My understanding is that there's a scaling effect in play here. At low gear levels - dungeon blues and the like - rating stats are very weak. However, as you gear up, they appear to scale pretty rapidly. So right now, strength may be a little better, but with a few weeks of gearing with raid epics, mastery should reclaim the lead.
That said: it hasn't yet been determined *why* rating stats are scaling so quickly relative to the historical norm. It could have to do with balance changes that have been made - the weapon damage rebalance being an obvious suspect - or it could be indicative of a bug in the modeling. So take this analysis with a grain of salt.
Before being able to run heroic 5-man content I reforged hit and expertise to 6-6.5%. When I started running heroics I reforged to 7.5% hit and expertise and saw horrendous DPS.
After reforging to 7.5% hit, then mastery and then haste if mastery was not possible and taking the expertise that was left I ended up with ~4/5% expertise and ~4% more mastery than I had previously.
On the training dummy I saw my DPS double. This was at iLevel 450 with a 450 MH and / 440ish OH (green quest dagger).
My thinking is that at this lower iLevel, especially with the low damage weapons that mastery scales better than expertise - in other words poison damage gained was much higher than the direct damage lost via misses.
I don't have logs for this and my dummy sessions were short - one full cooldown of shadow blades - just to see where the DPS was settling. Obviously this does not take into account execute damage, but for me it was enough for the short fights in heroic 5-man content.
I am going to revisit this now that I have 460+ gear with logging and better analysis, but I wanted to give my informal experience.
Apparently I must be confused if we are supposed to open with Garrote. My typical opening is:
[Stealth]Mutilate (Shadow Focus)
Dispatch(if proc'd and at 4CPs)
Dispatch(if proc'd from before and not used)
Normal Rotation 4-5cp envenoms while keeping Rupture up.
The more I think about it, I should probably use SnD first before Mutilate due to our overwhelming amount of damage from passive, but I've gotten so used to this it's automatic. Until I see math that says otherwise, I'll continue to do this rotation.
Opening with Mutilate (unless I am overthinking this) gives:
2x Chance to proc Poisons
Chance to proc Blindside
Chance to give 2-3 CPs
Garrote would start energetic recovery and delay our first Rupture - allowing us to get SnD up a little earlier in the opening. Does this outweigh the benefits of an extra CP, Blindside Proc, and Poisons? (all not 100%, of course..)
I had been using Garrote to open for a while, but have since switched to Mutilate. Here are a couple points as to why:
Garrote's damage is somewhere in the 15k-20k range, give or take (and depending on gear). That's about as much damage as your mutilate's main hand. At least it is for mine. At any rate, the entire mutilate is going to be higher damage.
Mutilate produces an extra combo point naturally, and has a chance to produce a third.
Both cost the same energy: 0.
Garrote ticks every 3 seconds as opposed to every 2 seconds of Rupture. So your energy regen is not as strong. Even if you started with garrote, prioritized a quick SnD, and then went for a 5 CP rupture, you may only be looking at 1 tick, 2 tops. Where as if I started with mutilate and rushed a 5 point Rupture, I could have that up in 3 global cooldowns. Prioritizing SnD, I could have it up in 5.
Ambush hits a lot harder for me than both garrote or mutilate and also generates 2 or 3 combo points. The only downside is one less attack to apply poison, but ambush still hits harder for me even if you assume off hand mutilate gets an extra poison proc every single time.
I have been following the Simcraft opening... Ambush -> SnD -> Mutilate -> Rupture to get everything going, and then into normal rotation.