[Edit]: Apparently, I misread the new meta gems requirements. The previous post is now outdated. I'll let it remain in spoiler tag, if anyone wants to follow the old thought for their own information.

The Chaotic meta gem now requires more blue than red gems. Ouch. That's actually enough to make the Ember meta more desirable in many situations. If you want to use a Chaotic meta, you should gem one green in a yellow socket, then pairs of green/red gems in yellow/red sockets (pick those yellow sockets with the best bonuses first) and then fill the rest with purple gems. The gem parts being red = int, blue = hit, yellow = best of crit/haste/mastery. The details, break points and reasons are listed in detail in this post.


Chaotic Shadowspirit Diamond - Item - World of Warcraft - more blue than red.
Bracing Shadowspirit Diamond - Item - World of Warcraft - at least 1 blue, 1 yellow.
Ember Shadowspirit Diamond - Item - World of Warcraft - at least 2 yellow.
Fleet Shadowspirit Diamond - Item - World of Warcraft - at least 2 yellow.

• The run speed meta is not good, since it has mastery and you can gain run speed on your feet at a loss of 15 mastery. 54 intellect easily beats 54+15 mastery and similar combinations of ratings.
• The intellect meta gems can be activated by intellect spam and 2 adjustments. The current gear allows that while even gaining DPS from socket bonuses.
• The Chaotic diamond is painful to activate and requires some planning, but it's not too bad. Assume you have R red, Y yellow, B blue sockets on your gear.
  Fill Y yellow sockets with green gems, Y-1 red sockets with red gems and the remaining sockets with purple gems. This is the maximum you can get, since "rating>intellect" is an implicit constraint (green + red = 2*purple in that regard). Now, the above process fails if you have at least 2 more yellow than red sockets (R<Y-1, R<=Y-2). Call D=Y-1-R the amount of yellow sockets that cannot be paired up with reds. You can fill them with purple gems, or fill them pairwise with orange+green gems, trading 20 int for 20 rating + 1-2 socket bonuses. As those socket bonuses will be weak since we fill sockets with strong bonuses first (green/red combo), those remaining sockets are likely not worth it.

So, let's look at what gear is available to mages right now: Items - World of Warcraft
4/5 T11, non-set Headpiece and the rest of the good gear will give us 4 red, 4 yellow, 2 blue sockets.

Meta	Gems	Bonuses	Total:	Intellect	Rating
Ember Shadowspirit Diamond - Item - World of Warcraft	Red-; Red/Purple+; Red+; Red-; Purple+; Orange/Red+; Red+; Red-	70 intellect, 20 haste		464	80	40 hit, 20 haste, 20 choice
Chaotic Shadowspirit Diamond - Item - World of Warcraft	Green+; Red/Purple+; Red+; Green+; Purple+; Green/Red+; Purple+; Green+	80 intellect, 10 mastery, 40 haste		260	324	140 hit, 10 mastery, 40 haste, 54 crit, 80 choice

(Gems are listed in the order in which they appear in the link. +/- indicate getting/losing the set bonus. Also, the +20 haste belt bonus for the chaotic meta is likely too high.)

So, let's simplify this a bit. With Lhivera's numbers, 3 rating = 1 int (for haste/mastery, and hit can be traded after the cap. Our combinations then become "464 int + 80/3 int + 2% max. mana" and "260 int + 324/3 int + 3% crit damage", with the the crit damage now being a flat 1.03 multiplier for us (we have a X1.33 crit damage miltiplier instead of our old x2 crit bonus) and 120 intellect (the difference) being a 2% damage increase by SP/Crit alone. The Chaotic 3% can only be a damage increase if 2/3 of your damage is crit damage.

Not that this is at your current gear levels with the current amount of gems. When our gear levels increase, the value of ratings will likely increase (not as much as I initially thought though, since intellect and rating levels on gear are mostly the same), but the amount of gems will likely increase as well, which increases the loss from gemming.

• Arcane at 33% crit has 50% crit damage, 50% noncrit damage. So the Chaotic meta part adds 1.5% damage and should not be used. The Ember meta gem adds more damage through intellect alone, and grants extra mana through intellect and the 2% max mana bonus which synergises with burn cycles and Mana Adept as well.
• Frost is a little tougher. About 50% of the damage are DF/IL crits. The remaining 50% are FrB/FFB/FO, half of their damage being crits at 33% crit (soft cap target). So we have 75% crit damage and the Chaotic meta adds 0.75*3% = 2.25% damage, while the Ember meta adds 2% damage. This should put the Chaotic slightly above the Ember diamond since mana does not seem to be a concern.
• Fire has a crit multiplier of ~3 with Ignite/Flashburn. Assuming 40% crit, about 2/3 of the damage is crit, so the Chaotic diamond adds 2% damage, just like the Ember diamond by SP/Crit alone. There is some more benefit from the extra mana through intellect and the additional 2% max mana since you are limited by mana/required to Scorch.
• Since the given numbers are all estimates, we cannot make any reliable predictions yet. Arcane will probably err towards Ember, Frost possibly towards Chaotic, but they seem to be in similar ballparks. For Fire, anything I'd say could be a guess.

So, with the new "More blue than red gems." requirement, the value of the Chaotic meta takes a huge drop at level 85. The main reason being that intellect is such a powerful stat compared to the other ones. The second reason is that there are a decent amount of sockets available, 10 in the gear set above. The more sockets there are, the higher the loss for the Chaotic meta. If the amount of sockets should increase in higher tiers (like 3 in robes/leggings, 2 in shoulders/gloves/etc.), it requires even more spell power for the Chaotic meta gem to break even. Ironically, when you have just 2 or 3 sockets in weak gear, the Chaotic meta is significantly ahead until you get to tier sets.

[Addendum]: Newer numbers from World of Warcraft: Cataclysm Beta - (English) Forums -> Mage 2 Cont (Till Blizz extends megathread) bring the Chaotic and Embe closer togehter. Also, the poster below was correct, I actually messed up the numbers. They should be mostly fixed now, though.

I was quite busy this week, so I just quick-edited the OP without mentioning it in the thread itself. I just quote the current "Latest Changes" here, that are mostly the topics discussed in the latest posts. An exception is the new Raidbuff overview - make sure to have a look at this one, too. Correct me if I'm missing any caster relevant buffs entirely or in a specific Spec.


Edit: The Spell Coefficient table is still outdated. If someone has time to spare...

Regarding Tailoring vs other professions, it's important to keep in mind that the +int professions will also benefit from Blessing of Kings or MoTW in nearly every raid while Tailoring won't. With this in mind and adding Wizardry, if comparing a profession that adds 80 int and uses the 50 int cloak enchant, that profession would provide 130*1.05*1.05=143.325 int compared to the Tailoring proc. Tailoring would provide roughly 1.7 more spell power (145 vs 143.325) assuming the proc is up exactly 25% of the time, but it would lose out on 143.325/648.91=0.22% crit and 143.325*15=2149 mana.

Looking at it from that point of view, Tailoring looks like a weak profession compared the other ones, however Tailoring benefits from the proc coming early on in the once per minute that it can go off. In other words, on a 40 second fight you're pretty much assured that the proc will be up 15 of the 40 seconds or 37.5% uptime. On a 1 minute, 30 second fight you're likely to have it up for 30 of the 90 seconds or 33.3%. As the fight gets longer the actual uptime falls closer to the 25% mark but it's more likely that it'll come in just above 25% than it might fall below that mark. There is also factoring in the possibility of Lightweave being stacked with other cds, for example if happens to proc during a Bloodlust or is used with Combustion or Arcane Power it's value would be raised but if not used with those buffs it's value would be reduced.

Unless my math is off, it seems like Tailoring is slightly weaker than an 80 int profession in general, but has the potential to be stronger as fights get shorter and depending on how well it can be stacked with other cds.