Life in Group 5 – A Resto Shaman Blog
A resto shaman perspective on raiding


December 21, 2009

The Line Between Theory and Application

More articles by »
Written by: Vixsin
Tags: , , , ,

Months ago, when I transferred to Mal’Ganis to join the ranks of Aftermath, I was excited for two reasons—first, because I was joining an established and consistently high-ranked raiding guild; and second, because my new stomping grounds would be the #1 US Server (according to, home to one particular group of players that I’ve followed for a long time—Elitist Jerks. Technically speaking, it’s the forum community that EJ has fostered which has been the recipient of my star-eyed gaze ever since I entered into the raiding world; it’s a community that extends beyond the limits of their own guild and includes top players the world over. The EJ forums are known for hard rules, hard math, and hard facts. They’re known for having the answers; they are the epitome and the apex of theorycrafting. But if there’s anything that I’ve learned from following all of their discussions, it’s that there is a significant difference between theory and application, especially when it comes to healing.

The Theory

WoW is, at its very core, a game of variables, a series of conditionals all united by formulae. Stats, spells, effects, buffs, etc. all factor into determining your calculated contributions to the fight at hand. And, as with anything in which weighted values are applied to variables, an optimal condition exists. These optimal conditions are often referred to as BiS Lists (Best in Slot) and max dps rotations, designed for the sole purpose of providing players with the “best” possible available stats and output potential.

And such is the community of nerds that we live in that players have taken it upon themselves to design and modify tools to help others achieve the same goal. Over the years, I’ve tried and tested a variety of simulators and even utilized a few of them in making my own gearing decisions. I’ve created character sets in Rawr, looked over gear rankings in Maxdps, and toyed around with spreadsheets, programs and the like, all for the sake of optimizing my character. (Heck I’ve even started a blog to discuss all this theory with people I don’t know!)

Two of my favorite tools at the moment are Daidalos’ Resto Shaman Healing Calcs spreadsheet and Stassart’s Shaman HEP program, both designed to help me achieve higher levels of performance through gearing and stat choices. I recently highlighted the rotational calculations proffered by Daidalos’s spreadsheet, and concluded that some of my highest HPS rotations are based around CH spam and RT+HW. And even then, faced with the cold hard numbers provided by some of EJ’s finest minds, I still choose not to follow some of the suggestions they make. Why?

The Application

Every system model (be it financial, engineering or another) is only as good as its assumptions, and unfortunately for healers, the assumptions for our models are plentiful. Unlike dps, healing is a dependant raid role. While our ranged and melee friends have a finite goal for dps (equal to the boss’s health divided by a limiting time factor), healing has no upper output limit.* There is a general level of healing required for any fight, with peaks and valleys based on encounter cycles and variables, but total healing for a fight can vary significantly depending on the skill and strategy of your raid.

Looking at Aftermath’s kills for Heroic Northrend Beasts over the past month, you probably wouldn’t be shocked to know that 37k separated the upper and lower limits of the total damage done during the encounter. Spread out over the 17 dps we normally bring, this means that were everyone operating at the same dps output, each player’s damage would only need to vary by 2k apiece to cover the spread. Alternately, the total healing done over the past month has an upper and lower limit separated by almost 1 million healing. This means that each of our healers some nights had to pump out an additional 170k effective healing over the duration of the encounter. Other nights, there was simply less to heal.

The point here is that no matter how brilliant the theorycrafting is, it simply cannot expect to simulate a role where that sort of differential in output exists. Creating max HPS rotations is one thing, but when there are so many assumptions tacked onto that evaluation–movement, duration, incoming damage, distribution of the raid, area effects, boss mechanics, healing team, raid placement, strategy, etc.–then the value of the theory diminishes. That is not to say that the data is wrong, simply that it is incredibly subjective. For example, what good is it to use a relic which buffs my CH if I cannot maintain its uptime due to inherent movement in the fight? Or, what good is it to use a high HPS rotation if my target is taking minimal damage? In the end, it is up to the healer to understand why the minds at EJ and other sites offer the advice they do, and to use his tools appropriately, with careful consideration of the fact that they may or may not be applicable in every situation.


Early in my graduate studies, I took a class about system failure analysis. It was chock full of formulae, statistics, physics, etc., and although the work entailed a high degree of calculational rigor, the class was in fact designed to teach one thing—your modeling assumptions will never reflect reality. Instead of using a model to provide an absolute answer, our professor suggested, models should serve to help you better understand your element of focus, to define those things which will effect an impact, and disprove alternate hypotheses. Models represent but one set of variables in a constantly-changing environment. And the reality is, healing is more about finding the right fit now, than it is about finding the perfect fit always.

* It will be interesting to see how this plays out when we fight Valithria Dreamwalker, the new ICC boss where healing output capacity will play a role in completion of the encounter.


  1. Too true! Modeling is all well and good, but you gotta go get your hands dirty to really know it.

    However, I think in your post-script, you are referring to Valithria Dreamwalker, not Sindragosa.
    .-= Zigi´s last blog ..Updated Emblem of Frost Pull List =-.

  2. Right you are, sir. I attribute the incorrect reference to being positively giddy at the prospect of a healing-focused encounter!

  3. Good stuff. This is why I get frustrated when folks swing by asking for BiS lists and “best” this and “best” that, even for DPS. Reality is so much more nuanced and … messy. If you don’t understand how people have arrived at their conclusions for what’s “best”, then chances are you won’t really get the best out of the things they have concluded are “best”. If that makes sense.

    Though I take theorycrafting very seriously, I find that for my resto spec, my playstyle and choices for talents and gear can be rather different from what some folk think is “best” – and that again differs from what other folk think is “best”. But I understand how these conclusions were reached and can apply them appropriately and discerningly to my situation. There tends not to be much wiggle-room for DPS classes (though there’s more than some people think), but healing is definitely a very special beast where a great multiplicity of styles and solutions can be worth considering. And what works for one player in one situation can be simply terrible for another player in a different circumstance. I guess that’s one of the fun things about healing.

  4. I see your point. My conclusion is a different one though.

    Modeling stuff through shaman_hep or enhsim is as close as we can get to what’s really happening. I don’t think though, that the EJ, spreadsheets and sims you mention are a good example for people who fail to apply theorycrafting to the virtual reality. EJ and the authors you mention are very self-critical. Whenever new numbers pop up on EJ, people join the discussion, criticize, do their own math and also do exactly what you’re doing: think about the viability of the math in-game.

    When I read your post, I somewhat get the impression, that there’s a nerdy, very number-centric EJ world vs. the “us real players”-base. I don’t think that’s the case. The discussion whether numbers are actually viable in the game is very lively on EJ.

    What you describe, the unreflected application of some half-knowledge, gotten from some 5 mins cross-reading on EJ, is very common though. But this hasn’t anything to do with the viability of the numbers we get through a sim.

    If RT-HW is the highest HPS rotation, that’s how it is. There are a lot of good reasons NOT to switch from RT-LHW to RT-HW. But there’s also good reasons switch to the bigger heal. Predictable damage, enraged mobs, whenever you have the possibility to precast: excellent situations to switch to HW, not doing so is very often just laziness. Which is somewhat understandable, because the situations where HW is better than LHW are rare, but they are there and ignoring them isn’t something I’d recommend.

    A part of what you describe is a healer-specific problem: As long as we get a long, we often don’t really think about what we’re doing. DPS players have a great tool to min/max: DPS meters. In a pure single DPS fight, the meters just don’t lie. Healmeters though are crap for benchmarking performance. So as long as no one dies, everything is ok. I’d even argue, that it’s much much easier to get unnoticed as a slacking healer than as a slacking dps. A bad dps will always be at the bottom of the meters. A bad healer: not necessarily. Also, good healers will often cover up the mess bad healers make.

    So, if numbers show us, which the highest HPS rotation is, that’s just how it is. So if the fight needs high HPS, that’s your tool. Not much to argue there. Of course this doesn’t take into account the specifics of a boss fight. But why should it? It’s modeling the situation, where I’m standing still and high HPS is needed. Incinerate flesh -> Highest HPS is a good idea. Precasting HW on tank in P3 so the heal lands exactly when the mob gets out of the stun -> good idea.

    A completely other thing is the interpretation of numbers: shaman_hep is a good example: It values haste much higher than SP why? Because overheal is valued as bad and thus “good” haste leads to higher effective healing while “bad” SP just sinks into unnecessary overheal? This is debatable and leads to the old discussion: “is snipe healing good or bad?”.

    My thoughts about theorycrafting and in-game reality is this:

    I’d say there’s three things that define players: “knowledge”, as in theorycrafting, “skill” as in “I’m also a good FPS gamer” and “social competence” as in “brilliant players with the social competence of a slice of bread are often no worth the trouble”. I’ve seen people with horrible rotations pull impressive DPS, just because they are fast and have a high situational awareness. We all know the guys who do their math and know their stuff but somehow still suck. We all know that reading EJ doesn’t automatically lead to never standing in the fire.

    Some basic interest in theorycraft is good.
    Real interest for the numbers is even better.
    Making theorycrafting a holy grail is bad. It’s just one aspects of many, who define the game and you as a player.
    .-= drug´s last blog ..T10 Shoulders/UI WIP =-.

  5. Monsieur

    Very good post(s) Vixsin (and drug)! I usually split the two “schools” into two different parts of the game, so I don’t really think theres any difference, at least for healers. EJ and theorycrafting are for preperation. Intuition, speed, flameaviodance and great whack-a-mole skills are the execution. Its like stats! Haste(execution) is our best stat, but it wont work as well or at all without the basic int and mp5(theorycraft). And the other way around, being a slowpoke with mana pouring out your nose.

    Great job at stupidifying what the other guys said, Monsieur 😀 *pat on my own back*

  6. Monsieur

    Oh! I forgot my two sidetracks!
    1. Drug; its great how you write posts in the comments for Vixsin, you guys are like a married couple and this blog is what, 1 month old now? (but plox dont neglect shieldsup 😀 )

    2. About the dreamwalker fight, i havent really thought that much about it before, but in terms of raw hps, won’t other healers bust our chops on that fight? Not that im worried about my raidspot for now, but if the hardmodes are(ghasp!) hard, should we be worried guilds will just stack healbombing paladins for the only healercentric fight in the game?

  7. Very well said Drug, on all counts. I think my underlying concern with all of the modeling and simulation programs out there, and accompanying theorycraft, is that they’re too often taken to be acceptable representations of reality. I think back to the engineering students in my class who couldn’t quite grasp that their models weren’t wrong because of improper calculations (eg: the tension cables in the bridge weren’t going to fail because the student didn’t know the value of gravity), the models were wrong because of the variability inherent in what they were attempting to model. Thus, calculations for acceptability were better reflected as a range, or a minimum, rather than an absolute. As you pointed out, taking the numbers as absolutes, as end-all answers, is simply bad practice. (And BTW, I absolutely can attest, that EJ members do on occasion, stand in fire! *GASP*)

    And thank you Monsieur, for the (I think) compliments. As to Dreamwalker, I definitely am concerned that the tendency might be for guilds to turn pallies. But, from what I saw from the PTR testing, there are so many other elements to the fight that stacking all long-bomb healers isn’t viable. (Cuties Only brought 8 healers to their PTR kill, only 2 of which were pallies.) So I’m not sweatin my raid spot … just yet.

  8. In a sense, you are referring to what we would call in layman’s terms as Street Smarts vs. Book Smarts, I think. Well, at least that is what I took from it!

    A person can know all of the answers to things in life, but that doesn’t mean that they can apply them well or have adequate social skills to deal with what they may encounter. For example: I went to college with a violinist who started playing at a very young age under the “Suzuki Method”. This is a very successful method that teaches playing my ear, and focuses heavily on the technical aspect of music. She studied this method for the entirety of her childhood and early adulthood. By the time she entered college she had perfect pitch, could play just about anything by ear and had techinique that was to die for. But, she was missing something very, very important: The Music. She lacked the feeling, emotion and passion that should have gone into her performances. In the end, she ended up playing very well…but was a robot, essentailly playing without her own thought. So while she skill beyond many people, lesser skilled players brought more to their music and were far more thrilling to listen to.

    I think WoW is very much the same way when it comes to the “theory” or “BiS” conversations. You can plug the numbers into all the spreadsheets and calculators that you want, but it won’t teach you how to play…and more importantly it won’t tell you what *feels* right for you in any given situation. Sometimes your experiences and street smarts have got to kick in, and you’ve got to think a bit for yourself. After all, we aren’t all robots!

    I certainly want to be intellegent about my play, but like with everything in life I’ve learned that a little flexibility and listening to “my gut” are just as valuable.

  9. […] you’re looking for actual content: I love Vixin’s latest post on viability of theorycrafting in the real raiding […]

  10. Yep I agree with this.

    This is why I don’t get too worried about people complaining about assumptions that i’ve made in my theorycrafting. At the end of the day player skill and experience are at least as important as gear.

    Theorycrafting should only be taken as a guideline.

    Gobble gobble.
    .-= BobTurkey´s last blog ..Healing priest trinkets for 3.3 =-.

  11. […] The Line Between Theory and Application An interesting post over at Life in Group 5 that touches a bit on some conversations I’ve been having recently about theorycrafting.  […]

  12. […] question was posed back in a previous post whether the fight would be trivialized by stacking one of the healing classes and I think the […]

Leave a Reply

Your email address will not be published. Required fields are marked *

CommentLuv badge