It would not surprise me if mutilate offhand swing cannot be dodged/missed/parried by design. But whenever a mainhand swing is miss/dodged/parried the offhand swing follows. This would be a very easy computational way and works with everything we are seeing. You could test this theory with enough mutilate swings as well. If your offhand has a slightly lower critrate than your mainhand with mutilate, this would explain it.

If it was a two roll system (i.e. first check if attack lands, then check if it is a crit or a hit), you would lose a significant portion of your crit.
I even made a simple PHP script to "prove" this: http://sp00n.pytalhost.com/wow/attacks.php
Now of course you could program something that accounts for this, but why make it more complicated than it has to be? Assuming that there are thousands and thousands of combat events every second on each server, every single modification would have a huge impact on the performance of the server.

I figured out an easier way to show what I'm saying, and I'm hoping it's understandable.

What you're saying is not what I'm saying. I'm saying the table is rolled for avoidance, and then that number is used to determine the hit type. This helps to also explain how glancing/crushing blows happen, since they would be a constant on the hit table after crits. Regular hits are the area between glancing and crit. In the event of higher skill level, glancing or crushing is determined simply by the difference. Obviously, the combat mechanics are more complicated than many believe, because there has to be a check of weapon skill vs level to determine the diminished (or added) damage as a constant on the bottom of the 'hit' area. It's not two seperate rolls,

If Parry gets taken off the table, that just expands the hit area, which that number is then taken against. Since I haven't done any recent combat logging against a wide range of mobs lately, I'll have to wait until I get some raids in (tonight, Saturday) to get the data to show this in the working. It utilizes the 'one roll' idea, but shows the second table. Rather than crit being 'fixed' into the first table, it's part of an 'expanded' hit table, meaning that the more attacks you hit, the more attacks you also crit so that the percentage stays the same. Otherwise, when you get more +hit, you would lose crit percentage, not maintain it. Basically, using my table above, if 1000 swings were done, 240 miss, 50 get dodged, 50 get parried, 211/212 get in as crits, 185 get in as regular hits, and the last 263 get subjected to glancing/crushing rules as appropriate. It explains why critrate is constant, because the more you hit, the wider range of attacks actually crit, and it also explains why it appears that 'crit' eats 'hit', because the 40% constant of glancing is there, just undetectable in non-level-divergent combat. This also shows why having too much crit without enough hit means you lose crit, because the constant 40% eats into the only other 'thing' on the table, crit.

Maybe I'm saying the same thing everyone else is, I'm not sure, but this 'expanded' table idea is just how I see it working in my head. No secnod roll, but the same roll is taken across to determine it's 'true' effect.

I can't see the advantages of your system.
That is, I don't understand your system. ;)

It may just be that my way of seeing the system is just a different method of modelling it.

Basically, it's just a way of further explaining and allowing for the different combat mechanics we have in the game. If someone rolls, say, a 7885 for a hit, that would fall under the 'glancing' area of my table, since 3399-5512 would be a crit, 5512-73659 would be a regular hit, and 7360 to 10000 would be a hit subjected to glancing/crushing calculations.

The thing is, "meaning that the more attacks you hit, the more attacks you also crit so that the percentage stays the same." is *wrong*.

The more attacks you hit, your crit percentage (defined as crits per swing that hits or crits, the way most parsers define it) goes *down*. This has been demonstrated experimentally. The number of crits you receive is constant across the number of swings, not across the number of hits. In effect, your table *is* a one-roll system, you just use different rules for populating it, rules which don't match experimental evidence. It isn't 40% of connecting swings that glance; it's 40% of *all* swings.

Assuming Mutilate is a special case, another question occurs: Obviously, at least two rolls are made, since the two attacks can crit separately. Anecdotal evidence suggests that miss/dodge/parry occur together. Does this mean (A) a double miss/dodge/parry occurs whenever EITHER of two normal attack rolls results in a miss/dodge/parry, OR (B) Mutilate actually has three rolls, an "avoidance" roll and two "crit" rolls? Case A would result in increased miss rate for Mutilate as compared to BS/SS/Hemo, since you have to "connect" twice. Case B would preserve hit rate, but potentially change crit rate, depending on what adjustments are made for Mutilate.

An idea i have about surprise attacks and mutilate... Let's assume your specials got a 5% dodge chance; maybe the game simply does a 1-95 roll instead of 1-100. That would actually mean that the dodge part of surprise attacks is a benefit.

That would contradict the idea that if you have 30% crit, 30% of your swings will crit. Your new crit rate would be 30/95 = ~31.6%. Gets silly if your target is a rogue with Evasion up. If the pre-SA table was 70% dodge and 30% crit, the post-SA table will be entirely crits!

So, with the current info we have about the different tree's and the itemization we've seen so far; what are you guys thinking for specs? The question is intended as a nod more towards your "everyday" spec, as opposed to some 1v1/5v5/BG spec.

Thinking about Mutilate, at least to try it out; if I don't like it, probably a combat daggers spec. If I got heavily into PvP I could see going for a deep sub spec, and it might actually be entertaining to level as a Shadowstep backstab build to the extent which that's possible.

I should note: Shadowstep swap build until you hit the point where you have enough points to pick up everything (so only the last couple levels would be SS Backstab; before that you'd be looking at a Shadowstep-Ambush-swaphemo setup.)

I'm thinking I'll also be spec'ing into mutilate. First, b/c it's new, but add to that the fact that you can pick up imp KS which will be a really nice asset if you're planning on doing a lot of leveling in instances (as I do). I can't really forsee myself spending a whole lot of time outside of instanced content, with the size of my server what it is, as I would expect everything will be pretty heavily camped for an unknown duration.

Spec'ing into combat would also be pretty good, but without SF or Mutilate you're going to be relatively starved for CP - which isn't very nice for 5 man content in general (most of the leveling I expect to be doing). As far as sub goes, it looks really cool for PvP and will probably have some pretty neat PvE applications as well, but as Pf mentioned, you give up quite a lot by heading down to the 41st talent - talent points that will be especially noticable pre lvl70.

TLDR version: I'm having a hard time justifying any build that doesn't contain imp KS for the type of leveling I'm expecting to do - just wanted to see if anyone else was in the same boat.

Note: In light of the following post, it's very possible I'm thinking about the tBC leveling game in light of today's 5 man content/raid game, which I probably shouldn't be doing. I don't expect many fights in tBC to be tank and spank though, which (I would assume) would place an increased priority on utility, burst and other unique rogue-ish abilities. Combo points are obviously unique to rogues and with some sort of SF/mutilate spec you'd be maximizing that and be able to get 'more' use from our new finishers - ranging from the ranged throw business to envenom -edit- mis-read the envenom tooltip - it eats CP :(

I was thinking the same thing, though I don't know how much I honestly like relying on combo point generation for damage. I just recently tried a SF dagger build for giggles over a weekend to try for pvp and found myself running back to the trainer to go back to combat after only a day of pvp'ing. I've just become so reliant on heavy white damage and pvp just isn't the same without 3.5 minute improved sprint and evasion admittedly 4 'guilty pleasure' points in my combat daggers build.

I'm having a really hard time trying to figure out what I'll migrate to in tbc. Especially after you get used to the buttloads of raw white damage you get from CD, going to a heavy subtlety build or SF build. Maybe if improved KS still gave you a reduced cooldown on KS it would be different, but since you can't get any better than 20 seconds anymore it's hardly worth it for pvp or solo'ing. I've just gotten way too used to my rogue with no prep or fancy whiz-bangs ;)

Mutilate sounds like it will be worth a try, but I think the eventual winner will be combat, as it was in the original game. Combat Potency looks simply amazing, though I'll also give a chance to something like 29/27/5 with 4/5 Seal Fate and 2/2 Weapon Expertise, because I never feel like I have enough CPs in the current combat dagger build.

I must have missed something; when did it stop eating CP? That would suddenly make it good.

Dunno. CS-Mutilate, wait for energy, KS-Mutilate is going to be a *ton* of burst for a Mutilate build. That might make for a hell of a grinding build, especially if combined with Remorseless.

Very true, I guess I will probably look at things differently when I don't have to beat on mobs for 10 minutes at a time... I've almost forgotten what burst damage is good for. It's the little things I think I'll miss the most, the blade flurries, the AR's... Getting whistful of combat, someone hit me with a club.

It's funny. I'd mentioned the other week that I'd picked up a CotBD to play around w/ hemo. Ran AQ/naxx trash as it, and decided I didn't like the damage I did w/ it, so I respec'd to combat fist and noticed farming in tyrs hand was notably more tedious (easier w/ hemo)... so, I respec'd back to hemo (slightly modified, CB in lieu of Premed) and am loving it again. 2 weeks ago I got stomped on the twin emps damage meters, however last night I topped them with basically the same spec. There's definitely a lot to be said for burst damage, in basically everything BUT the current raid game. At least what I've seen of it, which is limited to very early naxx.

Glass, if you've been a combat spec for that long (assuming you haven't tried it and hated it), may I recommend you check out a hemo spec? Yeah, it was FOTM a few months ago, but it really is a pretty fun spec. Each mob can be a different CP cycle depending on your level of buffs, raid DPS that night, and any number of different factors - it's pretty fun. My gear isn't even good by any stretch of the imagination (I was missing specials left and right on melee emp) but the spec can still put out alright damage.

For a thread about talents, we haven't talked a whole lot about builds.

Hemo Shadowstep

and

Sword Combat

These are really the only two builds I'm interested in at this point. I know a lot of rogues are still underwhelmed by SA, but I think it'll be pretty solid for swords. The Hemo build seems very viable if you can pick up lots of gear with both agility and AP together, which seems to abound in BC with the new itemization.

There should be a tier 4 subtlety talent that costs 1/1 point: "Removes the global cooldown on weaponswapping" if they're not going to add any Hemo usefulness to Shadowstep.

Just let me say that mutilate in current form is not reliable at all. you can do 600 or 1800 damage (example).

EDIT: The real strength of mutilate is the ability to spam envenom in my opinion. Without dagger spec, dual wield and WE, spaming envenom (especialy on trash or in smaller dungeons) seems extremly effective compared to SnD and the DPE ratio is pretty good. My envenom rank 1 averaged at 1700ish after 3 auchindoun runs which is very good for just 10 energy.

I dunno why people keep saying this. You will want to open plenty of fights with 40 energy +20% damage garottes that silence. If your reflexes are 'that good(tm) you can actually counterspell people with this move in group PvP.