I was just wondering what gems people are using for the red,l yellow, and blue slots when it is worth getting the gem bonus. Also what to gem when the gem bonus sucks?

Using the weights from this guide, any socket bonus (except +60 crit and +60 mastery) that requires a blue gem, is worth socketing Glinting for. Provided of course that you are still able to reforge yourself so that you do not end up with a surplus of Expertise or Hit.

Using my current weights (taken from simcraft, epic 476-489 and some 463 blues), any +60 agility and any +120 rating socket bonus that requires a blue gem is viable. Again, provided you are able to still reforge to the optimal Hit and Expertise values.


As a rule of thumb, all blue gem socket bonuses are okay to take, except +60 in any rating. Using the weights from this guide, +60 haste does become viable however. Also make sure that you never grab any +Hit and +Expertise socket bonus that will take you over the suggested optimal values after reforging.

Using the weights from this guide, I'm pretty sure that Glinting (80Agi/160Hit) is inferior to Lightning (160Haste/160Hit). Also, saying that "+60 in any rating" is not worth going for is a blanket statement that is far from accurate. Since Blue sockets invariably mean gems with hit, that stat being capped or not should be the deciding factor. If it isn't capped, it's practically universally worth it to go for blue; if it is, it almost never is. The actual bonus itself will rarely matter (mostly if it's exceptionally large).

Would it be worthwhile to spec into Shuriken Toss and use it during Deep Insight instead of SS? It costs less than half the energy of SS, has no range, and generates one CP (it gets no benefit from RS, though).

Stat priority is: Agility > Expertise (7.5%) > Melee Hit (7.5%) > Haste > Mastery > Crit

Stat Summaries

Agility
Increases damage dealt by your abilities and slightly increases critical chance.
Expertise (7.5%)
The Expertise cap of 7.5% (2550 Rating) will eliminate your attacks from being dodged. Having this cap and always attacking behind your target is crucial for avoiding high DPS loss. To check your Expertise, open your character panel and then check under the Melee tab for "Expertise." Make sure it is at or just above 7.5% for your weapon(s).
Melee Hit (7.5%)
The Hit cap of 7.5% (2550 Rating) will eliminate the chance for any special ability (including poisons) to miss against a boss. This is the cap regardless of weapon type. Special abilities missing is the easiest way to lose DPS, so always make sure you have this cap. To check your Hit percentage, open your character panel and then check under the Melee tab for "Hit Chance" and make sure it is at or just above 7.5% for your weapon(s).
Haste
Increases attack speed and Energy regeneration.
Mastery
Mastery: Main Gauche increases the chance for your main-hand attacks to grant an extra attack for additional damage.
Crit
Increases the chance for your attacks to critically hit for 200% damage.


Taken from: Combat Rogue PvE Stat Priority (MoP 5.0.5)


Now with said information,

Reforging:

Make sure your hit and expertise are both at 2550

After that try to reforge into Haste, if you cannot then do Mastery.
Also, always preferably reforge out of crit.


And that is my two cents.

When we reach 'Deep Insight', do we apply Rupture or do Eviscerates?

Also, I believe that we should not build additional Anticipation stacks even in Deep Insight since Anticipation stacks do not reduce further the CD of Rush and KSp. Obviously, theroycrafting can prove otherwise.

1. Arguably, both. If memory serves from previous discussions, the damage modifier of Deep Insight applies only for the duration of Deep Insight. Meaning, if you apply Rupture with two seconds left of Deep Insight, it'd be a bad idea, and Eviscerates' front-end damage would be preferable. However, if you're able to weave in Rupture to fit the duration of Deep Insight, it'd be as much of a relative dps gain as it would to weave it in during normal rotation.

2. I'm not quite sure I understand what you mean with your remark about Anticipation charges. You don't use Anticipation charges to cast finishing moves; you use them to convert to Combo Points that you, in turn, use to cast finishing moves. Restless Blades reduce the cooldown of Adrenaline Rush, Killing Spree and Shadow Blades (among others) with each damaging finishing move based on how many combo points you use to cast it. To that end, Anticipation Charges get converted to Combo Points that get converted to a finishing move that reduces the cooldown of the above mentioned abilities. In that sense, Anticipation charges do reduce the cooldown of Adrenaline Rush and Killing Spree if you use them on a damaging finishing move.

I just wanted to restate your comment about Deep Insight because when I first read it I thought you were saying the opposite of your final conclusion. The % dmg buff from deep insight is dynamically updating unlike how many other damage buffs have behaved in the past. (I recall tricksing my warlocks back in the day so they could roll the same dmg buffed corruption for the whole fight). That is to say, the damage modifier from deep insight will only apply to rupture ticks that occur while deep insight is up. Once deep insight falls off, there will be no further damage modifier to ticks of that same rupture cast. I am unsure how many deep insight ticks of rupture you need to be able to get before it is better to rupture in deep insight rather than eviscerate.

That's how it used to work in Cata, but it's different now. Rupture's damage is determined by our insight level when it is applied and it does not dynamically update until/unless it is refreshed. So if you apply rupture during deep insight then all ticks will benefit from deep insight, even those that happen after deep insight has faded. There is proof of this in the Mists Mechanics Testing topic, post 38 by shadowboy813. His post was specifically talking about how rupture interacted with clipping, but his screenshot of combat logs show that rupture cannot possibly be dynamically updating as insight changes: Looking at his first screenshot where he starts with a no insight rupture and later clips the last tick with a deep insight rupture, we see that every tick of rupture up until the refresh does the exact same amount of damage when we know that his insight is increasing during this time (ie if it was applied at no insight and then refreshed at deep insight then he must have passed through shallow and moderate insights during the duration yet the rupture ticks did not change). Further, after he applies the deep insight rupture it continues to tick for the same damage for the entire 24 second duration when deep insight only lasts for 15, which means that deep insight had to have fallen off during that duration yet again the damage did not update with it.

This *should* mean that insight has no bearing on the decision of using rupture vs eviscerate, because regardless both will get the same multiplier at that given time (ie if you're in deep insight then you're choosing between a 30% buffed evisc or a 30% buffed rupture, during moderate insight you choose between a 20% buffed evisc vs a 20% buffed rupture, etc). Since they both get the same multiplier the multiplier won't determine which to use, if Evisc is better then it's better regardless of insight, if Rupture is better then it's better regardless of insight.

Re-tested and confirmed. I fiddled with the combat log, but for some reason I couldn't get it to merge the application of insight levels. Anyhow, I hope you'll take my word for the following: This is a statless Rupture with no buffs or auras aside from Deep Insight. Rupture is cast some way into Deep Insight, which is up for the first three Rupture ticks and then fades. However, this change has no bearing on the remaining Rupture ticks.

To that end: Rupture isn't dynamically updated based on Deep Insight fading; and so, whatever decisions you make regarding prioritizing your abilities when Deep Insight isn't present, should be the same decisions when it is.

I think you are misunderstanding how anticipation works. The charges of anticipation that we gain don't affect the finishers at all, and aren't consumed by them, the combo points are basically being 'stored'. So an evis at 5cp and 0anticipation will do the same amount of damage as 5cp with 5anticipation. Whenever you cast a finisher it takes your anticipation stack and turns it into combo points on the target. This allows us some room into when we want to cast our finishers, as well as them always being 5cp and full use out of SS's chance to produce an extra cp (casting SS at 4cp without anticipation has chance a to waste a cp). So in the end, whether you are delaying your casts with anticipation or using your combo points immediately will have no effect on the number of finishers.