Encrypted Text: Theorycrafting basics for the studious rogue

In the immortal words of our favorite crab, "WoW is a game about upgrading your stuff." While achievements and professions allow for your character to grow, they don't make you any stronger. If you want to increase your power, you need to upgrade your stuff. Gear improvements are often the difference between success and failure on cutting-edge encounters, when your raid needs to meet specific raw damage and healing numbers to pull off a win.
While it sounds easy, it's not always as simple as it seems to upgrade your stuff. While the advent of the ilvl system allows for us to quickly judge the rough quality of an item, it doesn't provide any help when comparing two items from the same dungeon. It also doesn't assist us in our gemming or reforging choices. Making intelligent gear decisions is a part of every rogue's life. In order to make the best choices for our characters, we use what's known as theorycrafting to evaluate the various stats in the game. The end result allows us to make educated gearing strategies and ensure we're doing the most DPS we possibly can. But how do we get from evaluating stats to choosing an actual piece of gear?
Damage is first and foremost
In order to figure out the potency of the game's various stats, like agility and haste, we look at how they increase the damage of our abilities. Let's start with a move like Sinister Strike, one of the simplest rogue attacks. The tooltip for a level 85 player indicates that it deals 100 percent of our normal weapon damage, while also adding an extra 200 damage on top of that. Thus, we can easily express the damage of Sinister Strike as:
Weapon damage + 200 damage
Now, let's start adding our other rogue stats to the formula. Notice that you don't see haste anywhere in the formula. Haste doesn't affect the damage of our instant attacks, and so haste won't increase the damage of our Sinister Strikes at all. Critical strike, on the other hand, gives our Sinister Strike the chance to critically hit. Every time we critically hit, our Sinister Strike does double damage. Thus, critical strike rating increases the average damage of our Sinister Strikes. Let's say our critical strike chance is 25 percent, which is easily possible. We'll be doing double damage 25 percent of the time and regular damage 75 percent of the time. This will change our formula a bit:
(Weapon damage + 200 damage)*0.75 + (Weapon damage + 200)*2*0.25
We would then repeat this exercise for every possible stat. In addition, we also have complex stats like agility, which grants attack power and critical strike chance at the same time. After we apply all of the various modifiers, we're left with a complex formula that represents Sinister Strike's average damage. That's just one ability, and we haven't even factored in the various talents that affect our abilities! In order to perfect our rogue DPS model, we'd have to repeat this for every single ability. However, once we're done, we have a complete collection of damage formulas for all rogue abilities.
Damage per energy
Now that we know how much damage each one of our attacks can do, we have to factor in the cost of each spell. While Backstab may do much more damage than Sinister Strike in every situation, we have to consider that Backstab also costs more energy to use. In order to compensate for varying ability costs, we look at our damage per energy point spent, known as DPE. We have a limited supply of energy on any given fight, and so figuring out how we can squeeze the most damage out of that energy is a very important aspect of playing a rogue.
A great example of this is the new Murderous Intent talent. For the majority of any encounter, Mutilate simply does more DPE than Backstab does. You'll be spamming Mutilate merrily, and then the boss' life drops to 35 percent. Now that Murderous Intent is active, the DPE of Backstab increases drastically due to its reduced energy cost. Backstab now becomes your move of choice. Every rogue rotation comes from an examination of the best DPE abilities for that spec and then sorting them into a priority system.
Damage per combo point
Similar to DPE, our damage per combo point is a very important factor in forming our rotations. Since nearly every rogue spell will be picking up Relentless Strikes, many of our finishers don't actually cost any energy. Since we can't examine the DPE of an ability that costs no energy, we have to look at the DPCP instead. It's important to note that the DPCP of many abilities scales differently based on how many combo points you have. For example, a 5-point Slice and Dice does not last five times longer than a 1-point Slice and Dice.
A great example of the DPCP system in action is the basic combat rotation. Combat rogues have a choice when beginning a fight. Either they can open with a Slice and Dice to increase their number of autoattacks, or they can start with a Rupture to get their target bleeding early. Both abilities have a duration and increase our damage done, and both end up being free after Relentless Strikes is factored in. However, as every rogue knows, Slice and Dice provides so much damage for its combo points that it is always worth using first. Because of its high DPCP, we always open with Slice and Dice. If we had to, we would spend all of our combo points on Slice and Dice, because it provides the best benefit per combo point spent.

Once we've been able to evaluate the damage of all of our various abilities, we can determine the DPE and DPCP of all rogue abilities. The fun doesn't stop there, though. Now that we know the damage per combo point of a finisher, we now know how much damage we get from a single combo point. We have to factor that damage back into our damage calculations for each ability. Sinister Strike not only does weapon damage plus the bonus damage, it also grants a combo point, and we have to account for that in its damage formula.
As you can see, nearly every formula refers to another formula, which refers to another formula. For example, as our critical strike chance increases, the value of attack power increases. As the amount of attack power we have increases, the value of haste increases. Each of these curves are in constant motion, which means we need a spreadsheet or program to track the constantly fluctuating damage values. Our stats work the same way, as there are tons of talents and buffs that affect how much critical strike chance or attack power we have at any given time. We also have to adjust our overall damage formulas for stats like hit and expertise that don't increase the damage of our abilities directly, but just let them land more often. Once we start factoring in things like cooldowns, procs, and latency, it's clear that theorycrafting is not something that can be done on the back of a napkin.
Distilling years of work into a half-dozen numbers
Trying to create a working theorycrafting model for rogues is an incredibly complex task, as there are simply so many moving parts. It's impossible to say that one piece of gear is better than another without first comparing the two in detail. The problem is that the value of each piece of gear varies based on our current gear, meaning that two different rogues can come to two different results. Even as we equip a new item, it changes the value of other items, making it impossible to always guess which gear is the best at any given time.
In order to solve this problem, rogues began using the idea of EP, or equivalency points. The idea is that we normalize the value of any stat to a single point of attack power, or 1 EP. If critical strike is twice as good per point as attack power, it has a value of 2 EP, since it's worth 2 attack power. Of course, the value of critical strike in relation to attack power is constantly fluctuating as we gain new gear, and so any general EP number is by definition an estimate. However, with small gear upgrades, the change in EP values of various stats doesn't shift so much that it changes the order or preferred stats for rogues.
Thus, we can use EP as a valuable tool for quickly evaluating the value of a piece of gear, but we have to temper that with the knowledge that unless we compute the EP values for our rogue at that exact moment, they're only an approximation. This is especially important if two stats are nearly equal in value, as we have to figure out if one has become more powerful after recent gear upgrades. Ghostcrawler was right when he said WoW was a game about upgrading your stuff, but I think that it's pretty clear that even that seemingly innocent task can be incredibly complex.
Filed under: Rogue, (Rogue) Encrypted Text






Reader Comments (Page 1 of 2)
S i N Nov 18th 2010 9:27PM
Wait, what?
Surtur Nov 17th 2010 9:40AM
Did I miss any information on the Gunship not working in ICC for many people? From what i know no one from Blizz is talking about it.
Surtur Nov 17th 2010 10:01AM
finally found it on the new forums!
Still surprised nothing was said about it.
xiani Nov 18th 2010 4:22PM
Awesome stuff!
Don't forgot about procs . And set bonuses. And cooldowns & haste stacking multiplicatively, and the relative value of a stat depending on whether it's affected by raid buffs, and all the rest...it's enough to make you want to be a healer or something! Though I expect they're probably up to their necks in maths too.
Makes my head hurt just thinking about it anyway, and I'm very glad there are people out there doing all the hard work so us mere mortals can get on with killing stuff faster!
guest42 Nov 17th 2010 10:11AM
Is it possible for someone to figure out things like hard and soft caps on stats, and how long our GCD is based on our haste rating without wading through 137 pages of obsolete elitistjerks posts?
Tinwhisker Nov 17th 2010 10:16AM
All of the cap info is in the "combat ratings" thread, first post.
The rogue GCD is unaffected by haste. It is 1 second at zero haste and 1 second at 1000 haste.
Chase Christian Nov 17th 2010 10:48AM
In Cataclysm, there will only be caps on Hit and Expertise, although even those may change as more difficult dungeons are unlocked. We shouldn't run into any haste, crit, or mastery caps during Cataclysm.
Metalupis Nov 17th 2010 11:53AM
Technically doesnt combat does have a mastery cap (unless they changed main gauche when I wasn't looking) caps/ when you get a 100% chance to proc an extra attack
Xayíde Nov 17th 2010 12:22PM
Yes, as there will still be a theoretical crit cap. What I believe he meant is that we probably won't be able to achieve them realistically even with the latest tier gear in Cataclysm.
Boobah Nov 17th 2010 7:28PM
"Technically doesnt combat does have a mastery cap (unless they changed main gauche when I wasn't looking) caps/ when you get a 100% chance to proc an extra attack"
Technically? Yes. But if it looks like that's going to happen (we'd have to have some serious stat inflation beyond what is planned) they'll 'fix' the mastery so it's not an issue, whether by making it not worth stacking that much anyway, allowing a second chance at an attack if you've more than 100% Main Gauche, or maybe just rescaling what the % chance per point of Mastery is.
We've already seen the devs' reaction to threatened mastery caps for both Prot specs.
CrossEyed Nov 17th 2010 10:22AM
So... Battleground Rogues... Combat with its MUCH improved survivability, does it look good for Cata?
Hank Nov 17th 2010 10:55AM
I want to play a game, not do math homework...
CrossEyed Nov 17th 2010 10:59AM
That's why you let OTHERS post spreadsheets... then you use theirs... and profit
Kurash Nov 17th 2010 12:32PM
Then play the game and don't worry about the math. Easy enough.
But if you enjoy doing number crunching there is no reason not to do it. Hell, I know history isn't everyone's cup of tea, but I read history books for fun. To each his own, eh?
Eyhk Nov 17th 2010 2:21PM
That's probably one of the main reason why one rogue with ilvl 277 gear does 8k damage while another rogue in 251 gear does the same 8k damage.
prae Nov 17th 2010 10:58AM
Why are things made equivalent to attack power when agility has always been the primary stat for rogues, and, secondly, when the concept of Agility Equivalence Points have been around for ages?
Xayíde Nov 17th 2010 12:36PM
I believe this is because numbers between 0 and 1 are not very friendly... Attack power, point-by-point, is worse that any other stat; on the other hand it costs half the gear budget. So, when you compare other stuff with AP, they will (almost) always have an EP greater than 1. This way makes it easier for people to understand.
Let me try a crude example:
If Agi had double the value of AP:
Using Agi Equivalence Points - Agi: 1, AP: 0.5
Using AP Equivalence Points - Agi: 2, AP: 1
Both are saying exactly the same thing, but the second is easier to understand for people who are not comfortable with numbers.
Jsbaldo Nov 17th 2010 6:09PM
Probably because agility also grants crit, making the calculation more difficult. AP very simply makes everything hit harder on every swing.
Xayíde Nov 18th 2010 7:23AM
That can't be the reason because calculating the value of Agi based on AP is the same as calculating the value of AP based on Agi.
Chase Christian Nov 18th 2010 9:33AM
The value of agility changes more rapidly than the value of attack power, which is another reason that AP makes more sense.