jbuschell Posted April 9, 2012 Share Posted April 9, 2012 (edited) http://ambient.enjin.com/home/m/2994108/article/669668 Madness sorc parsing at 1400 second highest dps, is this an abberation or is everyone crying for a reason, I'm just curious. Edited April 9, 2012 by jbuschell Link to comment Share on other sites More sharing options...
Tenacity Posted April 9, 2012 Share Posted April 9, 2012 From my sorceror's perspective at level 31 right now, 1.2 is almost a pure improvement with my spec. The chain lightning change to wrath is a nerf, but since I never played with wrath + chain lightning, I'm not used to using it and it wont feel like a nerf for me. Beyond that, the 20% damage bonus from wrath will apply to the entire effect of crushing darkness now (which is what I primarily use wrath procs for), which is a damage increase. We'll be losing a bit of damage on force lightning, and that'll add up considering it's our primary damage spell for madness spec, but overall I think madness is still in good shape with 1.2's changes. Link to comment Share on other sites More sharing options...
Rikeryo Posted April 9, 2012 Share Posted April 9, 2012 it's pve. everyone actually cries about pvp. for a reason. Link to comment Share on other sites More sharing options...
warultima Posted April 9, 2012 Share Posted April 9, 2012 (edited) doesnt matter... Say the class average dps is 1500... for all dps... guess what I will take a 1300 dps sorc thats crazy bursty in PvP over another sorc with 1800 dps that cant burst at all. Hopefully people are smart enough to understand why the above statement. I want my PvP damage to be something like 1800... 3500... 3000... 3800... 3500... 2000 over 15 second then 300.250.200.250.200.400.250.200.200.350.375.250.200.150.200.240.230.250.400.320.400.300.180.220.250.200.180.200.230.240.200.250.300 over 60 seconds. Edited April 9, 2012 by warultima Link to comment Share on other sites More sharing options...
Wheelerific Posted April 9, 2012 Share Posted April 9, 2012 doesnt matter... Say the class average dps is 1500... for all dps... guess what I will take a 1300 dps sorc thats crazy bursty in PvP over another sorc with 1800 dps that cant burst at all. Hopefully people are smart enough to understand why the above statement. I want my PvP damage to be something like 1800... 3500... 3000... 3800... 3500... 2000 over 15 second then 300.250.200.250.200.400.250.200.200.350.375.250.200.150.200.240.230.250.400.320.400.300.180.220.250.200.180.200.230.240.200.250.300 over 60 seconds. Then you rolled the wrong class, as Sorcs have the least burst of anyone. Derp. Link to comment Share on other sites More sharing options...
Daellia Posted April 9, 2012 Share Posted April 9, 2012 This is why I hate parses, and why I've been cursing at the community opinion that "parsing makes everything clearer". It doesn't. That is one statistical iteration, where each player's gear, skill, and knowledge of the fight is not only variable but completely unknown. The Madness player's DPS is around 85% of the simulated maximum, but the sniper is only doing about 65% of the calculated maximum and the Madness assassin is only doing about 75% of the simulated maximum. Variations in gear, skill, affinity for the fight, RNG mechanics, and simple luck are the basis behind these differences, and mean that parses such as these given essentially zero reliable information on the way classes are performing. Once we have a large stock of archived parses (on the order of 10k or more), we can start to look at average, but even then average difficulty of the class (and therefore how many people are in the "low skill" category on the parses) and demographic distribution of the classes in high-end raids will affect the totals. Fact of the matter is, parses are incredibly unreliable when it comes to actually determining the dps placement of each class. Parses are best used to identify abnormal game mechanics for integration into simulators. Link to comment Share on other sites More sharing options...
jbuschell Posted April 9, 2012 Author Share Posted April 9, 2012 (edited) Regardless it does show that madness specced sorcs won't be complete dookie and wont be a worthless piece of flaming dog poo, in the hands of a good player they will still be effective to a point. I personally do not believe that one sample parse should be the end all be all of everything, I was just directing it to those who think sorcs will be completely rubbish Edited April 9, 2012 by jbuschell Link to comment Share on other sites More sharing options...
fungihoujo Posted April 9, 2012 Share Posted April 9, 2012 Then you rolled the wrong class, as Sorcs have the least burst of anyone. Derp. Yes... but it wasn't this bad at launch- and you could mix builds to get decent burst and decent healing- which is what a sorc was supposed to be, a mix of both. Now- you pick only one thing... and you're not good at that one thing you're mediocre. But, here we go, another thread trying to tell people talking about how bad a PVP NERF this is by saying 'no, you're ok in PVE'. Fantastic. Link to comment Share on other sites More sharing options...
Hokonoso Posted April 10, 2012 Share Posted April 10, 2012 doesnt matter... Say the class average dps is 1500... for all dps... guess what I will take a 1300 dps sorc thats crazy bursty in PvP over another sorc with 1800 dps that cant burst at all. Hopefully people are smart enough to understand why the above statement. I want my PvP damage to be something like 1800... 3500... 3000... 3800... 3500... 2000 over 15 second then 300.250.200.250.200.400.250.200.200.350.375.250.200.150.200.240.230.250.400.320.400.300.180.220.250.200.180.200.230.240.200.250.300 over 60 seconds. i agree, the 1350 powertech will be doing what you said over 15 seconds then be out of heat doing nothing for the next 90s while vent heat is coming back off cooldown, and certain marauders do even more burst than anyone while doing next to nothing the next 45 seconds, but no one seems to have posted that spec so ill ignore it for now as pvp is about burst not dps, no one cares about dps, currently on live we have burst with recklessness and CL/DF, ofc we cant do it but once every 2min but to drop someone fast it's there. madness will be fine but it will still lack burst, just like warlocks are fine in wow. Link to comment Share on other sites More sharing options...
LordZanos Posted April 10, 2012 Share Posted April 10, 2012 (edited) Then you rolled the wrong class, as Sorcs have the least burst of anyone. Derp. Interesting then that Sorcs do have a spec that's meant to be a bursty turret, but is almost completely useless in PvP. Edited April 10, 2012 by LordZanos Link to comment Share on other sites More sharing options...
Boissi Posted April 10, 2012 Share Posted April 10, 2012 (edited) Atm a hybrid sorc has pretty huge reliable burst potential actually. Double LB FL followed by instant CL and deathfield with recklessness. That's at least 12k damage, potentially much higher with relic/adrenal, in 4 gcds. This of course takes a bit of setting up, like an affliction proc and recklessness up, but I do this combo often, and it usually results in a very dead person unless someone manages to interrupt me. Can follow with a shock crit too if you go for the stalker set bonus, which I have. Well, this is now, but won't be happening in the next patch anymore. Full madness doesn't even get LB. I just can't see myself playing a dps sorc in pvp in 1.2. I don't like dots. Lightning burst doesn't really even compare, but I suppose I'll have to try it one of these days. Edited April 10, 2012 by Boissi Link to comment Share on other sites More sharing options...
XxErebusxX Posted April 10, 2012 Share Posted April 10, 2012 Holy... I can't understand half the stuff on forums cause people like to destroy the english language. Right now none of you said anything smart to me because I can't understand you. People need to realize others actual prefer to go to school and speak correctly. SPREKENZE ENGRISH POR FAVOR? Link to comment Share on other sites More sharing options...
RDeanOU Posted April 10, 2012 Share Posted April 10, 2012 (edited) This is why I hate parses, and why I've been cursing at the community opinion that "parsing makes everything clearer". It doesn't. That is one statistical iteration, where each player's gear, skill, and knowledge of the fight is not only variable but completely unknown. The Madness player's DPS is around 85% of the simulated maximum, but the sniper is only doing about 65% of the calculated maximum and the Madness assassin is only doing about 75% of the simulated maximum. Variations in gear, skill, affinity for the fight, RNG mechanics, and simple luck are the basis behind these differences, and mean that parses such as these given essentially zero reliable information on the way classes are performing. Once we have a large stock of archived parses (on the order of 10k or more), we can start to look at average, but even then average difficulty of the class (and therefore how many people are in the "low skill" category on the parses) and demographic distribution of the classes in high-end raids will affect the totals. Fact of the matter is, parses are incredibly unreliable when it comes to actually determining the dps placement of each class. Parses are best used to identify abnormal game mechanics for integration into simulators. So what is, in your opinion, a reliable and accurate way to determine the dps placement of each class? I'm just curious because I always get suspicious when someone throws out numbers and the response amounts to a simple dismissal of those numbers. Some people do that every time the numbers don't back up their preconceptions. I'm not accusing you of that. I have found your posts to be very reasonable. I'm just curious what kind of facts would be persuasive to you. Edited April 10, 2012 by RDeanOU Link to comment Share on other sites More sharing options...
Daellia Posted April 11, 2012 Share Posted April 11, 2012 (edited) So what is, in your opinion, a reliable and accurate way to determine the dps placement of each class? I'm just curious because I always get suspicious when someone throws out numbers and the response amounts to a simple dismissal of those numbers. Some people do that every time the numbers don't back up their preconceptions. I'm not accusing you of that. I have found your posts to be very reasonable. I'm just curious what kind of facts would be persuasive to you. I tend to put my stock in simulators, who's mechanics are verified by careful analysis of parses. The issue with parses as a raw numbers tool is simply one of statistical variance and uncontrollable variables. Simulators negate both of these issues simultaneously, particularly if they are carefully programmed and verified with current game mechanics. Parses cannot overcome these issues, even if you have an extremely large (talking tens of thousands) repository you're averaging over. Edited April 11, 2012 by Daellia Link to comment Share on other sites More sharing options...
Tenacity Posted April 11, 2012 Share Posted April 11, 2012 The problem with simulators is that they assume you're going to be using an exact rotation without ever messing up or adding/removing abilities from that rotation - and in both pvp and pve that never holds up. Players can try to set up ability rotations as much as they want, but when you really get into a fight, your rotation never holds consistently because you're constantly having to move, use other abilities reactively as the encounter requires, and adjust for anything that happens to the rest of the group. A combat log parser accounts for THOSE variances, by recording the player's performance as he's actually doing the encounters. Link to comment Share on other sites More sharing options...
Daellia Posted April 11, 2012 Share Posted April 11, 2012 The problem with simulators is that they assume you're going to be using an exact rotation without ever messing up or adding/removing abilities from that rotation - and in both pvp and pve that never holds up. Yes, but when evaluating the DPS potential of each class, you need to factor out mistakes and errors. That said, sims are really only truly valid in PvE. If you're looking for your DPS, sure, parses are the way to go. If you're trying to compare classes or specs, however, parses just don't have a sufficiently large or controlled sample size to be statistically valid. Link to comment Share on other sites More sharing options...
Mjolnyr Posted April 11, 2012 Share Posted April 11, 2012 Yes, but when evaluating the DPS potential of each class, you need to factor out mistakes and errors. That said, sims are really only truly valid in PvE. If you're looking for your DPS, sure, parses are the way to go. If you're trying to compare classes or specs, however, parses just don't have a sufficiently large or controlled sample size to be statistically valid. Simulations aren't nearly as accurate as a real long term parse. For rotations that are extremely simple they can be somewhat accurate but on a proc based rotation like the sorc rotation its impossible to get proper numbers from a simulation because of all the rng and variables in between. Plus there are even more mechanics involved than just rotation and user error. That's why we need parses to get REAL numbers not predicted numbers from a spreadsheet or poorly constructed simulation. Once 1.2 goes live and we get the masses submitting results we will get a better understanding but even one person with a standard gear set parsing for 5+ minutes straight will give much more accurate numbers than a simulation. Link to comment Share on other sites More sharing options...
Daellia Posted April 11, 2012 Share Posted April 11, 2012 (edited) Simulations aren't nearly as accurate as a real long term parse. For rotations that are extremely simple they can be somewhat accurate but on a proc based rotation like the sorc rotation its impossible to get proper numbers from a simulation because of all the rng and variables in between. Wait, what? Why would you think those can't be modeled? The game itself is nothing more than a mathematical simulator with a fancy graphical shell wrapped around it. If they game can do it, so can we. In fact, the very point of simulators is to take over where rotations and mechanics become to complex to easily or accurately compute formulaically. Frankly, a computer is a lot better than a person at managing extremely complex or proc-based systems, if for no other reason than that it can compute and react orders of magnitude faster than a human. Plus there are even more mechanics involved than just rotation and user error. That's why we need parses to get REAL numbers not predicted numbers from a spreadsheet or poorly constructed simulation. Poorly constructed? Have you even looked at SimC's source code, or are you just one of those anti-simulator zealots? Point me out a specific example of SimC's "poor design", and we can talk, but otherwise you're just deriding it based on your own personal prejudice. Once 1.2 goes live and we get the masses submitting results we will get a better understanding but even one person with a standard gear set parsing for 5+ minutes straight will give much more accurate numbers than a simulation. You've really no idea how statistics works, do you? Lets compare: Parse Extremely small sample size (single 5 minute fight). Large number of unknown or uncontrolled variables (skill, gear, latency, etc) Simulations Extremely large sample size (10000-50000 3-5 minute fights averaged together). Zero unknown or uncontrolled variables. If you went to anyone that's ever taken a class in statistics (and actually remembers it) and tried to convince them the former was more reliable, you'd be laughed out of the room. The one weakness of simulators is that they need to be very carefully programmed to properly and accurately emulate the environment they are simulating, which is why parses aren't completely relegated to personal evaluation alone. Parses find areas where the simulators may not be accurate. However, given the amount of time and effort the SimC staff has put into making sure it is so, and the amount of time we at sithwarrior have spent double-checking the mechanics and making sure everything works as we think it does, you're going to have to put forth some form of supported argument if you want to suggest that's the case. Simply assuming the simulator is wrong because you dislike simulators is insufficient. Now if you just flat don't like them or don't trust them, fine. Take it elsewhere. We're trying to have a rational discussion here, though, so comments that are so blatantly ignorant of the nature of simulators just waste space. Edited April 11, 2012 by Daellia Link to comment Share on other sites More sharing options...
Mjolnyr Posted April 11, 2012 Share Posted April 11, 2012 Wait, what? Why would you think those can't be modeled? The game itself is nothing more than a mathematical simulator with a fancy graphical shell wrapped around it. If they game can do it, so can we. In fact, the very point of simulators is to take over where rotations and mechanics become to complex to easily or accurately compute formulaically. Frankly, a computer is a lot better than a person at managing extremely complex or proc-based systems, if for no other reason than that it can compute and react orders of magnitude faster than a human. Poorly constructed? Have you even looked at SimC's source code, or are you just one of those anti-simulator zealots? Point me out a specific example of SimC's "poor design", and we can talk, but otherwise you're just deriding it based on your own personal prejudice. You've really no idea how statistics works, do you? Lets compare: Parse Extremely small sample size (single 5 minute fight). Large number of unknown or uncontrolled variables (skill, gear, latency, etc) Simulations Extremely large sample size (10000-50000 3-5 minute fights averaged together). Zero unknown or uncontrolled variables. If you went to anyone that's ever taken a class in statistics (and actually remembers it) and tried to convince them the former was more reliable, you'd be laughed out of the room. The one weakness of simulators is that they need to be very carefully programmed to properly and accurately emulate the environment they are simulating, which is why parses aren't completely relegated to personal evaluation alone. Parses find areas where the simulators may not be accurate. However, given the amount of time and effort the SimC staff has put into making sure it is so, and the amount of time we at sithwarrior have spent double-checking the mechanics and making sure everything works as we think it does, you're going to have to put forth some form of supported argument if you want to suggest that's the case. Simply assuming the simulator is wrong because you dislike simulators is insufficient. Now if you just flat don't like them or don't trust them, fine. Take it elsewhere. We're trying to have a rational discussion here, though, so comments that are so blatantly ignorant of the nature of simulators just waste space. I was un-aware that SimC finished their swtor version, but that doesn't change the fact that simulations only go so far. unless it's in game the numbers mean nothing, and by basing your goals off of a simulation that no matter how detailed its portrayal is still doesn't compare to real parsing. The SimC team does a great job to give a general number but even admit themselves it has it's limitations. Even with a comparison to someone parsing just DPSing a dummy the simulation wouldn't be as accurate and its already been shown that there is a large gap between peoples parses and the simulation numbers. When it comes to comparing the 2 the parse always wins because that's what has actually be done not simulated, parse always trumps the simulation. If we did have the ability to log and parse like before 1.2 then simulations would be the best marker for comparing classes dps, but now that we have logs it's just fun numbers to look at. Link to comment Share on other sites More sharing options...
PutinDoesJudo Posted April 11, 2012 Share Posted April 11, 2012 (edited) I made a thread about how sweet creeping terror is, and basically madness altogether. People seem to have rolled sorcs for the wrong reasons. The whole notion that pvp has to be 100% about burst is pure BS. We melt people - simple as that. No amount of dot cleansing will stop that. It's our force lightning that melts people as much as any dot, it's our true workhorse. Deathfield also does the job of what you should be using chain lightning for anyway, pick off low hp'ers in a group.. it easily does 3k+ using relic and reckless. If you do well now, you will do well in 1.2. Edited April 11, 2012 by PutinDoesJudo Link to comment Share on other sites More sharing options...
Daellia Posted April 11, 2012 Share Posted April 11, 2012 I was un-aware that SimC finished their swtor version, but that doesn't change the fact that simulations only go so far. unless it's in game the numbers mean nothing, and by basing your goals off of a simulation that no matter how detailed its portrayal is still doesn't compare to real parsing. The SimC team does a great job to give a general number but even admit themselves it has it's limitations. Yes, simulators have limitations, but that doesn't mean parses are infallible, nor is the existence of limitations of simulators sufficient reasoning to believe parses are superior by default. Parses have their place, but in terms of accurate comparisons between classes or specs, parses just flat do not possess statistical validity, regardless of being "real" numbers (which, for the record, is an implicit suggestion that SimC's are "fake"). Even with a comparison to someone parsing just DPSing a dummy the simulation wouldn't be as accurate and its already been shown that there is a large gap between peoples parses and the simulation numbers. When it comes to comparing the 2 the parse always wins because that's what has actually be done not simulated, parse always trumps the simulation. How on earth could you sustain the argument that a random person of unknown skill or experience and the smattering of gear they've managed to collect thus far playing the game (since basically no one is in true BIS), DPSing a target dummy without the majority of raid buffs or debuff present, and comparing that to another person of similar unknowns, could possible be more accurate than a completely controlled simulator? Of course there's going to be a gap between simulators and parses. In fact, that's my very point: people are not perfect. While simulators cannot accurately model how a certain person would perform, they can remove that imperfect, so you are directly comparing one class or spec to another class/spec. With parses, you are inherently comparing Person A with Spec/Class X to Person B (or even A) with Spec/Class Y. You can't remove the person element from it. From a formal logic standpoint, you can't take such a pair of parses and derive that Spec/Class X is superior or inferior to Spec/Class Y. All you can derive is that Player A playing Spec/Class X is superior or inferior to Player B (or A) playing Spec/Class Y. We're trying to determine raw class and spec placement here. We need to know how the classes and specs perform at their absolute maximum, without any mistakes or player-based imperfections. Only then can you actually say something about the comparative balance between the classes or specs. Parses cannot give you that, only simulators. Link to comment Share on other sites More sharing options...
RogDodge Posted April 11, 2012 Share Posted April 11, 2012 Daellia, I think that some people are thinking that you are talking about what they should have for a parsed DPS instead of what the class/spec should have. If someone wants to figure out what their DPS is, they should try it out on PTS or wait until 1.2 comes out and try it out live. Please correct me if I'm wrong, but what you are doing is trying to figure out what our DPS should be, if we are under "perfect" conditions, which is a great mark to compare what our max DPS should be. I know I wouldn't want to be the one who says "All Sorc DPS should be based off of my parse." Mainly because I am human and would most likely not have the best of the best gear/buffs/ect. Keep up the great work, love your site! Link to comment Share on other sites More sharing options...
jbuschell Posted April 11, 2012 Author Share Posted April 11, 2012 I'm full madness on my sorc, and my Dmg in wz is perfectly fine, usually 400k, no you may not reach the redic numbers you've seen ranging from 600k to 1 Mil but its defintily still effective Link to comment Share on other sites More sharing options...
Daellia Posted April 11, 2012 Share Posted April 11, 2012 Daellia, I think that some people are thinking that you are talking about what they should have for a parsed DPS instead of what the class/spec should have. If someone wants to figure out what their DPS is, they should try it out on PTS or wait until 1.2 comes out and try it out live. Please correct me if I'm wrong, but what you are doing is trying to figure out what our DPS should be, if we are under "perfect" conditions, which is a great mark to compare what our max DPS should be. I know I wouldn't want to be the one who says "All Sorc DPS should be based off of my parse." Mainly because I am human and would most likely not have the best of the best gear/buffs/ect. Keep up the great work, love your site! Exactly. If they are trying to determine their dps, or which spec/class they are best out, simulators simply aren't an option. Parses are the only source that matters in that case. I was speaking mostly about comparing the raw root capabilities of each class, to determine whether Sorcs really are behind or ahead of other classes (which was what the OP was trying to do with a parse), and that's an area where parses simply cannot hold their ground against a well-built sim (like SimC), or even a well-designed formulative calculator. Thanks for translating my rant into terms perhaps more easily understood by others. Link to comment Share on other sites More sharing options...
RDeanOU Posted April 11, 2012 Share Posted April 11, 2012 (edited) Exactly. If they are trying to determine their dps, or which spec/class they are best out, simulators simply aren't an option. Parses are the only source that matters in that case. I was speaking mostly about comparing the raw root capabilities of each class, to determine whether Sorcs really are behind or ahead of other classes (which was what the OP was trying to do with a parse), and that's an area where parses simply cannot hold their ground against a well-built sim (like SimC), or even a well-designed formulative calculator. Thanks for translating my rant into terms perhaps more easily understood by others. Why is a measurement of what our potential dps is in a perfect world a better way to gauge where we stand than the actual results of performance in the actual game? It seems to me like the people arguing for parses are saying that simulations only measure a potential that may or may not be achievable and that isn't as relevant as the numbers people are actually achieving. If a rotation is really unforgiving and making mistakes really hurts your dps then the simulations will not reflect what we can achieve because we will all make mistakes from time to time. Actual fights in game will force movement and make our rotations mess up too. Why would you not want to account for this when comparing us to other classes? It just seems intuitive to me that what matters is where we actually are coming out rather than where we could come out if everyone playing the game was a computer. Edited April 11, 2012 by RDeanOU Link to comment Share on other sites More sharing options...
Recommended Posts