Wishlisting for Warlords: Somewhat Less Sinister Strike

In a discussion on twitter last week I made my standard comment about bandit’s guile, it has the potential to be an interesting mechanic but it isn’t right now. I’ve been somewhat more positive on bandit’s guile than most in the rogue community since Blizzard added it in Cataclysm because I see it as a mechanic that could be a solid foundation for the spec. To begin lets identify the two obstacles preventing bandit’s guile from being an interesting mechanic.

  1. Bandit’s Guile is uncontrollable. Blizzard has said many times that they link RNG in rotations, how a player handles RNG distinguishes a skilled player from a less skilled player, bandit’s guile takes that too far. There is nothing to handle or react to with bandit’s guile, it simply moves through its cycle as you move through your rotation and you hope things line up. Anyone who has played combat is familiar with this type of scenario, you have 1 or 2 sinister strikes left before red insight and a boss burn phase is starting in ~20 seconds. Nothing, short of not hitting abilities and letting your energy cap is going to let you have red insight for that burn phase. This isn’t an example of RNG enhancing a rotation by letting the player wrangle it but rather the player passively accepting RNG as their lot and praying it comes up better next time.
  2. Restless Blades/Adrenaline Rush+Shadow Blades smothers any potential for bandit’s guile to be interesting. Since the introduction of these mechanics is cataclysm people have noted the inherent conflict. This isn’t inherently a bad thing, conflicting mechanics can lead to interesting choices but when one mechanic is a very powerful pair of cooldowns and the other mechanic is basically uncontrollable the conflict doesn’t last very long.

In this post I am not going to address point (2), combat’s cooldowns are a complex issue and fixing bandit’s guile can be decoupled. As Ninjablaze pointed out on twitter, use cooldown during red insight is hardly an interesting interaction. Both sets of combat mechanic’s need rethinking, in this post I focus only on bandit’s guile.

Lets begin with a simple, mostly uncontroversial statement, revealing strike is a boring ability, a hit this button every 24 seconds maintenance buff that is functional if not particularly inspired (a phrase which describes most of rogue PvE design). I propose we replace revealing strike with what Rfeann jokingly christened on twitter “Somewhat Less Sinister Strike” (SLSS). The idea of SLSS is simple, it has the same energy cost as regular sinister strike, generates the same number of combo points, does more damage than sinister strike, lets say 20-30% more and it does not advance insight.

Lets consider the example in (1) above, 1 or 2 sinister strikes from deep insight, burn phase coming up in ~20 seconds. The use case for SLSS here is obvious, you can use SLSS until you are just about to go into the burn phase and then use an SS to push you into red insight at exactly the right time. While this represents a simple use case it also has all the qualities that a good rotation should. It allows a skilled player who knows the encounter timing to optimize their damage for the encounter while simultaneously leaving a relatively simple priority system for newer players. The interaction with cooldowns is also reasonably straightforward. Since SLSS has the same combo points per energy as sinister strike using It doesn’t delay cooldown cycling and again it becomes about lining up high insight phases with cooldowns. Also to emphasize what I said earlier SLSS works whether or not restless blades is changed.

One objection I saw when I proposed this in the abstract on twitter was that this wouldn’t add a lot of rotational complexity. While its certainly true that on a patchwork style fight the theorycrafting community will work out optimal rotations based on how much time to spend in each insight phase per insight cycle most rotations are reasonably simple on patchwork and the vast majority of fights are not patchwork. One of the qualities that distinguishes skilled players is doing high dps on non-patchwork fights, knowing how to use the fights timing’s to your advantages. To utilize SLSS optimally requires the player to think ahead a full insight cycle, approximately 30 to 40 seconds. That is intuitively a somewhat challenging task given how many relevant raid mechanics there may be within that cycle. Yes, on progression those timings would be learned but that happens with every hard fight, ask anyone who has recently finished heroic Garrosh progression to map out the fight and most of them could do it in their sleep.

As with the other wishlist posts this idea is rough and would need testing before anyone could know if it would be a fun change. The damage of SLSS would need to be carefully balanced so it was worth using more often than just during red insight but that is the kind of numbers tuning that Blizzard has tended to be pretty good at handling this expansion. Clearly SLSS isn’t the only change combat needs, restless blades and combat cooldowns need to be discussed but this change would help liven up the combat rotation and help bandit’s guile finally fulfill the promise it has held since its creation.

Homogeneity and Flex

Reading through the comments on a recent WoWInsider Encrypted Text article I saw a similar sentiment expressed by a lot of players, “Rogues are just generic mDPS.” This isn’t the first time I’ve seen this sentiment, it pops up quite often on various forum posts but this time I saw it I started thinking about why this sentiment is so common when in my experience it is simply not true. The answer I believe is an issue of level of play, the exact things that have opened up raiding so well to casual players have resulted in this sense of homogeneity. Before I go further I quick note on the use of the word “casual,” I’m not a fan of the word in general, it has too many definitions to be broadly meaningful, in this post I am using it as a convenient shorthand for flex, normal and lower progression heroic raiders who do not min/max their raid comp.

“Bring the Player not the Class” was the watchword of Wrath of the Lich King and almost every mention of class balance brought up the ugly acronym “BtPNtC”. Blizzard mentions BtPNtC much less now, primarily because it worked so well. While I would be a fool to claim that class balance today is perfect, the degree that classes are balanced to today is incredible even from the perspective of late Wrath. There will be misses occasionally (Hello Destruction Warlocks) and questions about how BtPNtC applies to high-end heroic progression but overall Blizzard succeeded. Left unsaid in many of the discussions of BtPNtC is how it enabled the other major Wrath philosophy change, casual raiding.

The link between the two is obvious in hindsight, casual raiding, pug raiding and the new hotness of flex raiding basically require raid comps to flexible where one a melee dps is a melee dps and a healer is a healer. To be sure certain dps may be better for one encounter or another but the encounter is balanced such that groups don’t need to care about those differences. Lets consider flex style raiding because come Warlords flexible raiding is going to be what most people experience. Flex is built around a simple idea, the raid must go on, if someone can’t get online, the raid must go on, if someone has to leave early, the raid must go on. It isn’t perfect, a tank leaving is still a problem but flexible raiding loosens up comp restrictions about as far as they can go while still preserving the basic concepts of raiding and the trinity.

It should be clear why BtPNtC is essential to flex, BtPNtC allows raid leaders to think of people in terms of their role, with skill and gear being deciding factors about whether or not to bring someone rather than having to think about their classes. It wouldn’t be an exaggeration to say BtPNtC at a high level turns players into roles rather than classes. This brings us full circle, the reason many rogues feel like generic melee dps is self-apparent, the design of raiding today requires it. The question casual players need to ask themselves is not “Would you like less homogeneity?” but rather “Would you be willing to give up flex raiding in exchange for reducing homogeneity?”

I try to end my posts on a positive note, if I identify a problem I like to at least point in the direction of a solution, here I have nothing. The conflict between casual/flexible raiding and class diversity is fundamental and in many ways built into the underlying design philosophies Blizzard has been pushing since Wrath with great success. When I first started thinking about this I called it “Blizzard short-changing casual raiders” but I’m not sure that assessment is accurate. Blizzard looked at this seemingly intractable conflict and deemed access to raiding more important than class diversity and I suspect a large portion of casual raiders would agree.

Wishlisting for Warlords: Tricks of the Trade

A couple weeks ago Haileaus made a nice post over at his new rogue blog about why Tricks of the Trade is a failed ability. Tricks has been a pet issue of mine for a while now so I want to emphasize Hail’s points in a few places. If you want to read my full anti-tricks argument it is still on the forums.

Tricks is a flat out boring ability, pretty much every raiding rogue has a set of macros for frequent tricks targets (other rogues, tanks, high dps) and then a focus macro for uncommon targets. Every 30 seconds you hit that button, there is no interaction there, it is hit this button on cooldown to not lose damage. With threat what it is using tricks on the tank, even on the pull, is a waste so target selection, using it on a target with a good threat drop, isn’t relevant anymore.

Beyond that tricks suffers from a damage and utility dual use. Blizzard correctly recognized long ago that having an ability that did damage and had important utility was asking for trouble because players will always feel penalized if they have to use that ability for utility and since then Blizzard has been attempting to remove dual use damage and utility abilities. Tricks is in theory an awesome ability for when things go badly but since I’m using it on cooldown for dps it is almost certainly not up.

To finish this discussion of why the tricks damage buff should go look at this post by the great rogue theorycrafter Aldriana from more than two years ago during cataclysm which discusses again why tricks and more broadly tricks based set bonuses should be reevaluated.

Now that I have hopefully convinced you that the Tricks of the Trade damage buff should be removed the question becomes, now what? A few ideas.

1) Remove Tricks of the Trade. Threat is basically irrelevant in WoW today, with 500% threat modifiers and big AoE damage abilities like keg smash its generally safe to assume that if an add is in range of the tank, the tank has threat on it. Unlike hunter misdirects which can occasionally be useful rogues as a melee class generally can’t help pick up new adds that the tanks can’t reach themselves. Tricks is useful in random heroics with tanks you outgear by 50+ ilevels but that seems like a very specific use case to keep an ability around. If Blizzard is looking to clean up ability bloat tricks would be a good place to start.

2) Make Tricks an offensive raid cooldown. I don’t like this idea because I feel offensive raid cooldowns undermine class balance at a very fundamental level. Most class or spec specific raid utility is at least in some form situational, death grip for instance while very powerful on many encounters is sometimes not useful. Offensive raid cooldowns do not have a fight where they are unneeded, more dps helps on all fights. To preempt some criticism I freely admit that defensive raid cooldowns while often limited in some ways are similarly too widely usable.

The common response to this is, Blizzard already has a perfectly serviceable solution to the problem, sated style debuffs. My favorite variant of sated style debuff was put forth by Shadowboy on twitter. Change all offensive raid cds (Skull Banner, Stormlash Totem, now Tricks of the Trade) to provide the same with buff and normalize the cooldowns. These raid cooldowns share a shorter sated style buff to give this offensive raid cooldown a different character than Bloodlust et al. Whether Blizzard makes tricks an offensive raid cooldown or not I’d like to see them go to this design for skull banner and stormlash totem.

3) Make Tricks an external personal cooldown. One of the more interesting ability trends we’ve seen in MoP is the addition of external personal cooldowns, these have always existed in tank external cooldowns such as pain suppression but MoP has brought a number of more varied personal external cooldowns, monk’s Tiger Lust, priest’s feathers and in a slightly different vein symbiosis. This design has a number of useful features, it loosens comp restrictions by allowing more people to fill a given role and it creates some interesting comp choices by allowing one player to cover for another’s weaknesses. For these reasons personal external cooldowns are something I hope Blizzard does more with in WoD, possibly as a replacement for the defensive raid cooldowns Blizzard has been handing out like candy for the past two expansions.

Tricks of the Trade could make a very interesting personal external raid cooldown. One potential idea that I posted on twitter last week:
Tricks of the Trade
2 Minute Cooldown
When Tricks of the Trade is cast on a friendly target the next time the rogue uses Feint, Cloak of Shadows or Sprint the friendly target receives a weakened version of that ability instead of the rogue.
Feint- Reduces AoE damage by 30% for 5 seconds
Cloak of Shadows- Reduces magic damage by 40% for 6 seconds
Sprint- Increases movement speed by 70% for 4 seconds

The basic idea is the rogue can trade some personal utility for increasing the utility of a raid member. A number of people on twitter suggested this idea might be overpowered, looking at these three abilities I don’t think any of them are that overpowered, cloak of shadows and feint are very specific damage reductions that will probably be situationally powerful but probably not generally overpowered at least compared to other raid cooldowns with 30% general damage reduction. If they are overpowered the cooldown of tricks could be increased of the damage reduction of the abilities could be decreased, the numbers here are ballpark attempts but the general mechanic is the real idea I want to get across.

Another complaint I saw on twitter was people don’t like having to trade personal utility for raid utility both from a thematic and gameplay perspective. I don’t understand this complaint from either perspective, thematically tag team or synergistic ability usage is a hallmark of many genre fiction. Beyond that while the idea of a rogue as a thief or assassin who doesn’t play well with others is one part of the rogue archetype it is by no means the only rogue archetype. From a gameplay perspective raiding already involves trading personal utility for raid utility, if you raid asks you to use cloak to soak a specific mechanic you cannot use it when you need it for a defensive cooldown. This idea of tricks simply allows more potential uses for raid utility. You are sacrificing a use of one of your abilities for someone else but that kind of coordination and team play seems like something that should be encouraged by mechanics in a team based game like raiding or pvp.

Tricks doesn’t really fit in WoW today, 500% threat multipliers and high threat abilities like keg toss have rendered threat transfers basically pointless and the simple on cooldown damage buff is hardly compelling gameplay. Tricks needs to be reevaluated to make it fit in the modern raiding game or removed in the interests of button bloat. I can’t say which option is better, it depends heavily on Blizzard’s class design goals for WoD but something needs to be done.

Wishlisting for Warlords: Injection

One of the big questions rogues were asking about Blizzcon (full thoughts on all the annoucements this weekend) last weekend was, when will the big rogue revamp happen? Listening to the panels I think a big revamp is unlikely however Blizzard did suggest the direction they want to go. The level 91 through 100 passives offer a place to add additional game play mechanics without adding too much to button bloat. This somewhat changes a lot of the wishlisting people have been doing about what WoD rogues will look like. We probably aren’t getting new resource systems or a ton of new abilities rather we will probably get a number of passives, some fun, some boring. In this post I propose an idea for an assassination passive that would add some, although not enough, much needed complexity to the spec.

The big myth of assassination dps is envenom matters. People like to talk about how assassination is all about maximizing the envenom buff when in reality it is minimal at best. Consider 3 players, one player who never clips the envenom buff, one who wastes on average 25% of the envenom buff and one player who wastes about 50% of the envenom buff. It should be noted that wasting 50% of the envenom buff is a huge amount and probably not something that a player could do if they wanted most of the time, 25% is a bit more reasonable but still somewhat unlikely. The difference between 100% and 75% of maximum envenom utilization is about 1.2% dps and between 100% and 50% is about 2.3%. That is more than zero but it’s effectively a rounding error. At realistic levels of envenom utilization (~85%) the difference in dps is less than 0.7%. Note these are not uptimes but rather uptime utilization, if a player can achieve 80% envenom uptime playing optimally then 50% utilization corresponds to 40% envenom uptime.

That said envenom is at least in theory an interesting mechanic, the idea of having the filler finisher not be spammable is a nice take on a combo point class and probably will be the basis for whatever changes Blizzard makes to assassination. The simple solution to adding complexity to assassination is to make the envenom buff more meaningful, make it buff poison damage as well as increasing application rate or add an additional effect to the envenom buff to make it more important. These changes are immediately appealing partially for their simplicity, I championed them for some time but considering them further they aren’t really enough. The problem is envenom buff utilization doesn’t really scale with skill, once you know how to not clip the envenom buff you are done. More broadly simply not clipping the buff seems like a basic prerequisite to playing the class much like keeping up SnD for combat or subtlety rather than a reflection of skill.

So then the question becomes if maximizing envenom buff uptime is too simple how do you properly reward skill for an assassination rogue. The answer I believe has two parts, timing restrictions and superlinear scaling. The problem with a simple uptime based scheme is it doesn’t tax your resources, if you don’t have enough energy to get another envenom out immediately then no big deal, buff uptime is still buff uptime. There are minor effects due to envenom pooling but as with raw uptime the impact of those is negligible. A more challenging system that would better reward players for good resource utilization should tax the player’s resource pool more. This is where timing requirements come in, by providing small windows where the player needs to use resources this encourages resource pooling and spending. For an envenom uptime based scheme a possible solution is to use buff continuity rather than raw uptime, that is how long can you keep the envenom buff up without letting it drop?

This is where the second requirement comes in, superlinear scaling. Keeping the envenom buff up for 12 seconds and letting it fall off then putting it up for 12 seconds again doesn’t require as much skill (yes and RNG) as keeping it up for 24 seconds straight despite being the same in raw uptime. The reward structure, that is damage amount, should scale proportional to difficulty.

This leads to injection, a simple passive that achieves both of these with some caveats/weaknesses discussed later.

Injection-Every second the envenom buff is active the player deals N damage to the target for every stack of injection and gains a stack of injection, max 30 stacks. The injection buff lasts 2 seconds.

The scaling on this may not be immediately obvious but the number of 1 stack injection events follows the triangular numbers and is shown in the graph below. The important thing to note is that the damage scales strongly superlinearly. For comparison consider three players with different levels of continuity with the same total envenom uptime, 30 seconds, assume each injection proc does 1 damage. Player 1 lets the envenom buff drop after every envenom and over 30 seconds of uptime does 75 damage. Player 2 has an uptime of 12 seconds and 18 seconds, this player does 219 damage. Player 3 has a single 30 second uptime and does 435 damage. This scales very fast with duration, to reduce the scaling additional stacks could be gained every 2 ticks or injection could occur every 2 seconds rather than 1. Additionally the impact of this scaling will depend heavily on the percentage of total damage injection does and this scaling pattern clearly won’t account for the majority of assassination damage.

Triangle numbers

Triangle numbers

Some people reading are probably up in arms about the RNG impact of this mechanic, to them I say, RNG is not inherently a bad thing. Reacting appropriately to RNG is an indication of skill, a skilled player will turn bad RNG into acceptable performance and good RNG into exceptional performance. I do not believe this system will add significantly more RNG into assassination dps than already exists through mechanics such as blindside and general encounter to encounter variance in crit rates and the like.

There are a couple of issues with this mechanic. First it increases the dependence on anticipation. This is not a unique issue, one of the powerful components of anticipation is that it allows to players to smooth out combo point RNG. This is going to be hugely valuable in any scheme relating to the envenom buff and a number of other schemes that modify combo point usage. The second issue is gear which again will be an issue with most envenom based systems. At low gear levels when envenom uptimes are in the 40-50% range there is a fair degree of additional buff continuity to be gained by careful play however at higher gear levels when envenom uptimes can increase to upwards of 80% like they are today envenom buff continuity is basically trivial. This will be an issue with most envenom buff based schemes, we will basically scale out of the difficulty and that does seem problematic. A possible solution is to nerf the duration of the envenom buff by 25% to 50%. This would make it very unlikely that we could achieve envenom buff uptimes of over 70% and 50% respectively without external factors.

I’m not saying this is a perfect idea, or that it would in practice even be a fun idea but I believe that a time limited and thus resource limited system with superlinear scaling like this injection mechanic is something Blizzard should pursue. Even if this idea ends up being a bust I hope this idea sparks some conversation about what kinds of passive abilities could liven up and increase the skill cap of the assassination spec.

Thoughts on 9/24 Rogue Changes

Since Blizzard simply changed global dps modifiers the numbers are pretty straight forward 4% for assassination, +8% for subtlety.  What this means and why it happened is another question.

The assassination nerfs make sense on their own, remember assassination came into this tier at approximately +3% because of the dispatch and mutilate buffs. Many rogues (myself included) wondered if those buffs were really needed as assassination didn’t seem to need single target dps buffs. In that context these nerfs make a ton of sense, we the playerbase were correct assassination didn’t need those nerfs and Blizzard has come around to that conclusion.

Where things make less sense is with respect to spec balance. Assassination is a very strong single target spec but so is combat with Assurance of Consequence(AoC). Combat probably does lose more dps on interrupted damage fights but the fights where assassination was clearly over the top were pure, uninterrupted single target fights, Norushan, Juggernaut, Malkorak and Siegecrafter. Combat and assassination were very well balanced this tier, combat has cleave, assassination has execute, combat has regular burst, assassination handles interruptions better. This change undermines that, nerfing assassination without nerfing combat means that combat with AoC is probably the stronger single target dps spec. Related to my earlier point since Blizzard buffed dispatch and then nerfed damage across the board assassination is now a weaker spec for non-execute fights, how much this matters is a bit unclear but its worth bringing up.

You don’t see combat tearing it up on Raidbots because there are just a lot fewer combat rogues out there. There are a few reasons for this, first a lot of people don’t have an AoC yet and until you do combat is a weak spec, second people who played assassination all T15 may not have a good combat weapon so even if they want to switch they are limited. Finally assassination remains an utterly braindead spec with an insultingly low skill cap which makes it a very appealing option. Until now assassination rogues didn’t necessarily have a good reason to go for AoC or a combat weapon because assassination was just as good, after these changes I expect most rogues to move AoC to the top of their list.

The real head scratcher in these changes is the subtlety buff. Subtlety was not hurting on raw single target damage, given the buffs to subtlety on the PTR (approximately 4%) plus today’s buffs subtlety with 4pc and AoC will likely be a very strong single target dps spec, almost certainly stronger than either assassination or combat. Subtlety is still saddled with positional requirements and weaker AoE/multitarget but if the concern is raw dps on uninterrupted single target fights then buffing subtlety by 7-8% seems completely counterproductive. To emphasize, it isn’t that the nerf to assassination is wrong per-say, I’d say its pretty well supported. The problem is nerfing assassination while leaving combat where it is and heavily buffing subtlety just shifts the problem of high single target dps it doesn’t actually fix anything.

All of this doesn’t answer the question most people are really curious about, what spec should I play on Tuesday? As always if you aren’t pushing hardcore progression it doesn’t really matter, all three specs will do competitive dps. For heroic progression my suspicion is if you don’t have a good combat weapon and AoC your best bet is assassination or subtlety. Subtlety’s AoE weakness and positional requirements will make it a weak spec on fights like Nazgrim, Galakras and Spoils so it probably isn’t a good full time spec but for tight single target dps checks it is going to be very strong. If you do have AoC combat you should probably go combat as your primary spec and it is probably worth getting a subtlety spec for big single target fights where it will likely be very strong.

5.4 Rogue PvE Changes Guide

Before 5.2 I wrote a guide to the rogue changes in that patch for players who hadn’t been following all the PTR builds and theorycrafting (like the other rogues in my guild) and it was pretty successful so here we go again. Since I am a progression raider this post will be written from that perspective but I will note things that are high level optimizations and not relevant for LFR/Flex/Normal mode raiders when possible.

This guide is split into two parts. The first part focuses on 5.4 changes and the second is set up as a FAQ that hits the high points of what you need to know going into 5.4.  As always if you have any questions leave them in the comments or ask me on twitter.

Changes:
For the full 5.4 patch notes look here, rogue specific notes here. I am only going to be focusing on the big stuff that changes gearing and playstyle, numbers tuning info you can get from the official notes.

RPPM:
The first big set of changes was to RPPM mechanics. If you are still unclear on RPPM mechanics Hamlet’s RPPM 101 and RPPM 201 are probably the best sources you’ll find. The big change with RPPM is that certain items that already scale with haste no longer scale their proc rate with haste, in practice this means trinkets or other items that have a stat proc rather than a damage proc. If you are looking at the formula’s in the above linked posts just take the “H” term out of those formulas. However items with procs that do not scale with haste, in practice this means static damage procs, do continue to scale their proc rates with haste.

The second major change to RPPM is that on the pull the time since last proc term used for bad luck correction is set to 120 seconds upon entering combat. During 5.2 you could increase the chance of an RPPM trinket proccing on the pull by waiting for long enough since the time since last proc term would increase, this is discussed in RPPM 201 linked above.  In practice this means that RPPM trinkets will have a fixed chance to proc on the pull no matter how long you wait between pulls.  This fixed chance is generally lower than 100% so you will no longer be able to guarantee a proc on the pull.  This makes RPPM trinkets somewhat weaker on fights with a lust and burn phase on the pull.

Finally 5.2 and 5.4 RPPM trinkets had their RPPM values doubled and durations halved. This means that RPPM trinkets should proc more often for less time, the end result should be more consistent proc uptimes over the course of an encounter. ICD based trinkets also had their ICDs increased slightly so RPPM trinkets now have slightly higher average case stats to make up for their greater variance.

Glyphs:
Blizzard cleaned up some of the underutilized glyphs, the biggest change is the baselining of the Glyph of Adrenaline Rush and replacing it with Glyph of Redirect which reduces the CD of redirect by 50 seconds. This frees up a glyph spot for combat rogues and creates a very useful glyph for all rogues.
Glyph of Crippling Poison has been replaced by Glyph of Sharpened Knives which allows FoK to apply the armor debuff, this may be useful for rogues in raids without a hunter to provide AoE armor debuff.
Glyph of Sap has been replaced by Glyph of Hemorrhaging Veins which allows the hemorrhage bleed to count for sanguinary veins. This, combined with the glyph of redirect, will help subtlety’s major target swapping weaknesses.
Check the official notes for other glyph changes.

Talents:
Rogue mobility talents were buffed across the board but gameplaywise I don’t expect many changes. Check the official notes for full talent changes.

Combat:
There are a number of major changes to combat in 5.4, the biggest change is the increase of sinister strike energy cost to 50 energy. To compensate for reduced combo point generation Blizzard added ruthlessness which is a 20% chance per combo point to generate a combo point when using a finisher. Since combat typically uses finishers at 5 combo points this means in practice you will only need to generate 4 combo points.

The other big change for combat is the T15 4pc nerf, the energy reduction during Shadow Blades is 15% down from 40%. This is relevant to other rogues specs but it is primarily relevant to combat because of how the T15 4pc synergizes with restless blades.

Assassination and Subtlety:
Both specs received various dps buffs, check the official notes for full details.

=====================
5.4 PvE FAQ
What spec should I play?
Assassination will remain a strong single target spec. The dispatch buff is a substantial increase to assassination’s execute range dps. On patch day dps for an assassination rogue without T15 4pc will increase by 5%, for an assassination rogues with T15 4pc will increase by about 3%.

Combat will start 5.4 at an approximately 10% dps deficit from live on patch day if you have T15 4pc, about 2% without however once you get the cooldown reduction trinket (any level will do, even LFR) Assurance of Consequence (AoC) combat will make up that deficit. T16 also has a number of fights that depending on strategies could be very good for combat regardless of the relative strength of combat on single target.

Subtlety will gain about 4% dps on patch day, target swapping will be much improved because of the new hemorrhaging veins and redirect glyphs however the spec is still hamstrung by positional requirements and a general weakness on add fights. Subtlety does appear to do comparable single target damage to other rogue specs but the lack of a compelling niche hurts the spec.

One final consideration on spec choice for T16, stat and gearing priorities between combat and assassination in particular diverge heavily during T16. The exact details are discussed below but the big takeaway is that switching between combat and assassination of a fight to fight basis is going to be very hard and require substantial regemming and reforging. It will likely be better to play a single spec primarily and switch only if your guild is having a lot of trouble on a specific fight.

Recommendation (Heroic Progression):
Either combat or assassination should be fine especially once you get an AoC as combat. An important consideration is your raid comp, if you raid lacks strong cleave specs like ele shamans combat may be stronger than it is on paper since it fills a hole in your roster. In general playing either of these specs will be unlikely to make you a liability to your raid assuming you are skilled in the spec you choose to play.

Recommendation (LFR/Flex/Normal Progression):
All three rogue specs are close enough on general dps output that you should feel free to play the spec you enjoy most without being a liability to your group.

Should I change my talents?
Nope. As noted above mobility talents were buffed across the board. With the reduced cost burst of speed is a more interesting in some scenarios but it will still remain a very niche talent because of the energy cost. The buff to nightstalker still doesn’t appear to be enough to move it beyond the category of leveling talent.

Should I change my glyphs?
Probably yeah.
Combat- The big glyph change you should make will happen automatically on patch day as the Glyph of Adrenaline Rush is replaced by the Glyph or Redirect. This will make the Glyph of Redirect and Glyph of Smoke Bomb your default choices with the 3rd slot rotating based on fight specific choices.
Assassination- Smoke Bomb and Vendetta will likely remain your default glyph choices with Redirect being added to the 3rd glyph slot rotation, on some of the add heavier fights you may want to drop Vendetta for another glyph on the 3rd slot rotation.
Subtlety- Glyph of Hemorrhaging Veins and Smoke Bomb will be your default choices with Redirect added to the 3rd slot rotation.  Glyph of Vanish is a potential dps increase of around 0.2-0.3% if used with subterfuge, whether that is worth a glyph slot over a more utility oriented glyph is a matter of personal preference.

How will playstyle change?
Combat- 50 energy sinister strike and the nerf to T15 4pc will slow combat down somewhat. During Adrenaline Rush/Shadow Blades you will no longer be riding a .5 second GCD so combat will be somewhat friendlier to players with poor latency. Ruthlessness also interacts oddly with anticipation which will change bandit’s guile pooling slightly, instead of finishing at 5 cps, 4+ anticipation stacks you will want to finish at 5 cps, 3+ anticipation stacks, because if you have 5 anticipation stacks your ruthlessness combo point will just vanish. It is worth noting here that the slower combat rotation will slow bandit’s guile cycling however it will not slow finisher frequency for the purposes of restless blades as Shadowvenom shows here.  Rupture remains stronger than eviscerate so optimally it should be used however the dps increase for using remains in the 1% range and may not be worth using.

Assassination- Playstyle won’t change much, the nerf to and eventual loss of T15 4pc will make energy overflow and envenom clipping during Shadow Blades less of an issue. How to optimally stack Shadow Blades and Vendetta with T16 4pc is an open issue if T16 4pc is even optimal to use.

Subtlety- Target swapping will be much easier, instead of trying to get a low cp crimson tempest or rupture up on a new target you can simply put a relatively cheap hemorrhage on the target. T16 4pc will also change cooldown usage slightly, simply put if you have a ambush proc shortly before you were going to use Shadow Dance, Vanish or Shadow Blades you should delay that cooldown till the end of find weakness from the ambush proc.

How should I gear?
The following are very general recommendations based on general trends, gearing this tier has a lot of complicating factors so I highly recommend using ShadowCraft of SimCraft. If my recommendations below conflict with what ShadowCraft or SimCraft are telling you, trust the tools.

Combat- The big question for combat gearing in 5.4 is when to replace your T15 4pc bonus because you will be holding onto it for quite some time. As a general rule of thumb is it takes 4pc of the same quality of T16 tier to replace T15 4pc. That means if you have LFR T15 4pc it will generally take at least LFR T16 4pc to be dps neutral, similarly if you have heroic T15 4pc you will need heroic T16 4pc to replace it. This assumes you have an AoC, if you do not have an AoC, T15 4pc appears to be replaceable with 2-3 pieces of equal quality tier but make sure you save your old tier because you may need to use it again once you get an AoC.

In terms of stat priority combat will stick with stack haste to the gills and then mastery>crit. The big point to make here is that haste is becoming even more valuable for combat next patch to the point where the value of haste is as high as 2.5 EP compared to the 1.5 or so EP it is now. As mentioned previously this will make switching to combat on a fight to fight basis require a heavy regem.

Gearwise you need to get your hands on an AoC immediately if you are planning on running combat as your primary spec. AoC is an approximately 12% dps increase for combat which is I believe greater than any other agi spec. Even if you are a normal or heroic raider you should run LFR and flex for an AoC, an LFR AoC is better than any other trinket available to you other than higher ilvl versions of AoC. AoC drops from Sha of Pride, the 4th boss of SoO so it will be available in flex week 1 and LFR week 2. I would also recommend using coins to get an AoC because it is probably the largest upgrade you can get. Conversely tier gear isn’t high priority for combat because it will a long time before you can use it, if the goal is to maximize gear distribution efficiency it may be best to pass on tier tokens until they are going off spec.

Assassination- Given the nerf to T15 4pc replacing T15 4pc is really easy for assassination. In general a single piece of higher ilvl tier will be enough to break T15 4pc, for example a single piece of normal T16 gear, not necessarily tier gear will be enough to break T15 4pc. T16 2pc is also a very strong set bonus and should be a high priority. T16 4pc by contrast is very weak and may not be worth using compared to several pieces of higher ilvl. Note that ShadowCraft somewhat undervalues T16 4pc so if the difference between T16 4pc and offset gear are very close T16 4pc may be worth using.

Stat priorities for assassination next tier are a bit complex. T16 2pc increases the value of crit substantially, and the legendary meta gem and cloak procs which together make up about 12-14% of your dps (higher on multi-target fights) increase the value of haste. The stat priorities for various conditions is shown in the table below.

Legendary Meta Gem Legendary Cloak T16 2pc Stat Priority
N N N Mastery>Haste=Crit
Y N N Mastery>Haste>Crit
Y Y N Haste>Mastery>Crit
N N Y Mastery>Haste>Crit
Y N Y Mastery>Haste=Crit
Y Y Y Mastery=Haste=Crit

Again these are very general recommendations and you should check ShadowCraft to be sure.

Subtlety- Both the 2pc and 4pc bonus for subtlety are very strong and should be your top priority. As with assassination the nerf to T15 4pc makes it very easy to replace, a single T16 ilvl item from the same difficulty is typically enough to break T15 4pc.  Subtlety stat priorities depend primarily on whether or not you have AoC, without AoC it Haste>Crit>Mastery, with AoC Haste>Crit=Mastery however there are some gear sets where this may not be true.  Additionally it appears optimal to cap expertise for subtlety in T16 which is a change from previous tiers although many players may have capped it by accident.  AoC is a strong subtlety trinket, not as strong as for combat but the flex and sometimes the LFR version will be an upgrade over heroic ToT trinkets.

What Trinkets Should I Use?
For all specs the cleave trinket Sigil of Rampage is weak on single target but quite powerful on multi-target, on a constant 2-3 target cleave fight Sigil of Rampage should pull ahead. It probably shouldn’t be a priority to get because it is situational but for cleave fights it can be very valuable.

Combat: As mentioned above AoC is very powerful for combat. The difference between other trinkets is less important, Harom’s Talisman is the next strongest trinket followed by Ticking Ebon Detonator.

Assassination- Assassination gets surprisingly little out of AoC, both Harom’s Talisman and Ticking Ebon Detonator are stronger trinkets on paper. In practice AoC may be worth using at a slight (<0.3%) dps loss because of the additional utility granted by shorter Cloak of Shadows and Sprint cooldowns in addition to lining dps cooldowns up with fight mechanics. Additionally since Ticking Ebon Detonator is an RPPM trinket and is not guaranteed to proc on the pull on fights where you will use bloodlust on the pull AoC’s proc may be more valuable.

Subtlety- AoC and Harom’s Talisman appear to be the strongest trinkets for subtlety. Ticking Ebon Detonator is close to Harom’s Talisman on paper but as noted in the assassination section it may be weaker in practice than it is on paper because of the randomness and lack of proc on pull.

What is BiS?
I don’t know and I don’t care. In the age of reforging and when we have such easy to use tools like ShadowCraft at our disposal BiS lists have basically outlive their usefulness. I’m sure someone will put one together because some people stubbornly hold to them but I won’t be.

UPDATE 9/6: Updated wording based on some comments on twitter.  Also updated subtlety section based on feedback from Haileaus one of the few sub focused theorycrafters.

T16 Trinket Values

NOTE: This post is based on a mid 5.4 PTR beta build and is no longer accurate. For more accurate T16 trinket reccomendations refer to my 5.4 Rogue Changes Guide or better yet ShadowCraft.

Initial Values for T16 Trinket from ShadowCraft. This contains the change to RPPM mechanics haste scaling as well.

For RSC and Vicious Talisman two versions, one with full yellow hit/exp usage, one with no exp/white hit usage. For cleave trinket (sigil of rampage) showing value for 2 and 3 target fights. Also including Thok’s Tail Tip the strength amplification trinket and a generic int/spirit amplification trinket. Agi_5 is a so far unnamed trinket, possibly vp trinket if those will be available this tier. All T15 trinkets are unupgraded. All T16 trinkets are normal mode only.

Standard caveat: These trinkets have been hacked into ShadowCraft by me, it is possible I made a mistake implementing a mechanic or entering data. If you find any questionable results let me know in the comments or on twitter.

Assassination:
sigil_of_rampage_3_target: 8828.42217528
heroic_thunder_bad_juju: 8035.79761154
agi_5: 7928.23465741
sigil_of_rampage_2_target: 7774.28213533
vicious_talisman_of_the_shado-pan_assault: 7535.95974888
assurance_of_consequence: 7533.65954962
heroic_bad_juju: 7462.26394224
ticking_ebon_detonator: 7186.50403248
heroic_thunder_renatakis_soul_charm: 7140.92245691
thunder_bad_juju: 7112.46400267
heroic_renatakis_soul_charm: 6751.2583449
bad_juju: 6720.44993564
sigil_of_rampage: 6720.14209538
heroic_thunder_talisman_of_bloodlust: 6465.78982612
haroms_talisman: 6396.79219178
thunder_renatakis_soul_charm: 6322.44012299
heroic_talisman_of_bloodlust: 6109.32167314
renatakis_soul_charm: 5974.79302864
thunder_talisman_of_bloodlust: 5727.89798635
lfr_bad_juju: 5572.13094005
talisman_of_bloodlust: 5414.16485598
heroic_thunder_renatakis_soul_charm_exp_cap: 5380.83571471
vicious_talisman_of_the_shado-pan_assault_hit_cap: 5257.82669224
thoks_tail_tip: 5100.56272937
heroic_renatakis_soul_charm_exp_cap: 5086.61009614
lfr_renatakis_soul_charm: 4956.08118694
thunder_renatakis_soul_charm_exp_cap: 4763.27652488
renatakis_soul_charm_exp_cap: 4501.02176677
lfr_talisman_of_bloodlust: 4494.94603713
lfr_renatakis_soul_charm_exp_cap: 3732.45923945
amp_int/spirit: 3100.29102165

Combat:
assurance_of_consequence: 18851.0897391
sigil_of_rampage_3_target: 12486.1209147
heroic_thunder_renatakis_soul_charm: 10594.9574319
sigil_of_rampage_2_target: 10254.2145738
heroic_renatakis_soul_charm: 9761.84363479
vicious_talisman_of_the_shado-pan_assault: 9546.41026602
thunder_renatakis_soul_charm: 9142.6711216
heroic_thunder_bad_juju: 8681.16944015
renatakis_soul_charm: 8640.87883416
agi_5: 8315.9016962
haroms_talisman: 8240.20122134
sigil_of_rampage: 8022.30823285
heroic_bad_juju: 7903.67178692
ticking_ebon_detonator: 7527.75468059
thunder_bad_juju: 7401.32884976
lfr_renatakis_soul_charm: 7170.38749653
bad_juju: 6994.01817235
heroic_thunder_talisman_of_bloodlust: 6908.51769254
heroic_talisman_of_bloodlust: 6533.50839589
thoks_tail_tip: 6497.66492059
thunder_talisman_of_bloodlust: 6120.64699331
vicious_talisman_of_the_shado-pan_assault_hit_cap: 5836.83191694
heroic_thunder_renatakis_soul_charm_exp_cap: 5818.28652342
lfr_bad_juju: 5800.5202182
talisman_of_bloodlust: 5785.60373412
heroic_renatakis_soul_charm_exp_cap: 5244.18170824
thunder_renatakis_soul_charm_exp_cap: 4911.28228032
lfr_talisman_of_bloodlust: 4803.78389225
renatakis_soul_charm_exp_cap: 4641.23487143
amp_int/spirit: 4484.82809926
lfr_renatakis_soul_charm_exp_cap: 3849.61970744

Subtlety:
sigil_of_rampage_3_target: 13332.6207081
sigil_of_rampage_2_target: 11433.4057235
heroic_thunder_bad_juju: 10042.1324585
sigil_of_rampage: 9534.19073898
assurance_of_consequence: 9311.33984487
heroic_bad_juju: 9303.58367649
thunder_bad_juju: 8887.22644834
heroic_thunder_renatakis_soul_charm: 8872.73855642
ticking_ebon_detonator: 8812.13405076
heroic_thunder_talisman_of_bloodlust: 8460.74530901
bad_juju: 8396.96900462
vicious_talisman_of_the_shado-pan_assault: 8388.52873295
heroic_renatakis_soul_charm: 8388.18038583
agi_5: 8086.86255759
heroic_talisman_of_bloodlust: 8001.51007423
thunder_renatakis_soul_charm: 7855.13815132
haroms_talisman: 7811.05720048
thunder_talisman_of_bloodlust: 7495.52526384
renatakis_soul_charm: 7422.95930356
talisman_of_bloodlust: 7085.12306664
heroic_thunder_renatakis_soul_charm_exp_cap: 7021.74740265
lfr_bad_juju: 6960.99896718
vicious_talisman_of_the_shado-pan_assault_hit_cap: 6893.36639497
heroic_renatakis_soul_charm_exp_cap: 6637.55690307
thunder_renatakis_soul_charm_exp_cap: 6215.44735757
lfr_renatakis_soul_charm: 6156.58895283
lfr_talisman_of_bloodlust: 5882.58964728
renatakis_soul_charm_exp_cap: 5873.07116282
thoks_tail_tip: 5279.88748421
lfr_renatakis_soul_charm_exp_cap: 4869.76976031
amp_int/spirit: 3274.70388414

UPDATE 8/7:
Most recent PTR build had a number of trinket changes. Many RPPM trinkets had RPPM increased and duration halved. 5.4 ICD trinkets had their ICD’s increased by 10 seconds. Full changes below.
5.2 Trinkets:
Renataki’s Soul Charm: 10 second duration, buff increases every 1 second (down from 20, buff increases every 2 seconds), 1.21 RPPM (up from 0.7 RPPM)
Bad Juju: 10 second duration (down from 20), 1.1 RPPM (up from 0.55 RPPM)
Vicious Talisman of the Shado-Pan Assault: 115 second ICD (up from 105)

5.4 Trinkets:
Agi_5: 115 second ICD (up from 105)
Sigil of Rampage: 85 second ICD (up from 75)
Assurance of Consequence: 115 second ICD (up from 105)
Thok’s Tail Tip: 7% Amplification (down from 14%), 115 second ICD (up from 105)
Ticking Ebon Detonator: 10 second duration, buff decreases every .5 seconds (down from 20 second duration, buff decreases every 1 second), 1.01 RPPM (up from 0.57 RPPM)
Harom’s Talisman: 10 second duration (down from 20), 0.92 RPPM (up from 0.46 RPPM)

EP Lists updated accordingly.

The Unofficial ShadowCraft Documentation-Resource Budget Modeling

One of the most basic questions when writing a model is what is the model actually modeling? Those of you who have been around the rogue community for a while will remember the Wrath (and earlier) notation of “3s/5r/5e” corresponding to 3 cp slice and dice, 5 cp rutpute, 5 cp eviscerate. This fixed rotation is clearly an abstraction however it exposes a useful concept which is still used in ShadowCraft today, resource budget modeling (my term).

There are three basic ideas of resource budget modeling:
1) Rogue ability selection can be defined in terms of finisher usage.
2) Certain finishers must be kept up 100% of the time and can be considered a fixed cost.
3) The duration of one of these 100% uptime finishers defines a time slice or cycle.

Bringing these three ideas together we have a pool of resources per cycle. Since we are being finisher centric the next step is to define the energy cost of one finisher, that is the energy cost of both the combo points and the finisher. This can actually be a bit more complex than it seems as I’ll discuss in a later post but the simple solution is reasonably apparent. Simply divide the target number of combo points per generator by the total combo points for the finisher noting that the total number of combo points per generator includes things like blindside procs.

Now we have energy costs per finisher and we have a resource pool of energy per cycle. By assuming 100% uptime on certain finishers those finishers can be considered a fixed cost in our resource pool and whatever is left over can go into the “spammable” finisher, eviscerate or envenom. With numbers of each finisher per cycle, a cycle duration and combo generators per cycle we can generate attacks per second for each ability. There are some complexities the above discussion ignores, honor among thieves combo points as subtlety and revealing strike rather than sinister strike usage as combat but both of those can be handled using basically the same approach.

Now some code. Since this is the first post in this series a quick word on how ShadowCraft is organized. The major theorycrafting function in ShadowCraft is the attack_counts function for each spec. These functions drive ShadowCraft generating a list of attack counts per second for each ability that is then used to do the actual damage computation.

For this code example we’re going look at the assassination anticipation attack counts function because it is the simplest. From line 1648 in ShadowCraft. From here on the line numbers will refer to line numbers in the code block below.


cp_per_finisher = 5
avg_rupture_length = 4. * (6 + self.stats.gear_buffs.rogue_t15_2pc_bonus_cp()) # 1+5 since all 5CP ruptures
avg_wait_to_strike_connect = 1 / self.geometric_strike_chance - 1
avg_gap = 0 + .5 * (avg_wait_to_strike_connect + .5 * self.settings.response_time)
avg_cycle_length = avg_gap + avg_rupture_length
attacks_per_second['rupture'] = 1 / avg_cycle_length
rupture_ticks_per_second = 2 * (6 + self.stats.gear_buffs.rogue_t15_2pc_bonus_cp()) / avg_cycle_length # 1+5 since all 5CP ruptures
attacks_per_second['rupture_ticks'] = [0, 0, 0, 0, 0, rupture_ticks_per_second]

energy_regen_with_rupture = energy_regen + attacks_per_second['rupture_ticks'][5] * vw_energy_per_bleed_tick
energy_per_cycle = avg_rupture_length * energy_regen_with_rupture + avg_gap * energy_regen
cpg_per_finisher = cp_per_finisher / avg_cp_per_cpg

energy_for_rupture = cpg_per_finisher * cpg_energy_cost + self.get_spell_stats('rupture', hit_chance=self.geometric_strike_chance, cost_mod=ability_cost_modifier)[0]
energy_for_rupture -= cp_per_finisher * self.relentless_strikes_energy_return_per_cp
energy_for_envenoms = energy_per_cycle - energy_for_rupture

envenom_energy_cost = cpg_per_finisher * cpg_energy_cost + self.get_spell_stats('envenom', hit_chance=self.geometric_strike_chance, cost_mod=ability_cost_modifier)[0]
envenom_energy_cost -= cp_per_finisher * self.relentless_strikes_energy_return_per_cp
envenoms_per_cycle = energy_for_envenoms / envenom_energy_cost

envenoms_per_second = envenoms_per_cycle / avg_cycle_length
cpgs_per_second = envenoms_per_second * cpg_per_finisher + attacks_per_second['rupture'] * cpg_per_finisher

Since slice and dice is refreshed by envenom we use rupture to determine cycle length. Lines 1 through 8 above determine the length of the cycle. Line 2 handles the length of rupture so it checks for the existence of T15 2pc. Lines 3 and 4 handle the additional time caused by miss rates and player reaction time.

Lines 10 through 12 determine the resource pool, note that line 11 makes a distinction between the portion of the cycle with rupture up and the portion without rupture up.

Lines 14 through 20 determine the total costs of each finisher. Lines 14 and 18 account for the additional cost added by misses and lines 15 and 19 handle relentless strikes. Line 16 computes the remaining energy available for envenom and line 20 uses that number to determine how many envenoms are used per cycle.

Finally lines 22 and 23 convert those envenoms per cycle into envenoms and combo generators per second.

Beyond this the computed values are written into the attacks_per_second list for damage calculations. I did not show the computation of combo points per combo generator but that can be found above the discussed section.

If you would like to look at how other spec’s resource budgeting works you can look at sections starting at line 1522 (assassination no anticipation), 1919 (combat) and 2225 (subtlety).

Since this is the first post in what I hope will be a long running series I’m still trying to work out some of the formatting. Any feedback on where things are unclear would be welcome. ShadowCraft is also a big topic so if there is some part of it you are especially curious about let me know in the comments on twitter.

The Unofficial ShadowCraft Documentation-Introduction

In my very first post on this blog I talked about the difference between simulations and models. Simulations most people get, they correspond pretty closely to how most people think about theorycrafting and testing. Run a test of some behavior and look at the results. Models are a bit trickier because models rely on simplifying assumptions which can often be somewhat obtuse. The goal of this series of posts is to to explain those simplifying assumptions and show examples of them within the ShadowCraft codebase.

Another goal of this series is to get more people looking at and contributing to the ShadowCraft codebase. I know from experience the ShadowCraft codebase can be very intimidating to start working through partially because there isn’t an obvious starting point. By showing and explaining pieces of the ShadowCraft model I hope to make the prospect of digging into ShadowCraft a bit easier.

ShadowCraft is written entirely in Python, for the purposes of this series I am going to assume the reader has a decent understanding of Python. There are some pieces of code which are a bit obtuse and I there I will explain the syntax to an extent. For people who don’t know Python I will still explain the underlying concepts in plain English (and possibly a little math). If you do want to learn Python I recommend the free book A Byte of Python. Python is one of the easier programming languages to learn and is about as close to plain English as a programming language can get. For those with experience with C style languages (Java, C, C++, C#, etc.) a lot of Python syntax should look very familiar.

A final caveat, ShadowCraft is an actively developed project so the code samples I use may not be what ShadowCraft actually uses at the time you are reading this however the focus of this series are the underlying modeling techniques which should still remain broadly applicable even if the code changes.