The first post in this thread points to the PvE DPS article, which would answer that question. Please use it.

From what I read Fist MH / Sword OH is slightly better than the opposite. Thanks.

The short answer is.. Download the spreadsheets and use them

Based upon Spiteblade, I'll assume that you are in 2pT4. If you Keep sword spec with Spite/S2OH merciless and spend badges on the great hit gear available from the Sunwell badge loot (Chestpiece > kara lewt), you will probably come out ahead. again.. probably. do the spreadsheets.

What he said:

What we heard:

Just a general comment:

When people start making useless posts, the correct way of addressing this is not to make useless replies. Instead, I'd far prefer you to simply report it, at which point I will infract it and/or move it to The Thread of Ultimate Suck, and we can forget about it. Having half a dozen responses to a question that didn't really need to be asked in the first place is not really an effective use of people's time. Particularly those responses that do little besides pointing out the uselessness of the initial post without answering it. So moving forward if people could try to cut back on the useless responses and use the report functionality instead, I'd appreciate it. Thanks.

The Rogue DPS Spreadsheet models Blinkstrike the same as Sword Spec. It does not yet have WF -> WF and used a 5% proc rate. Honestly, I haven't found any compelling data on what the actual proc rate is. I would be interested in the results of any testing, but it's still far behind other typical 25-man raiding weapons, so I wouldn't go out of your way, but if someone happens to have one in the bank...

Taking into consideration the WF and SwdSpc proc research showing the chaining possibilities of these, as well as the possible testing of Blinkstrike and it's chance on hit proc, It occured to me that this "new" information may also lead to a slight increse in the stat-value of Haste Rating for sword spec rogues.

Just based upon what we know now, does this WF and Swd Spec changes affect Haste enough to warrent a boost in its stat value?

Any increase that haste gets, hit will get roughly 2x as much, due to WF and Sword Spec double dipping on hit (once to proc, and another time to land)

Really, it's *only* going to change the relative value of hit and expertise. 1 point of haste will do more damage as it's generating a few extra attacks... but attack power and crit will make similar gains, as the new WF mechanics give them more attacks to add on. So while they slide a little due to the slightly shifted balance of white and yellow damage, they don't change by very much.

For instance, with my gear + spec and the default buffs in RogueCalc, the old WF model gave EP values as follows:
Str: 1.1
Agi: 2.281
Crit: 1.871
Hit: 2.898
Haste: 2.426
ArPen: 0.358
Exp: 3.224

With the new WF model, but otherwise the same gear and buffs, this becomes:
Str: 1.1
Agi: 2.286
Crit: 1.879
Hit: 3.099
Haste: 2.453
ArPen: 0.360
Exp: 3.465

So agility increases in value by about .02%, crit and arPen by about .5%, and haste by a little over 1%... while haste and expertise both go up by a full 7%. So the net result of this discovery is that hit is *even more* valuable than it was before, while nothing else changes very much.

I have some vague recollection that there used to be some sort of gear setup back in the MC/BWL days whose goal was basically getting massive proc-strings using various things like Sword Spec, Windfury, Thrash Blades, Hand of Justice, a Darkmoon deck, weapon enchants, and probably more that I never even heard of. I don't know if that was a successful spec or not, although I seem to remember it getting nerfed in some fashion. Is there any reason to believe that a viable build can be built out of this gimmick or is it mostly a curiosity that alters some EP values?

Well, I think one thing that they changed is that Sword Spec no longer procs itself, and I believe that's true of some of the listed items as well - which is why we had all been assuming it was equally true of Windfury (which turns out to be an unsafe assumption). So I think that theory behind such a build has been nerfed somewhat.

That said, it is true that if you get enough extra attack procs, you can theoretically start generating disproportionately large amounts of damage. For instance, if each swing has a 10% chance to generate an additional attack, the expected number of attacks you get per regular swing is 1.11; if you bump that chance up to 50%, the expectation climbs all the way to 2 attacks per base swing; and if you get get it to, say, 80%, the expectation shoots all the way up to 5 attacks per base swing.

So, what we draw from this is that extra attacks do exhibit positive scaling with themselves - that is, the recent discovery regarding Windfury increases the relative value of Blinkstrike and Sword Spec - which is certainly something to keep in mind. On the other hand, there really aren't that many good sources of extra attacks - I mean, you have Vindicator's Brand as the best available MH... and nothing else even remotely competitive. Hand of Justice was a nice trinket back in the day, but it doesn't compete very well anymore, and even with the positive scaling of extra attacks I just don't think it's anywhere close to catching up.

Now, stacking extra attacks in this fashion *does* increase the uptime of PPM effects - but again, I don't think there's enough to make this increase significant.

Hence, while we should keep an eye out for more extra attack abilities, and we should remember their scaling properties should such things exist... it's also true that at the moment, there's little practical application for this knowledge other than tweaking some EP values.

The scaling property of the extra attacks makes for an attractive scenario. Perhaps, if Blizz is serious about setting up Itemized Gear Sets in WotLK, based upon specs for rogues, extra attack procs of this type could be used to justify a piece of gear as "best used by" a sword spec rogue. The same way a piece of gear that increases the duration of your deadly poison applications may "best used by" poison dependent Assassination/Mutilate builds. The new talents in deep combat leave me worried about the direction Combat is headed, but armor sets and weapons that maximize the advantages of subtle differences in specs would be a welcomed change from our current options.

Unless you're talking about tier sets being optimized for different rogue specs, I don't see this as the direction Blizzard is going. They're trying to create more general gear sets (combining spell damage and healing, combining melee hit and spell hit, so on), rather than even more specific ones. The only place that I could see this happening would be set pieces, and the different rogue specs aren't different enough from each other to require drastically different gear sets. Perhaps the set bonuses, though, would have dual purpose, in the way that feral set bonuses have cat and bear aspects to them.

I'm watching the live stream OF WWI, Tom Chilton, Lead designer of WoW mentioned "Windfury will be a buff in WOTLK and no longer a weapon enchant" during the Q&A portion of the event. This was a response to druids being able to use mongoose/windfury while shapeshifted, he included that rogues will be able to use poisons while still maintaining the windfury buff. He also mentioned Strength of Earth will included the Grace of air affect, so totem twisting will no longer be required.

Another interesting change that they mentioned is that totems will now affect the raid, not just the shaman's party. That'll relieve some (but not all) of the pressure on raid/group composition that a lot of us were worried about.

Hopefully in WotLK they will provide more incentive for groups to be more mixed - it should no longer cause the optimizers in the raid to cringe in torment when a spellcaster winds up in a group with a rogue. Of course certain synergies will be "best" - but more group compositions should be workable.

This totem change is a great step in that direction. Big grin here.

Q&A session liveblogs indicate battle shout and unleashed rage will be raidwide as well. We're picking up a number of 'open slots' in the raid that could adequately hold a rogue.


On thinking about this, I'm happy that there is an awareness of the notably limited raid slotting availiable to us with the current group buff system. What worries me now is that by fixing these problems by making buffs raidwide, rather than reducing our dependency on them, we'll find ourselves unable to find a slot in smaller group content, as we still offer nothing but our damage, and being the best buff-sponge doesn't make a terribly strong case in a limited buff environment. Upgraded CC might be a significant help here but I don't know if the expansion of eligible targets for sap is quite enough.

Updated the first post:

Fantastic news on the shaman front, obviously.

I'd love to comment on the bolded part.

Your parse had 4k+ swings, 600 windfury procs, 200 sword spec procs.
But looking through the log, there are only those 3 occurences of windfury chains.

If windfury can truly chain proc normally, then we'd expect:
~100 windfury double procs, ~20 windfury triple procs and a few quadruple procs.
Instead, we get 3 occurences with 2 windfury procs and no triple or higher procs at all.

Also, playing the meta-game, if triple WFs were consistently reproducable, they'd happen in arena and someone would log them and it would appear on forums. But they don't.


So, instead of a clear answer of "yes" or "no", we get "windfury very rarely can chain proc".
It's a 20% chance on 1 extra attack and 0.1% chance on 2 extra attacks. If it could truly chain proc, the double proc chance would be 4%, it would happen 40 times as often as we currently see it.

It's simply a mistake to say "WF can chain proc" from a long test and change all spreadsheets, when there are no occurences of triple procs, and double procs happen forty times less often than they should.


I wish I could find an explanation for the chain procs, but I can't. I thought of things like some lag between the WF swing and the WF attack buff (attack power buff), so that some part of the server (that does procs) doesn't realise that the WF attack is a WF attack and accidentally lets it roll for another attack.
But that's probably complete nonsense

I WWSed your log and mine to make looking for procs easier:
Wow Web Stats Vulajin, 4k swings
Wow Web Stats Dhelas, 500 swings

No flukes in my short test (was resto, so no parries since I lack the talent).

After hearing the windfury change announcement, I decided to run some numbers to determine how useful Envenom would be in an assassination build with the new talents.

Feel free to point out if my math or understanding of game mechanics is flawed in any of this.

Assumptions: 1.8 Speed Mainhand with Instant / Deadly Poison (2/2 Deadly Brew)
1.4 Speed Offhand with Instant / Deadly Poison (2/2 Deadly Brew)
5/5 Improved Poisons Talent (+25% application rate = 55% application rate)
Slice and Dice Rank 2 is Up (+30% haste)



So Mainhand Attacks / Sec = 1.8 -> 1.38 Hasted 6/ 1.38 = 4.34 attacks / 6 seconds
Offhand Attacks / Sec = 1.4 -> 1.08 Hasted 6/ 1.08 = 5.56 attacks / 6 seconds
Mutilate = 2 attacks / 6 seconds

Each auto-swing and each half of the mutilate has a 55% chance to proc deadly poison.

Total Attacks / second = 11.9 / 6 seconds = 1.98 attacks / second.

55% of 9.09 is 5. So, on average, 9.09 attacks will result in a 5 stack of deadly poison.

This means, on average, you will have a full stack of 5 deadly poison (after envenom clears it) after 9.09 / 1.98 = 4.59 seconds.

Since deadly poison ticks every 3 seconds, you're looking at only missing out on 1-2 ticks of your DP stacks for envenom. Now, since we don't know the new ranks of deadly poison or envenom, it's hard to say whether or not 2 ticks will eclipse envenom damage, but I think that envenom is going to be a lot more useful than it is now (the windfury change and deadly brew are what really allows the fast stacking that makes this viable, I believe)

Edit: To comment on the poster above me... Improved backstab doesn't exist anymore. Do you mean puncturing wounds?

A good point, and one which I actually hadn't considered. To help me out, Neto of <Taurential Reign> on Mal'Ganis took his shaman out to Blasted Lands and performed the same test you performed for over an hour. (Thanks a bunch, Neto!) Here's the WWS:

Wow Web Stats

4128 regular swings, 784 Windfury procs, or an 18.99% proc rate. Using some simple statistics, we find that the sample standard deviation is about 0.61%.

If we hypothesize that Windfury can only proc once and not chain itself on a single attack, then our overall extra attack proportion should have been 20%. This is within 2 standard deviations of the sample proportion (Z = 1.65) and therefore the hypothesis can't be reasonably rejected.

If we hypothesize that Windfury can chain itself without limit, then each attack produces a geometric series of extra attacks: 1 + 0.2 + 0.2*0.2 + ..., or 1/(1-0.2), or 1.25. So our overall extra attack proportion should have been 25%. This is so ridiculously improbable given the data provided by the experiment (Z = 9.84) that we can reject it outright.

However, Neto's log still featured three chained Windfury procs. For example:

His swing speed was 1.6 and he was specced elemental at the time, so there's no possibility of parries or haste effects interfering. So clearly there is some other peculiar mechanic at work here -- but I believe you're probably correct in that I was probably wrong that Windfury attacks can fully proc off of themselves.

Well, I think what we can say, based on the overall proc rate, is that the behavior is probably better modeled by assuming WF can't proc off itself than that it can, hence models should be adjusted accordingly. What I suspect we're seeing in these apparent proc chains is something latency-related - that is, the WF was actually generated by some earlier attack and not reported for 1+ seconds, or some such.

This, of course, calls into question the data about multi-part chains, which we'll have to investigate at some point. One thing that might help is to use a slower weapon; with a 2.6+ speed weapon it might help separate the procs a little better.