The question of choosing a profession was discussed like a gazillion of times, e.g. here: viewtopic.php?t=12507 or there: viewtopic.php?t=12596 but here I'd just give a summary.

The main question is: What are the best professions combo for a raiding shadowpriest from the pure dps standpoint? Idd, herbalism allows you to farm mats for (then effectively free) flasks and skinning provides gold for repairs or consumables, but these benefits are not of interest for us now.

tl;dr version
Raiding 25-mans -> LW + Tailoring or Enchanting
Casual PvE -> Tailoring + w/e
Other cases -> w/e + w/e

tl;dr explanations
Here, despite of all flaming and swearing we came to the following conclusions:
#1 For end-game LW is superior to any other profession bringing 40+ raid dps in the best case of dps group or providing additional 20 mp5 for a group (you can benefit from only one effect at a time).
#2 Leatherworking brings trouble about rerolling it, about convenient group, about losing flexibility in using tailored gear or not needing to reroll enchanting each time you upgrade one of rings. Actually, rerolling enchanting for ring upgrades might be not hard as it seems. Look here

how to reroll LW
If you are enthusiastic about LW, but do not know what to begin with and what to do, check Leveling leatherworking F.A.Q.

how to reroll Tailoring
If due to some reasons you don't have tailoring but need it for Sunfire Robe or whatever else, but do not know how exactly to do this, I've composed a spreadsheet describing mats required and skillup order for getting 365 tailoring (getting 375 skill is pretty much repeating 360->365 skillups). You can download it from mediafire.

Summary
Shadow priests have the following breakdown of raiding professions usefulness for raiding. At the very start of raiding combo #2 is actually even worse than #3 and #1 is insanely good in comparison with both. However, as gear improves #2 becomes really close to #1, and #2 appears to be better and better than #3:
#1 LW + Tailoring
#2 LW + Ench
#3 Ench + Tailoring

Additionally, for the very end-game, ones who can afford releveling enchanting after changing the ring, might use alternatives to #1, if corresponding pattern drops, namely, LW + Engineering or LW + Jewelcrafting (former being better). However, these combos, while being situationally better than LW + Tailoring, in the very long run perform much worse.

Leatherworking (provides one of two insane party-wide buffs of your choice; evaluating these effects is still in progress)
Drums of Battle (recipe bought off NPC), a lw-only usable consumable provides party-wide static 20 spellhaste at the cost of GCD for activation. GCD costs us less mana returned from VT, and this means degrading of party mages'/locks' performance. Currently believed to be worth +7-10 personal pseudodmg per each dps in your group, not including yourself.
Drums of War (recipe bought off NPC), a lw-only consumabe that provides group with 15 static AP and 7.5 static spelldamage at the cost of GCD for activation. Clearly subpar to drums of battle.
Drums of Restoration (recipe bought off NPC), a lw-only usable consumable that replenishes 600 hp/mana over 15 sec each 2 mins. This equates to additional 25 mp5/hp5 at the cost of GCD for activation, which bring us to 20 mp5 for all group. Currently believed to be worth +20 mp5 for each mana-user in your group, which can be usable in mana-heavy fights and mana-lacking groups.

Tailoring (+13 pseudodmg in the long run, up to ~+70 pseudodmg in casual case)
BoP craft Sunfire Robe (BoE recipe, SWP trash drop). Worth extra +13 spellhaste, +14 spellcrit, +1 dmg, +- some stats, -31 spirit over M'uru sunmote trade-in.
BoP craft Frozen Shadoweave Robe (recipe bought off NPC). In T6 content beaten only by T6 robes and Vashj's Vestments.
BoP craft Mantle of Nimble Thought (BoE recipe, BT/MH trash) is on-par with best shoulders of T6 content.
BoP craft Frozen Shadoweave Shoulders (recipe bought off NPC), damage-wise pretty equal to T6 shoulders or Mantle of Ill Intent.
BoP craft Frozen Shadoweave Boots (recipe bought off NPC) - best boots of T6 content, closest competitors are Footpads of Madness (Jan'alai) and Slippers of the Seacaller (High Warlord Naj'entus), both are -5 pseudodamage away.
Spellstrike set 2pc bonus, a tailor-only proc, is worth +17 pdmg assuming both pieces of set are equipped

Jewelcrafting (+5 pseudodmg in the long run)
BoP craft Pendant of Sunfire (BoE recipe, SWP trash drop). According to pocketpriest it provides +2.71 dmg (due to spellhit lost in comparison with Twins' necklace).
BoP gem Don Julio's Heart (recipe bought off NPC) provides another +2 dmg.

Enchanting (nothing rly in the long run, +24 pseudodmg)
2x +12 spelldamage self-enchants for rings (recipe bought off NPC). However enchanted rings can be used even when enchanting is dropped.

Engineering (nothing rly in the long run, +17 pseudodamage before Kil'jaeden's loot)
BoP craft Annihilator Holo-gogs (BoE recipe, SWP trash drop) is inferior to Helm of Arcane Purity (Kil'jaeden's sumote trade-in), however it's +16.43 pseudodmg better than best pre-KJ loot.
BoP craft Destruction Holo-gogs (recipe bought off NPC) is slightly superior to T6 head w/o counting in set bonuses. However, it gets beaten by Hood of Hexing (Hex Lord Malacrass).

Alchemy (nothing at all)
BoP craft Sorcerer's Alchemist Stone (recipe bought off NPC) is sub-par to at least two easily attainable trinkets Icon of the Silver Crescent (41 Badges of Justice) and Timbal's Focusing Crystal (Heroic MgT, Priestess Delrissa).

Evaluating benefit of Drums of Battle

There are two thingies about drums:
- Personal effect: [+spellhaste][-gcd]
- Group effect: [+spellhaste][-dps lost due to loss of VT mp5]

As it is shown here loss of GCD of personal DPS is compensated by gained spellhaste, tho estimating personal effect isn't of our interest. However (thanks to pocketpriest for noting this) we DO need to account the lost manaregen to the group. It'll be: gcd/120*dps*0.05*5s = say, 1.5/120*1500*0.05*5 = 4.68 mp5. Assuming lower gcd will bring us to a fair compromise point of 4.5 mp5.

According to pocketpriest (need confirmation, cuz this sounds unbelievably high for me!), an end-game shadowpriest provides 400 extra dps for a lock and 300 extra dps for a mage. Assuming, 1500 dps shadowpriest, it's 375 mp5. Thus, 1 mp5 translates to 1.1 dps for a lock and 0.8 dps for a mage (to reiterate, unconfirmed info). Thus, 1 personal dps translates to 0.3 dps for a lock and 0.2 dps for a mage.

Regarding group effect. Starting from +1100 spelldmg, 1 spellhaste is very close to 1 spelldmg for locks and mages of raiding specs, and this triple or quadruple gain in dps from spellhaste might be very valuable.

Further calculations heavily depend on group composition and gear level of the raid. Let's study two cases:
#1 Casual gear, not so good grp composition 1x spiest (+1100 spdmg, hit capped), 2x shadow destro locks (+1100 spdmg, hit capped, 0 crit), 1x fire mage (same as lock)
#2 Awesome gear, good grp composition 1x spriest(1500 dps), 2x shadow destro warlocks (1400 buffed spd, hit capped, 400 spellcrit), 2x fire mages (same as locks).
#3 An occasional hunter or two in your grp In this case, I will dare to suppose that hunters will benefit as much as mages and locks from added spellhaste, tho I have no TC to ground this.

#1 Casual gear, not so good grp composition
locks gain 2x20x0.784 dps = 23 dps
mage gains 20x0.512 dps = 10.24 dps
---
total = 33.24 gained dps
at the cost of 4.5 mp5 to all the group, which is equivalent to 2x4.5x1.1 + 4.5x0.8 dps loss = 13.5 dps lost due to lacking mp5
---
total = 19.74 raid dps
amount of pseudodmg to provide the same effect on raid dps = (19.74 / 0.5) / (1 + 2x0.3 + 0.2) = 22 pseudodmg

#2 Awesome gear, good grp composition
locks gain 2x20x0.784 dps = 31.36 dps
mages gain 2x20x0.732 dps = 29.28 dps
---
total = 60.64 gained dps
at the cost of 4.5 mp5 to all the group, which is equivalent to 2x4.5x1.1 + 2x4.5x0.8 dps loss = 17.1 dps lost due to lacking mp5
---
total = 43.54 raid dps
amount of pseudodmg to provide the same effect on raid dps = (43.54 / 0.5) / (1 + 2x0.3 + 2x0.2) = 44 pseudodmg

References:
http://wiki.shadowpriest.com/index.php? ... aling/Mage
http://wiki.shadowpriest.com/index.php? ... ng/Warlock
http://wiki.shadowpriest.com/index.php? ... ing/Priest
pocketpriest's info about mage/lock dps gains due to VT mp5

Evaluating benefits of Drums of Restoration
Drums of Restoration replenish 600 hp/mana over 15 sec each 2 mins. This equates to additional 25 mp5/hp5 at the cost of GCD for activation, which bring us to 20 mp5 for all group.

And in case of 2x locks, 2x mages in the group this will equate to 2x 20x 1.1 dps+ 2x 20x 0.8 dps = 76 dps. That is much more than drums of battle provide us! Definately, pocketpriest's info needs to be thoroughly investigated

Evaluating tailoring for casual raider

If you cannot afford getting cutting-edge gear this does not seriously affect following professions: since LWing effects remain mostly the same, enchanting bonus is static, JC benefit minorly degrades, alchemist trinket can still be easily replaced, engineering can't afford t2 gogs, but t1 still gets beaten by Hood of Hexing. There's only one profession which value changes drastically: Tailoring. Let's have a closer look - find easily attainable counterparts for tailored items (badges, common crafts, ZA drops).

Sunfire Robe gets removed due to being unavailable outside SWP guilds for good 2-3 months). Same for Mantle of Nimble Thought since the pattern is not that common. Let's compare Frozen Shadoweave with 100% drop counterparts. For simplicity sake, let's assume that we socket everything with +12 spd gems available for 10 (or 15) badges at 100% drop rate. For the same reason intellect contribution to pdmg is ignored.

Shoulders
FSW shoulders = 50 shadow dmg, 2x sockets = 74 pdmg (pseudodamage)
Quest item = 40 spell dmg, 18 spellcrit, 0x sockets = 43 pdmg

Chest
FSW chest = 72 shadow dmg, 2x sockets = 96 pdmg
2.4 badges chest = 62 spell dmg, 50 spellcrit, 1x socket = 84 pdmg

Boots
FSW boots = 57 shadow dmg, 2x sockets = 81 pdmg
2.4 badges boots = 47 spell dmg, 1x socket, 23 spirit = 61 pdmg

Spellstrike set bonus
Spellstrike set 2pc bonus, a tailor-only proc worth of 17 pdmg according to SimCraft.
However, replacing Spellstrike Pants = 46 spell dmg, 26 spell crit, 3x sockets = 86 pdmg with 2.4 badges pants = 62 spelldmg, 43 spell crit, 2x sockets = 93 pdmg will net into +10 pdmg bonus.

Total benefit +73 pdmg. Please, note this is really a very worst case of drops, and the more you raid, the more u attend ZA/heroics, the less will be the gap between tailored and badges/quest stuff.

Evaluating Pendant of Sunfire benefit


Evaluating professions combos ranking

The absolute best combo is excellently described by SumFortis:


Actually, rerolling enchanting for ring upgrades might be not hard as it seems. Look here

But what if you don't have gold/time/desire to relevel enchanting+leatherworking each time you get your ring upgraded? Then you need to resort to suboptimal but less painful options, which are:

Tailoring + Enchanting
Traditional approach (I bet like 90% of raiding shadowpriests had, if not currently having, this setup) that complies common understanding of cloth caster's professions.

Leatherworking + Enchanting
Approach that became popular not so long ago, after a major spellhaste buff, tho its benefits were significant at least after first spellhaste buff (patch 2.2). However, the major hype about LW started shortly after 2.4 hit live.

Here's an objective comparison of previous two combos strictly for raiding:
T/E benefits: Player gains access to a best-in-slot robe and several best available casual gear pieces (+13 pseudodamage bonus for SWP raiders and +10-+40 bonus for casual raiders).
LW/E benefits: As it can be easily calculated, if you're residing in caster group, benefit of +20 static spellhaste by far overweights even the best estimate of tailoring benefit.

However, there might exist subjective reasons that players have when refusing to reroll leatherworking. They are not "lol? excuses, nab!", they are just subjective reason ppl might have. After all, most of us play WoW, not a minmaxing "WoW: Raiding Sunwell Plateau" game. Possible motivations:
- Leatherworking will be useless in WotLK.
- I have lots of rare recipes for tailoring/alchemy/blacksmithing and I totally cannot drop it.
- I'll get 2x RoAKs and drop enchanting. Well, maybe, not a second RoAK, but LoFP. Well, maybe a new, currently unknown, caster ring from Sunwell. In short, maybe in indefinite future.
- Cba.

That calculation was a fairly meaningless (but I thought interesting technical exercise - though I'm strange like that!) to compare the lost GCD to the drums benefit to the shadowpriest user only.

The overall benefit of the drums is 5 times as big because of course it affects all 5 party members and not just the drum user.

The net result is that LW comes out far ahead of any other tradeskill in terms of raidwide benefits provided.

Thanks for making an accent on this for me. However, I edited the original post to put an accent by myself.

Was
Became

Totally lolworthy, well done ^_^

yay, I got quoted

TC got a major update, and went through some cleanup (almost 90 edits of perpetual beta, lol). Changes are worth to check the topic again.

tl;dr post updated with a link to LW rerolling guide (based on my experience, nothing phenomenal, just to show that rerolling is swift and cheap).

Info about leatherworking got updated. Now its benefit is measurable in the terms of personal pseudodmg:


So, in case of end-game gear and good group composition (1x spriest, 4x mages/locks), LW will benefit your raid as good as ~+40 of your pseudodamage would do.

Note, this conclusion is based on unconfirmed TC about mp5->dps conversion for mages and warlocks. It seems for me LW benefit is underestimated, so I'll try to check that TC later.

Info about tailoring got updated, since I forgot to mention tailor-only 2pc spellstrike set bonus. So, now

Was:

Not really sure how this works. On average drums provides 20 haste to every caster dps in your group.

For caster classes 20 haste ~ 20 damage (in fact it is a little better for mages/warlocks with top end gear). Seems a pretty straightforward relationship so I don't understand why it is being reduced to 7-10.

#1 So, e.g. for a group of 2x mages, 2x locks our 1 pseudodamage translates into 0.5 personal dps + 2x 0.15 mages' dps + 2x 0.2 lock dps = 1 raid dps.
#2 From the other hand, 1 spellhaste for a mage/lock translates into ~0.7 raid dps.

Facts #1 and #2 along with taking into account mp5 lost from GCD of activating drums result in only +7-10 pseudodamage per dpser instead of seemingly obvious +20 pdmg.

Okay, I understand how this works now. The damage boosts from tailoring and enchanting (and other tradeskills) will boost our mana return as well as our damage. However the damage boosts from LW will only boost damage (because they are effectively boosting the damage of other party members rather than our own) and this needs to be taken into account.

I'm not sure I agree with the calculations for the impact of a shadowpriest on warlock/mage dps. While the overall number of 300/400 extra dps seems broadly reasonable, I think that shadowpriests provide excessive mana return to warlocks/mages.

As you calcualted, a high end shadowpriest provides 375 mp5 to his party. I think that the first, say, 200 mp5 helps boost mage/warlock dps most of the numbers listed.

Beyond that point, the additional mana provides much smaller benefits because the casters will already be casting at pretty much full capacity. So the fact that LW loses a bit of the top end mana regen compared to other tradeskills is probably not much of an issue.

Yeah, I agree, but have no theorycrafting to ground such conclusions. I'll stick to existing numbers until I can provide any TC. Tho, your version seems to me more reasonable than pocketpriest's.