• Announcements

    • Forum Rules (Updated 3/22/14)   07/14/10

      Welcome to the Elitist Jerks forums, a WoW discussion forum targeted towards topics regarding high-end raiding and analysis of game mechanics. We ask you to take a few minutes to read the following before making use of these forums.   First, a point of clarification. The name of these forums is not intended ironically; we have high standards for the discussion that occurs herein, and we're quite unapologetic about it. If you feel our rules are stupid or arbitrary, we don't really care. If you don't wish to follow them, you're welcome to return to the official Blizzard forums.   Following is a brief listing of our forum rules; note that it is by no means exhaustive, as in our experience people are quite innovative in finding new ways to be stupid. These are simply a set of guidelines to get you started in the right direction. If you follow them, you will generally do fine here; however, if you concoct some creative new form of stupidity, our moderators feel no need to restrain themselves in letting you know. All posters are to make an effort to communicate clearly. In particular, all posts should be made in a reasonable approximation of proper English. We realize that a significant number of you are not native speakers, and we do not expect perfection: merely effort. Please obey basic rules of capitalization and punctuation, avoid chatroom abbreviations ("lol", "imo", "u", and the like), and pay at least minimal attention to sentence and paragraph structure. This includes not starting a new paragraph for each sentence. All opinions should be stated as succinctly as possible. Do not make multiple consecutive posts; rather, multi-quote and include all your ideas in a single post. Do not quote huge blocks of text to add a short reply; instead, quote only what you need to to make your point. Do not break a single quoted reply into multiple blocks; doing so needlessly lengthens your post without aiding its readability. And don't provide unnecessary backstory: if it isn't relevant to the question you're asking or the point you're making, we don't need to know about it. All discussion should be both polite and civil. Trolling or flaming in any form is forbidden. Just because someone disagrees with you does not mean they are stupid or on drugs and their personal hygiene isn't really relevant to the discussion. Regardless of the merit (or lack thereof) of your argument, it should be made in a way that is neither insulting nor condescending. Whining in any form is forbidden. Blizzard is not incompetent or stupid and they are not intentionally screwing you over and neither is anyone else. If all you're going to do is complain, don't bother posting. Threads should be started if and only if there is some reasonable topic to discuss. If the issue you wish to discuss is covered in an existing thread, use it rather than creating a new one. If you are asking a simple question that you expect to have a simple answer, ask it in one of the "Simple Questions/Simple Answers" threads. But if you feel there is a topic of discussion not well-covered by existing threads, feel free to start a new thread to discuss it. Some sub-forums restrict new members from creating new topics unless they've made at least 10 approved posts, to prevent spamming or bad posts. If you really think it deserves a new thread before you have 10 posts, contact a moderator with your post content. Do not post unless you have something new and worthwhile to say. Do not bump, quote for truth, cross-post, or post only to say thanks. We don't want to hear your funny story about something that happened in your raid last night, your baseless speculation is unproductive, and your idea for a new ability really isn't that interesting. We don't care what gear you are hoping to get or just received. If you have an idea you'd like to share with the community, support it with analysis, testing, or both that indicates you've put some thought into it. (Note: Posting of a new untested spec falls under this rule, unless you have done the grunt work and have information to support your amazing new spec don't even bother posting it here.) Do not beg for hand-holding. These are forums for discussion and analysis, not for answering any question that you might happen to dream up. Search and read before posting--do not post a question unless you are fairly confident that the answer isn't widely known or easily attainable. In particular, we do not want to take a look at your armory or WWS to tell you what you're doing wrong and we're not interested in making your tough gear or spec decisions for you. We expect you to use the search function and also to read the first post as well as the last 5 pages of the thread you are posting in. Chances are your question has already been answered. Additionally, do not post asking for confirmation of a simulation result. If you think there is a problem with the Sim you are welcome to PM the author. All accounts must have a valid WoW profile. If you no longer play and have deleted all characters you used to have, you may select the "No WoW Account" option; otherwise, this information must be filled out for your main character. If you fail to observe this rule you'll be permanently banned from our forums. We do not permit anonymous posting. Do not sign your posts. People can see who you are from the profile printed to the left of each post, so signing your posts is redundant and simply takes up space. Similarly, you do not need to link your armory in your post, as if people wish to see it they can get it from your profile. Do not respond to terrible posts. Do not respond to a blatantly awful post (a post that is in clear violation of the rules) either in an attempt to moderate them or to answer a question they posed. Your post will just be warned/infracted and removed with the post you are replying to. If you feel that a post is in violation of these rules, please report it and the moderators will deal with it as we feel is appropriate. No Advertising. Do not make posts solely for the purpose of advertising your site/blog/twitch/etc. You may post such things if it's relevant and adds to discussion at hand. If you have information to share, share it here with a link back to your blog or whatever. Do not post "I have information, come to my site to get it". That will result in an immediate infraction and post removal. Also, we will remove any link to a site that violates a games TOS/EULA such as gold selling sites.

      If you violate a forum rule, you will receive an infraction. Most infractions are worth one point, although we reserve the right to give you more at any time if we feel you deserve them. If you accumulate too many infraction points, you will receive an automatic (usually temporary) ban which revokes your posting privileges, as follows: 3 Points - 1 Day 5 Points - 3 Days 8 Points - 1 Week 10 Points - 1 Month 15 Points - Permanent

      Familiarize yourself with The Banhammer, an archive of all infractions given by the moderators here; it will give you some examples of what not to do. Also feel free to take a gander over The Dung Heap, which will give you a good idea of what these forums would look like if we weren't such jerks.   Thank you for joining the discussion!
    • Site Maintenance (Updated 11-12-15)   10/01/15

      Site has been updated to IPS as of 12 November Known Issues So Far: Old wowhead "item" tags are depricated. They may be re-added to fix some issues with older guides, but for most part it's better to just use the modern wowhead.js which auto parses actual wowhead links automatically into tooltips/names. Just post wowhead links in guides and the js will do the rest. You don't need a special BB code for items. Old article links will be broken do to IPS 4 changing url path to remove some pointless "/_/" directory. The articles are still there and can be found in guides menu. You just may not be able to click links from any threads linking to them without editing out the "/_/" part of url. I'll fix most of this little by little when more important stuff is taken care of. External links/google search results may run into same problem. That will just rely on waiting for google to reindex new paths. After you login, you may be thrown to an error page saying page cannot be found if you logged in from the forum index. Harmless though. Just click "forums" and you'll get forum index since redirect after login is busted. The benefactors bar seems to hang/timeout sometimes.


  • Content count

  • Joined

  • Last visited


About Aldriana

  • Rank
    Mike Tyson
  • Birthday 07/08/81

Profile Information

  • Gender Not Telling

WoW Profile

  • Character Name Aldriana
  • Gender Female
  • Race Night Elf
  • Class Rogue
  • Realm No WoW Account
  • Realm Location United States
  • Guild Drow
  • Guild URL http://www.drow.org/
  1. For the record: I had always hoped to implement more phase-based calculations, but never had time to do it justice.  The initial implementation was little more than a direct port of spreadsheet logic, and spreadsheets can't really support phased solutions.  So kudos to Pathal (and any others that may have been involved) for finally putting together a working implementation (for at least some cycles).   Regarding time to compute vs precision: is there any way to make this tunable?  When one is rapidly iterating on many different sets of gear there is a lot of value in having an "approximate" calculation that will let you know if you're in the right ballpark; but once you're starting to get close - or want to optimize your strategy in detail for a given fight based on what your gear currently is - it could make sense to do significantly longer runs.   This, of course, depends on how different the logical path is between "fast" and "accurate", and also how large the accuracy different actually is.  But, conceptually: sometimes people want an answer in one second, and sometimes people want an exact answer; in an ideal world, it'd be nice to support both (although, if you're going to allow long, detailed evaluations that take more than a few seconds to run, you might need to think about server resources a bit - not sure what the traffic volume looks like).   It should be noted that this also applies to reforging to some extent.  An exact solution for reforging is incredibly time-consuming, which is why approximate solutions are used instead, which is why their is wonkiness.  Having both a quick and dirty solution for rapid iteration, and a "run as long as you need to" solution to get more accurate (and hopefully less wonky) answers would potentially have value.
  2. [5.4 Combat] I'm Not Dead Yet

    It would surprise me if weapon swapping triggered anything untoward with procs provided by other gear (trinkets, etc.). It would not surprise me if weapon swapping resets your "time since last proc" timer for the enchants on the weapons themselves. Of course, testing that's going to be sort of a nuisance - I think the best approach would be to equip the highest-proc-rate stuff you can, wait (without attacking) to get your RPPMs to max proc chance, and then weapon swap and open. Based on time average till first proc you should be able to estimate whether things are counting time since the weapon swap or giving credit for the pre-weapon-swap time as well.
  3. ShadowCraft for Mists of Pandaria

    To elaborate on this a bit: the problem with interrupted fights is that there are fundamentally two sorts of effects of combat interruption: those that are extremely hard to model, and those that don't matter very much. From the perspective of sustained abilities - stats, talents, and the like - the biggest effect of interruptions it that it tweaks your white/yellow damage ratio. You spend some time regenerating energy where you're not attacking, so you effectively have more energy to spend relative to the number of white attacks (and, of course, your overall damage goes down, but that doesn't affect relative values very much). And while there are some subtleties due to energy feedback mechanisms, this isn't fundamentally *that* hard to model. Its just that the last couple of times I tried such a thing, the conclusion was that it doesn't actually matter that much under all but the most extreme patterns of interruption. I don't have exact numbers in front of me, but my recollection is that no interruption pattern which reflected a real (PvE) fight effected stat weights more than about 5% in either direction. If you just use the regular Shadowcraft numbers for endless combat... you might be a little off what's theoretically optimal, but not by enough that you'll actually notice. On the other hand, cooldowns, procs, and other intermittent effects can have their value changed radically by the interruption pattern, and there's simply no way for ShadowCraft in its current implementation to even *try* to model this - mostly because the effects are highly dependent on the *exact* interruption pattern and the triggering conditions of the effect. This is not to say that one couldn't write a modeling framework to handle it - you'd just break the combat into segments based on what you're doing, build up an (extremely) complicated probability tree to keep track of the likelihood of procs and cooldowns being up at any point in time, and the computed the DPS as the weighted sum of each case, as opposed to computing a single case based on average-case effects. Thing is, writing such a thing would take a ton of time and effort, dwarfing all effort that has been put into ShadowCraft to date, and no one has felt it worth spending the time on so far - its not clear what the gains would be, and most of us have better things to do than spending a thousand hours rewriting ShadowCraft. Also, even if it did get written, it'd be a lot less user-friendly, as you'd need to manually specify how your guild does each fight in order to get accurate numbers out. And most people would rather have a simple tool where they get "roughly" the write answer than need to spend an hour spelling out every detail of a fight in order to perfectly optimize for it. So, in short... it hasn't seemed like an efficient use of time. This isn't to say that that the "unlimited combat" assumption isn't a problem - it is - its just a problem with no good way of solving it. At least, not without massively more time and effort than anyone has been willing to spend so far. The fundamental problem with theorycraft is that its our hobby, not our job, and that puts limits on what can be achieved.
  4. Rogue Simple Questions

    I doubt they'd specifically put in different refresh mechanics based on size, but they could, for instance, prevent replacing a longer buff with a shorter one, which would solve this problem in most cases.
  5. ShadowCraft for Mists of Pandaria

    At a guess: line 1452-4: if ar: self.attack_speed_increase *= 1.2 self.base_energy_regen *= 2.0 Assuming you pass through this section of code repeatedly (which is common for the inner loops of damage calculation) its pretty easy to see how this would result in vastly inflated energy regen.
  6. Assassination from the Mists

    Short answer: Pandas are new in Mists, and it hasn't been worth the money to people to change. But they really are that much better, assuming you have stat food on (which in raids, is most of the time).
  7. Assassination from the Mists

    And Mutilating at 3 CP can proc SF and waste a combo point as well - in fact, its more likely - yet we do it anyway. And I suspect its optimal to Dispatch at 4 as well for all of the same reasons, although I admit I haven't analyzed the problem in enough detail to say for certain. If you have evidence to the contrary I'd be interested to see it, but I might suggest not flatly asserting that the previous poster is wrong unless you have such evidence.
  8. ShadowCraft for Mists of Pandaria

    Because the final number at the end isn't the only thing that matters. If there are multiple reforges close together in quality, some people want to know that. And those people that do only care about the biggest number can still easily retrieve it by re-selecting the pre-reforge gearset.
  9. Assassination from the Mists

    Sure - but I don't think that changes the fundamental point very much. Even if you assume ST lets you do just as much damage from ranged as you do from melee (and it doesn't), you still need a couple of seconds per minute of time off target to offset a 5% deficit the rest of the time, and (with judicious use of movement cooldowns) that's usually more than you get from target switching alone. Its basically only going to come out ahead on fights where you have specific phases where you're stuck off the boss. Hence: Hagara Ice Phase? Sure. Atramedes? Great. And there's probably a couple of other examples that you can put together as well. But its definitely a fairly scant minority of fights.
  10. Assassination from the Mists

    Well, here's the ballpark figure for the "seconds at range" calculation. Lets say its a 5% damage loss for our initial estimate. So you do 5% less damage while at range, and (say) 25% of your baseline DPS while at range in exchange. In order to be worthwhile, you must spend at least 1/5 as much time at range as you do in melee - so, 10 seconds out of every minute of combat, or a minute or so over the course of a typical fight. If you tweak the assumptions - say, 3%, and 30% damage while at range - that number drops to 5 seconds per minute of combat, or ~30 seconds on most fights. Still not really enough. With Shuriken Toss, of course, your damage from range will be a lot better - perhaps 70% (as you get autoattacks and autoattack poison, but still not envenom, rupture, venomous wounds, envenom buff, etc.), which drops that number to more like 3 seconds per minute of combat, which is starting to get to the point where you'll hit it on the occasional fight... but its still going to be rare. And of course, if you're spending that much time at range, odds are good its because its a fight with lots of target switches, in which case Marked For Death might be a more relevant comparison than Anticipation. But, of course, we don't have any numbers at all on that yet, so its hard to say for sure. Long story short: I wouldn't bet on it being ahead from a purely numerical standpoint on many - or any - fights. With Sprint and Shadowstep, we just don't spend that much time away from targets. Blizzard just doesn't make that many fights where melee have to sit at range and watch, because those are the sort of fights that wind up with melee sitting in favor of more ranged. If you find it sufficiently more fun that you're willing to eat the DPS loss (and can sell your guild on the notion), so be it; but it is a DPS loss.
  11. Assassination from the Mists

    The short answer is that we don't have the modeling to answer this question in detail - at least, not in ShadowCraft. And frankly, I suspect that's unlikely to change - given the limited amount of time people have to spend on ShadowCraft improvements, modeling a clearly and obviously inferior spec is not likely to be high on people's priority list. Hence, here is the answer I will give: You clearly lose some damage. It is probably on the order of the 6.5% estimate above; you will gain some damage by doing more efficient finishers, but you also lose some by spending energy on less efficient CPGs. It is probably less than that 6.5% estimate, but whether its 3% or 5% or 6.4%, I really couldn't say. That said, I don't think it really matters that much. If you're in a competitive high-end guild, its too much - you're hurting your guild's progression. If you're in a highly casual, social guild, no one is going to care. If you're somewhere in between, you'll have to make that decision yourself, and no amount of quantifying the exact DPS drop is going to change that.
  12. ShadowCraft for Mists of Pandaria

    Not sure which branch you're looking at, but: https://github.com/Aldriana/ShadowCraft-Engine/blob/master/shadowcraft/objects/stats.py#L164
  13. ShadowCraft for Mists of Pandaria

    The thing to keep in mind with ShadowCraft is that its really not a single monolithic tool. There's at least 3 distinct components, written by different people, involved. Notable in this case is that the logic that computes DPS and EP values is entirely distinct from the reforger/gem optimizer. The reforger is a separate piece of logic that takes the EP values provided by that module and attempts to find the best possible reforge. So while the reforger has been known to make some slightly strange decisions, that doesn't mean the EP values and DPS estimate are in any way in doubt; a problem with the reforger and a problem with the DPS calculator are two very different things. I will also note that this isn't intended as a knock on the reforger; reforging is actually a very hard problem, and to guarantee the globally optimum reforge would be ridiculously computationally intensive. As such, by necessity it uses heuristics and simplifying assumptions to find a "pretty good" reforge in a reasonable amount of time. However, this also necessarily means that sometimes it will miss things, so there will be situations where you can manually fiddle the EP values to increase the DPS estimate.
  14. Assassination from the Mists

    Just to clear up a bit of misinformation that seems to circulate around the forums: there is no inherent benefit to pooling. Pooling to 85 energy (or any other amount short of cap) before you Mutilate is 100% equivalent to Mutilating the second you have enough energy for the move. There is no magic level of energy that you want to always pool to. The advantage of pooling is and always has been that it allows you to adjust the timing on your moves to be more advantageous. The benefit is entirely in terms of increasing Envenom uptime by spacing them out more, spending more energy while you have cooldowns or other buffs up, minimizing rupture downtime, and so on. Hence, advise like "if your previous envenom has not dropped wait until it does or until your energy is in danger of capping before casting the next one" will increase DPS. Advice like "go into Shadow Blades/Vendetta with high (but uncapped) energy" will increase your DPS. Advice like "try to have enough energy that you will be able to cast two Mutilates during the buff before casting Envenom" will increase your DPS. But any recommendation of the form "always pool to X energy before casting Y" is going to be wrong. I will also note, for the record, that there is nothing you can do to optimize your cycle that will increase your DPS by 10k unless you were doing things wrong to start with. That is: while the generally recommended cycle on these forums may not be perfect, its provably not 10k worse than perfect. If you're seeing a 10k improvement purely through cycle adjustments you've made, one of two things is happening: 1) RNG 2) You're better at playing your "new" cycle than you were the old one, such that you're letting energy/CP cap less, wasting less energy/CP, etc.
  15. Assassination from the Mists

    But it can proc Seal Fate and thus give 2. And since wasting CP - even Seal Fate procs - is bad...