Page 3 of 3 FirstFirst
1
2
3
  1. #41
    Simcraft default settings are 25-man + BL at pool.
    Last edited by mmokri; 2013-12-25 at 10:26 AM.

  2. #42
    Never said it's impossible, Whugzz just very unlikely =) good players will often reach 90% and above yes, but this is barely possible on most fights, that's the point.

  3. #43
    Quote Originally Posted by miffy23 View Post
    Never said it's impossible, Whugzz just very unlikely =) good players will often reach 90% and above yes, but this is barely possible on most fights, that's the point.
    Ok, let's be honest you did say it was impossible The point is that you can achieve 100% and beat simcraft. You are constantly being negative and putting limitations on not just your dps but the people looking to improve's dps as well.

  4. #44
    Quote Originally Posted by whugzz View Post
    Ok, let's be honest you did say it was impossible The point is that you can achieve 100% and beat simcraft. You are constantly being negative and putting limitations on not just your dps but the people looking to improve's dps as well.
    You just don't seem to fully understand what "100% of Sim" means. You're reaching a high percentage of it, and that's good, but it's not an absolute comparison (conditions will ALWAYS be slightly different).
    You can't "beat" Simcraft. It's not a challenge, it's a mathematical approximation of the theoretical optimum.
    You know how it's easy to "beat" Simcraft for example? By having your raid DPS be that much higher and outgearing a fight so it's duration is shortened. Simcraft simulates a 5 minute Patchwerk style fight. If you beat the fight in less, your bursts weigh much higher in your average, for example. That's not "higher than Simcraft" by definition, since it wasn't the same timespan.
    Last edited by miffy23; 2013-12-26 at 10:42 AM.

  5. #45
    Quote Originally Posted by miffy23 View Post
    You just don't seem to fully understand what "100% of Sim" means. You're reaching a high percentage of it, and that's good, but it's not an absolute comparison (conditions will ALWAYS be slightly different).
    You can't "beat" Simcraft. It's not a challenge, it's a mathematical approximation of the theoretical optimum.
    You know how it's easy to "beat" Simcraft for example? By having your raid DPS be that much higher and outgearing a fight so it's duration is shortened. Simcraft simulates a 5 minute Patchwerk style fight. If you beat the fight in less, your bursts weigh much higher in your average, for example. That's not "higher than Simcraft" by definition, since it wasn't the same timespan.
    ok more nonsense from miffy, i'm done arguing.

  6. #46
    Scarab Lord Raiju's Avatar
    Join Date
    Apr 2011
    Location
    England
    Posts
    4,506
    Quote Originally Posted by miffy23 View Post
    You just don't seem to fully understand what "100% of Sim" means. You're reaching a high percentage of it, and that's good, but it's not an absolute comparison (conditions will ALWAYS be slightly different).
    You can't "beat" Simcraft. It's not a challenge, it's a mathematical approximation of the theoretical optimum.
    You know how it's easy to "beat" Simcraft for example? By having your raid DPS be that much higher and outgearing a fight so it's duration is shortened. Simcraft simulates a 5 minute Patchwerk style fight. If you beat the fight in less, your bursts weigh much higher in your average, for example. That's not "higher than Simcraft" by definition, since it wasn't the same timespan.
    If we're using the simcraft calculated dps then why can't you? I was under the impression it ran perfect play for say 10000 iterations to account for RNG. Say you got very good RNG on a fight (that was same length effectively patchwerk etc.) and played near perfectly, could you not beat simcraft?
    Quote Originally Posted by Endus View Post
    which is kind of like saying "of COURSE you can't see the unicorns, unicorns are invisible, silly."

  7. #47
    @miffy23 - its incorrect comparison, because you should set time of fight according to real time of fight.
    @raiju - you are not going to achieve rightmost number ever except AOE or better rotation.
    Last edited by mmokri; 2013-12-27 at 05:20 PM.

  8. #48
    mmokri has the correct answer. You can, of course, see yourself make higher numbers than SimulationCraft when any of a few things are happen:

    1. The fight time is shorter than the sim you're comparing with. (Or it's just long enough to let you use another phase of cooldowns.)

    2. Your raid composition is different from the sim, so you're benefiting from more cooldowns than were simulated (Skull Banners, etc.).

    3. There are AOE or increased vulnerability phases in the fight you're doing, so you get to do more damage to additional targets.

    4. Extreme RNG.

    So, yes, mmokri is right about making incorrect comparisons. By default, SimC gives you your basic, single-target DPS output vs. Patchwerk with an ideal (all buffs) raid composition. Anything that is different from those conditions will skew your numbers up or down.

    Y'all are arguing different points about the simulations. Yes, you can do more DPS than SimC says you should. But SimC is making an estimate in a specific set of circumstances. If your boss fight has different circumstances -- and most do -- then your damage output may be vastly different from what SimC predicts.

  9. #49
    Scarab Lord Raiju's Avatar
    Join Date
    Apr 2011
    Location
    England
    Posts
    4,506
    Quote Originally Posted by mmokri View Post
    @raiju - you are not going to achieve rightmost number ever except AOE or better rotation.
    I don't believe I've ever seen anyone citing the top number when using simcraft, personally. They use their mean result from their chars sim over 10k iterations. See Hekili's point 4.
    Quote Originally Posted by Endus View Post
    which is kind of like saying "of COURSE you can't see the unicorns, unicorns are invisible, silly."

  10. #50
    @Raiju seems I speak not clear enough, I am going to rephrase.
    If simcraft is bug-free then you can reach every number bellow rightmost with small help of RNG.
    Last edited by mmokri; 2013-12-27 at 09:04 PM.

  11. #51
    Quote Originally Posted by whugzz View Post
    ok more nonsense from miffy, i'm done arguing.
    Others explained it better than me above this post, I tend to use a lot of big words sometimes.

    It's the holidays so I haven't had much time to test anything, but I will be testing SS pre burst and reintroducing 4MW LB again and log it when I get the time.

  12. #52
    Quote Originally Posted by miffy23 View Post
    Others explained it better than me above this post, I tend to use a lot of big words sometimes.

    It's the holidays so I haven't had much time to test anything, but I will be testing SS pre burst and reintroducing 4MW LB again and log it when I get the time.
    Your reply did not make any sense because you can reach the simcraft mean and you can breach the simcraft mean IE: Beating simcraft (Which I just proved in my previous posts).

    Yes, i'm treating it as a challenge (you don't have to but I do because I like having goals to constantly achieve in raiding) to beat the simcraft average if I know I can set the conditions of the fight to be about the same as the fight itself.

    I thought that was also assumed, you have to set the conditions of simcraft to match the fight conditions in order to get an accurate dps number goal to reach/breach.

    Obviously, if the conditions of simcraft are way off from the fight you will not get an accurate number to reach.

    The fight in particular we we're talking about is iron jug which you can set it pretty much perfectly to.

    You can set the conditions of simcraft to work for many fights in SoO using the overrides section, not just iron jug.
    Last edited by whugzz; 2013-12-29 at 06:18 AM.

  13. #53
    Since some apparently people don't understand what SimulationCraft results actually mean, let me explain a bit:

    The big huge number at the top of the SimC report is the mean (or average), as in, if you were doing everything 100% the same as the sim, that would be your average DPS. Assuming a normal distribution (which DPS is not, but for most classes it's somewhat close to a normal distribution) that means that 50% of the time you would beat that listed DPS value.

    Again, this is if you're doing everything 100% the same as the sim: Exact sim priority list (decision making in your rotation) with no mistakes/inconsistencies, the same reaction time, the same latency, the same fight duration, the same buffs and debuffs, same lust timing, same target uptime, etc.

    Now if your average DPS does not align with the SimC DPS value, then that means that one of these factors in indeed not the same in reality as you have it in the sim. Whether this is because the fight is not truely Patchwerk, because the fight duration you simulated does not match your parse, because you had higher or lower latency or reaction time, because there is bugs/inaccuracies in the sim. It is also likely that some of the damage differential is from a difference ability prioritization (whether intentional or not) but that alone doesn't account for anything close to 10 to 15% of your DPS.

    As already pointed out, most fights are not going to match sims simply because they are not patchwerk and therefore have different numbers of targets, or different amounts of target uptime, or targets that die and your flame shock ends, etc. But on the fights that are at least fairly comparable (Malkorok normal and Siegecrafter normal not doing belts [less on heroic because of adds, in both cases]) it's completely absurd to say that it's impossible or very difficult to achieve the average DPS value simply because the simulation plays "perfectly" (protip: It's not perfect, at all.). In a lot of cases it's going to be that you failed in setting up the simulation in a way that closely parallels the actual encounter, both in content and duration, but beyond that it's completely ridiculously to just say "oh I played really well but the sim is perfect so it's just going to beat me." If you actually did an appropriate job in setting up the simulation in a way that most closely parallels reality, then the next most likely reason is either:

    A) There's some bug or inconsistency in the simulation affecting results (and if you can find such an issue we're always grateful if you can report it to the issue tracker).
    B) You're not playing as well as you could be.
    C) You played well but got poor RNG (or things outside of your control did not work out in your favor).
    (or some combination of B and C)

    Assuming A is not the case, that's where a lot of guesswork comes in really. From there an informed simulator needs to objectively try to determine what could cause such a gap between their parse and the simulated average: whether it's just playing poorly or if there's an actual reason outside of your control, and that can be something that's extremely difficult to determine on an objective basis.

    I can say that as a feral druid (which sims quite highly in simcraft, not compared to in-game but compared to the other specs) it is very very possible to beat the simulated number, and on a regular basis too when you throw out the parses that you got terrible trinket RNG on.

    - - - Updated - - -

    Anyways, my point is:

    Quote Originally Posted by Bzzap View Post
    Simcraft isn't accurate, you'll never reach max potential due to all of the different variables. You're pulling about as much as you possibly can, give or take.
    Not only is this kind of post factually inaccurate, it's not helping the person requesting assistance in any way and you're probably just writing off someone's mistakes or disadvantages on something that doesn't actually exist.

    /endrant
    Last edited by aggixx; 2013-12-29 at 07:16 AM.

    Feral Druid Maintainer of SimC // Moderator at Fluid Druid forums
    Pawkets' Armory
    // Pawkets' Pro Raiders // <Forgotten Aspects>

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •