus, a 50% crit rate, and a whopping 50% haste to make it a 2 second cast. You would need the damage sum of a non-crit FFB and a crit FFB to total 32,000 damage for 8,000 DPS. That requires an 8,000 noncrit FFB followed by a 24,000 crit FFB in 4 seconds. That is simply an off-the-wall damage amount that contradicts any conceivable maximum scaling level including all double-dipping... you would need an absurd amount of spell damage near 4k and stacked multipliers (that isn't even possible anymore) to come up with a noncrit FFB for 8000, not mentioning the ludicrous amount of haste item-budget required. My biggest question has always been not that of how a mage produced 8k dps... but how they could only get a rogue to produce 3k max.

CS isn't on GCD so your example would work, any other spell would still have to wait for the GCD.

No, it wouldn't work on Live either. Keep in mind CS is completely off the GCD. It can be cast WHILE the GCD is rolling (just after an instant cast spell, for instance). Trying to cast a spell that activates the GCD, even if there's a reason it won't work (such as currently casting another spell) activates a GCD which doesn't cancel until the server spits back an error about trying to cast something when you're not ready. /cast AnythingThatActivatesGCD in a macro will, if not on cooldown or OOM, functionally cause the macro to ignore any further /cast lines, regardless of the use of /stopcasting. This was part of the changes going way, way back to limit the use of "intelligent" macros.

I appreciate all the math you do -- the time you invest into exploring and theorycrafting all of this. It definitely adds to the usefulness of this thread (and forum in general). With your calculations above, is that to say that in pure DPS terms, 1% crit is only slightly behind 1% spell haste? Would any other factors tip the favor for crit rating exclusively?

Consider [Blade of Twisted Visions] and [The Maelstrom's Fury]. In question then is 1.33% haste vs 1% crit and 7 spell damage. Rawr already shows it as a .5 DPS increase, but come 3.0, would it be significantly higher? And would that transfer over to gems, perhaps, as well? Will [Potent Pyrestone] be favored over [Reckless Pyrestone]? Essentially, isn't this why recommended talent builds for 3.0 favor clearcasting + Living Bomb over Icy Veins? And if we're dropping icy veins, then I would imagine haste isn't all that important in other cases, and crit rating helps MoE all the more.

I don't want to rush out and begin regemming and what not, but it's interesting to think about and plan ahead.

The problem with crit isn't that 1% crit is a lot worse than 1% damage or 1% haste (it is most often a bit worse though but this might change with new talents). The huge difference is that 1% spell crit requires a lot more rating than 1% spell haste. Wowwiki says 22.08 spell crit rating for 1% crit and only 15.76 spell haste rating for 1% haste.

That explains it better, thanks. Wasn't aware of the specifics to the macro rules. I suppose the most practical solution would simply be the addition of stopcasting before Pyroblast and use it based on discretion. If you're spamming scorch and are quick enough to recognize the 2nd crit to make it worth cancelling the next cast, it seems reasonable. Would really make it feel more comfortable if Fire Blast was included in the spells that proc HS. Without an instant cast to proc it you're essentially assured to be mid-cast of the 3rd cast when the proc occurs.

I guess you could really get button-mashing-happy and quickly hit Pyro before every Scorch/Fireball cast in a stopcasting macro to the point of near simultaneously hitting each key (or through a simple castsequence, but that would get messy occasionally). Just seems like if you want to optimize the procs you can't be wasting casts on potentially overlapping crits (e.g. 2x Scorch Crits, Mid-cast next Scorch when proc realized, 3rd Scorch Crits followed by Pyro ... instead of 2x Scorch crits, Pyro, begin next scorch).

Latency pretty much ruins all such plans. You want smooth chaincasting and /stopcasting messes with that. Even scorch is good enough DPS you don't want to throw away half its casting time to do zero DPS for .75 sec just to get the pyroblast out a little sooner. This is why interrupting casting for things like clearcasting is a bad idea too.

Just have pyroblast in another key and slip it into the next chaincast opportunity when it procs. The buff lasts long enoguh for you to mix it into your rotation after the fireball you already have queued up. Use it like you use fire blast today, when you're close enough to use fire blast.

If you take the 35% health/faster pyroblast talent you may want it on a nearby key anyway. There are now two "faster pyroblast" scenarios, both situational but together enough to justify having a key for it somewhere if you take the talents.

Just did a couple deep fire builds, with and without icy veins. I'm pretty happy with the way it looks. I might even be able to keep my favorite PVP talents in the build at 80 without giving up raid DPS. The DPS portion of the build is a little less bloated than it used to be, relative to total talent points, and the impact/molten shields went from 8 talent points to 4-5. Much easier to squeeze in.

One interesting thing about molten shields is that if it stays as the tooltip says, nobody will take more than one point of it. (it's the ranged molten armor+impact that matters, fire ward stuff is jut icing).

The frost tree leadup to icy veins is also more useful than it used to be for a deep fire build. The ice armor fire/frost ward buff is occasionally helpful, the reduction in cooldown of ice block etc is useful, elemental precision is useful and you have two points left...permafrost for your frostfire bolt is nice as an opener on anything that can be chilled.

I've also been trying to figure out how to use firestarter as a replacement for blast wave in opening up on AOE trash at close range.

Today the sequence is Dragon Breath, Blast Wave, then spam Arcane explosion till I have to stop due to incoming damage or the mobs are dead.

With blast wave being a "move enemies" tool, it goes over with frost nova as "not a routine thing in a raid", but I will miss having all my fire talents on the first two zaps.

Fully talented firestarter gives me an instant flame strike to replace blast wave with, as dragon breath will proc it 100%. So far so good. Trouble is it is a ranged AOE, so I need a mouseclick to target it. Possibly ok if i figure out the twitch.

Maybe I can practice with blizzard today. It starts channeling instantly and has similar targeting mechanic. There are some attractive aspects to flame strike on mobs in a close range AOE situation...the DOT isn't capped and they don't tend to move off the DOT if you're already at point blank range.

Does anyone know if you can tell a macro "target an AOE right where I'm standing?" That would be ideal for this kind of approach and wouldn't require me to use the mouse. (or heck, any method of keyboard targeting a AOE would be nice)

I don't think you can macro the targeting reticule, as it requires a click for placement. However, I think you might be better off if you made a simple macro like this:

#showtooltip
/cast Dragon's Breath
/cast Flamestrike

With it macroed like that, you just have to remember to click immediately after casting Dragon's Breath. The /cast Flamestrike in the macro will open up the targeting reticule ready for you to place. I'm not sure how well it will work in practice, but "DB, click" doesn't seem that unweildy. You also have to make sure that your mouse is kinda hovering over your own body.

You'd have to castsequence that and use two clicks. You can't macro stuff with GCD, and unless Dragon breath has changed, that'll end the macro.

But yeah, a castsequence is what I had in mind. It's how I do the close range AOE spam today. I suspect I'll need to just learn the mouse twitch (or give up on firestarter if I just can't get it right). I do have a GCD to get my hand on the mouse or move it if hand is already there between Dragon breath and flamestrike, so it might not be that bad.

A typical arcane spec (53/18/0 or so) has 17 points on effective DPS talents (at least 1% DPS/point), the rest are subpar, fillers or utility.
A typical fire spec (0/51/20) has 44 points spent on effective DPS talents.

Shitty trees produce shitty DPS, it's just that simple.
The only thing that makes arcane viable is a seemingly randomly given inherent +75% DPS bonus for one spell.

Does travel time affect the ignite bug? (fire blast is instant with no travel time. Pyroblast is instant with travel time. I was under the impression that you needed a simultaneous "landing" of the fireball and fireblast to trigger the bug. It seems less likely given the distance-based time gap that Pyroblast has over fire blast. But it depends on whether it's triggered "as it leaves hand" vs "as it arrives")

This progression seems to assume travel time, as you can't cast the instant pyroblast
before fireball 3 crits unless you are assuming it is cast after you finish fireball but before
the fireball hits.

I am not totally sure. I was under the impression that there is a small window of time that makes Ignite bug if there are 2 crits during the window. But, the Instant Pyroblast will leave your hand at essentially the same time as your 3rd Fireball.

Sec 1 - Begin Fireball
Sec 1.5 - GCD Ends
Sec 2 - Still Casting
Sec 3 - End Fireball
Sec 3.05 - Cast Pyroblast

The 0.05 is a made up, but I think the point is clear.

Exactly, you cast the Instant Pyroblast after you cast the 3rd Fireball but BEFORE the 3rd Fireball hits the target. Using the Hotstreak buff before the 3rd Fireball hits opens the door for your next set of crits to count toward Hotstreak.

I haven't updated the non-mage specs yet, mostly because Balance druids are now hell to model and warlocks will see major changes.

Balance Druids
Spamming glyphed Starfire while a glyphed Moonfire is ticking is said to be the mana efficient cycle.
In actual play depending on haste, you will weave in glyphed Insect Swarm and maybe the occasional Wrath if you have the haste/cast time for it.
If you have more mana available, one should switch to mainly Wrath casting for even more DPS.

4250 DPS is a ballpark estimate for Starfire with Moonfire DoT ticking.

Elemental Shaman
I only have Lightning Bolt modeled for them.
As I understood it, it's very beneficial to cast Flame Shocks and Lava Bursts in addition with proper timing.
I also don't have Flametongue weapon modeled, or the new weapon oil, or the most effective glyphs.

4000 DPS is a ballpark estimate for the Lightning Bolt part without Glyph and Flametongue weapon and further effects.

Warlocks
I only have Destruction with Emp. Imp and ISL modeled, and that only roughly.
The addictional crit from Empowered Imp should be modeled correctly, ISL is modeled as "no Life Tap" which the sheet suggests for higher gear levels.

4800 DPS is a ballbark estimate for 7/13/51 Shadow Destruction. I'm aware that that build is outdated and a 26/0/45 cross spec would be roughly 8-10% better. And the top end Destro talents are not in yet.

5100 DPS is a ballpartk estimate for 7/13/51 Fire Destruction. The calculations don't use Conflagrate or Chaos Bolt at all. Maybe F+B neither, not sure.


But so far, caster balance looks okay'ish.
Keep in mind that the above are just rough estimates to give an idea on the relative placemen of classes.


Fire Scaling with Hot Streak/Burnout

The base setting has about 50% fire crit, 3k spell power and 5049 DPS.
It's the full throttle rotation of only Fireball, Living Bomb, Pyroblast.

10% haste get us to 5477 DPS for the cost of 328 rating points. 1.30 DPS/point.

10% crit gets us 5509 DPS for the cost of 459 rating points. 1.00 DPS/point

10% crit while ignoring the added benefit via Hot Streak gives us 5437 DPS, 0.84 DPS/point.

300 spell power gets us 5428 DPS for the cost of 257 rating point equvalent. 1.47 DPS/point.


So barring mistakes that are very possible, crit is still pretty bad. But it's gets less bad from expansion to expansion!

Hot Streak does improve crit scaling by roughly 20%, but haste rating being 65% better for pure DPS makes it too hard to catch up.

It might make a difference when taking mana into account.

You spend about 1/4 of your mana on Living Bomb which doesn't seem to get the MoE mana returns, but MoE then still affect 3/4 of your mana spent. I think MoE didn't work at all recently anyway, not sure.


Hot Streak is the mechanic that finally makes 1% crit more of a DPS increase than 1% haste.

It also gives you good returns at high levels of crit which is important when fire gets more than 30% crit from talents and buffs alone.

If you are fast with the instant pyro it will fly together with the previous fireball, as such they ll both hit simultaneously and bug ignite - provided they both crit. Beta realm's overall latency doesn't allow for that much timely precision atm (i've only managed to see that 3-4 times only), but I assume in live it will be the normal to have both balls of fire flying and landing at the same time. It looks like a frostbolt + icelance really.

He's talking about the overall design of the tree, and the fact that so few of the points you spend there actually improve the tree's performance. The Fire Tree has more than 2.5 points that increase DPS by 1% or more for every 1.0 points the Arcane tree has to provide a similar benefit. That's a pretty significant scaling issue.

No need to get defensive. I don't think the arcane talent tree has feelings that can get hurt.

The arcane tree is a shitty tree. That doesn't mean anything about arcane's playstyle or arcane's value.

If the majority of talent points spent in a tree do not meaningfully increase that spec's DPS (and an Arcane mage's major contribution is DPS, right?) then it's a shitty tree. Compared to pretty much every othe talent tree in the game, the Arcane talent tree does little towards making the Arcane mage better.

He's talking about the design of the tree, not the playstyle. And he's correct, by the numbers that affect dps >1%.

So what's the justfication for keeping Ignite as a DoT in the next expansion? Surely the documented TBC bug where a quick second crit after an initial crit causing an overall loss of Ignite damage and the fact that we lose damage all the time on trash mobs when the mob dies before our Ignite DoT finishes ticking would mean that it's better to change fire crits to be front-loaded. Has this point been brought up and discussed on the official beta feedback forums? If we're not even allowed to roll our own ignites in a manner blessed by Blizzard, then there's no reason to retain the DoT mechanic in Ignite, is there?

Threat. A front-loaded crit will cause threat to ramp up much faster. In a raid, it's better to have the ignite, even if buggy. (And who cares about trash, we're all about bosses). The vast majority of your damage will not be bugged out.

The Arcane tree, to me, seems like it's filled mostly with "effect and defensive" talents rather than DPS improving talents. I wonder how worthwhile the additional threat reduction is in terms of actual game DPS. Sometimes, things like that can tip the scales in favour of Arcane, despite being contrary to what we theorycraft.