I have Focus magic turned on. Given that on my guild their is at least one other mage specced into Focus Magic. The only buff I do not have turned on; given that it's rarely any use to us, is Amplify Magic.
If your talking about Focus Magic Rate, I have that at default (0.2)
So the assumption is that at least one other mage will stay arcane or TtW/FB spec to provide the FFB mage the extra 3% crit? I would think that if they are close, and not having to worry about mana in the longer boss fights to come, you may not have that arcane or TtW/Fb mage in the raid. Of course FFB being close counts on that extra 3% crit to get there. Whats the difference in DPS if the FFB mages doesn't have the FM buff?
With my gear in Rawr 2.2.06 the difference is almost 600dps if I spec FFB and don't get a FM buff, vs TtW/FB. But my gear is scewed towards arcane and FB specs.
I think you are still doing something incorrectly, check my values on the previous page (in the character sheet) against yours. Also you never specify what fight length you are using. The gap between fire and ffb is largest in the middle, ~5 mins which is what most fights in the past have been balanced around (other than ezmode naxx plow). On short fights FFB pulls closer because of IV, and on longer fights it pulls closer because fireball spec has to evo. So fight length makes a noticable difference, however fireball always comes out ahead, that never changes. If I had to take a stab at it I'd say your fight length is shorter than 5 minutes right now, probably 3 mins if I had to guess.
I think you are still doing something incorrectly, check my values on the previous page (in the character sheet) against yours. Also you never specify what fight length you are using. The gap between fire and ffb is largest in the middle, ~5 mins which is what most fights in the past have been balanced around (other than ezmode naxx plow). On short fights FFB pulls closer because of IV, and on longer fights it pulls closer because fireball spec has to evo. So fight length makes a noticable difference, however fireball always comes out ahead, that never changes. If I had to take a stab at it I'd say your fight length is shorter than 5 minutes right now, probably 3 mins if I had to guess.
How good is the Rawr optimizer? Running the optimizer to gear the specs individually I got the following results:
I'm surprised that FFB pulled ahead and wondering if I did something wrong. These were run with identical buffs, and the optimizer was run seperately for each spec. Rawr version is 2.2.0b6 with 3.1 mode enabled.
Don't run the optimizer to gear the specs it is bad. Generally the more hit the spec needs the worse the optimizer does. We already have the two optimal gear sets that were figured out months ago on page one of the optimal gear sets thread. Also I'm assuming the optimizer is using new gear from 3.1 by the high values you have listed, and using 3.1 gear is kind of a bad idea at this point since none of it is final and we don't know the complete loot tables.
...what are you talking about, Duravi? The Optimizer does an awesome job of finding the best gear setup for a spec, and would only use 3.1 gear if you specifically marked 3.1 gear as available.
if you're comparing specs and *not* using the Optimizer to find the optimal gear for each spec, you're doing it wrong.
Don't run the optimizer to gear the specs it is bad. Generally the more hit the spec needs the worse the optimizer does. We already have the two optimal gear sets that were figured out months ago on page one of the optimal gear sets thread.
As Astrylian said, I'm really not sure how you think I arrived at the 'optimal' sets that I listed. I assure you, I've used the optimizer in every build of Rawr. It's an incredibly useful and powerful tool. I mean, if you want to hand sort through all the items and gems to arrive at the same numbers that it is getting, then I guess that's your choice, but if your argument is that the optimizer isn't properly showing you upgrades, then you'd be quite mistaken.
From what I've gathered, when you switch out an item, or enable a buff, or change anything, Rawr is computing your DPS based on your settings (and whether you have things like Sequence Reconstruction turned on to show you how it's getting those numbers). Optimizer is undoubtedly doing the same thing, but it's looking at unfathomably more combinations due to the number of variables possible, number of sockets, number of items selected per slot, and so forth.
While you can say, "This item has a lot of hit on it, and I don't need 40 hit, I only need 30 hit, so I'll switch out this item and switch in this other item," it's a limited perspective. You could look at Dying Curse and say it's great because hit rating is valued high, it has a ton of hit on it, so you say to yourself, if I can get a lot of hit from one item, I can free up two other slots, and yes, making that switch yourself could show you a higher DPS, but what the Optimizer is doing is saying, "One of those items you'll keep, the other you won't, then you use this gem, and replace Dying Curse with this other trinket, and it's an even better combination."
The only thing I can really think of concerning the optimizer tool is that in older builds (pre-2.1.6) it wasn't always working the greatest with gemmings, but the new gemming template is amazing and it works quite well now. I'm very impressed with how far Rawr has come, and of course, very appreciative (and vocal about it).
A quick detour to clear some possible misconceptions. First the mage solver and optimizer work in a very different way. The mage solver is deterministic, meaning each time you'll get the same result given the same input. The optimizer on the other hand is not and is not guaranteed to find optimum on any given run (though increasing thoroughness and running it several times improves the chances a lot).
Sequence Reconstruction is a post-process step. It plays no part in how the solver gets its numbers. In fact when you run the optimizer sequence reconstruction is never done. It is only used to display the cycle sequence and its corresponding graph. What it does is it converts a description from the solver (which is of the form in this time interval cycle 1 is used for x1 seconds with these cooldowns, cycle 2 is used for x2 seconds with other cooldowns and so on) and rearranges those cycle segments so that there are no gaps in cooldowns. This is most hard to do if you don't use segment cooldowns option since in that case there is only one very large time interval with all cycle-cooldowns inside it. Once again, this has no effect on the values that are displayed.
For most accurate evaluations I would suggest having both integral mana consumables, segment cooldowns and increasing advanced constraints as far as it's not too time consuming. Only do this for display values, not for comparisons because otherwise you'll be waiting years for optimizer to finish. The basic solver is setup so that comparisons should be close enough, but if you want the final answer once you have selected the gear to be as accurate as possible, then it's very important to use the advanced solver. In general the basic solver will give overestimates about 0.1-1% based on complexity of the fight parameters.
As Astrylian said, I'm really not sure how you think I arrived at the 'optimal' sets that I listed. I assure you, I've used the optimizer in every build of Rawr. It's an incredibly useful and powerful tool. I mean, if you want to hand sort through all the items and gems to arrive at the same numbers that it is getting, then I guess that's your choice, but if your argument is that the optimizer isn't properly showing you upgrades, then you'd be quite mistaken.
Maybe my settings for it are incorrect but I often get setups with it that end up not to be optimal. Remember that setup that I talked about a while ago that only used 2 PC t7? Thats what I ended up with using optimizer not your setup that is in fact better. I have never had success using optimizer to get the best possible setups, it comes close but never gets there from my experience. I havent used it since like rawr 2.14 - 2.16 so it could be greatly improved, but that was my experience with it in those versions. I'm going to give it a go in the newest beta and see what it comes up with, apologies if it has been upgraded since those older versions. Anyways the point of my original post was that we already have the best current setups atm for live, there is no reason to run the optimizer at this point.
Thanks for the information Kavan. It does make sense, and is probably why when you switch items around, they always result in the same DPS being listed. One thing with the Optimizer that I noticed early on was that some of the choices it came back with were, as you noted, not the 'best' per-say. I said this back when I originally made that "Optimal Sets" thread. I generally always run Optimizer under DPS Rating with Thoroughness at full. I don't know what has been done behind the scenes, but it certainly returns better results in 2.2 then it did in say, 1.6 or previous versions.
What I envisioned Sequence Reconstruction as was an "after-the-fact" type of deal, in that it was showing how the solver was reaching what it gets, and why it's optional. When I first started using Rawr, I never enabled things like Sequence Reconstruction, nor did I use Integral Mana Consumables, Segment Cooldown options when they were first implemented. I suppose what most users of Rawr want to see is just a DPS number when they enter their gear, right out of the box. And undoubtedly that's why I never bothered with the options at first.
To make matters more interesting, when I first started toying with those options, I was enabling Comparison as well as Display, which would always freeze Rawr indefinitely. And I haven't yet used Advanced Solver Log, nor have I used anything like Lower Bound Hint, Segment Duration, and so forth. I read your guide though, Kavan, on setting things in regards to SMP Computation Limit and Search Method. That was quite helpful.
Keep in mind Duravi that a lot of factors can influence why certain items are being valued over others. I think the main reason I even made a thread was that a lot of other threads were being cluttered with pages of discussion on whether 4-piece T7 was better than some other items (such as leggings of the wanton spellcaster), and discussion regarding just how much DPS the set bonuses were providing. My goal was to provide a convenient answer to this discussion, to help divert some of the traffic.
But as you'll notice, there are many issues that people share in common -- simple things like enabling 3.1 mode, or manually editing trinkets (back when Illustration was changed from 26 to 20, it was throwing a lot of results off), or turning off automatic armor switching. There was the case of someone having infinite mana turned on. All of these things combined probably make the optimizer come out with different results even though the "input" looks the same on the surface. But none of that detracts from it being a great tool when properly used.
Has there been a definitive test of the Living Bomb glyph vs. Improved Scorch glyph? I've been trying to dig through the thread after being away for a bit and my eyes have started to bleed >.<
We also have no raiding Warlocks atm (apparently they were crappy at the start of Wrath and re-rolled) so if the assumption that no warlock in the raid has any effect, note that that is in effect.
Has there been a definitive test of the Living Bomb glyph vs. Improved Scorch glyph? I've been trying to dig through the thread after being away for a bit and my eyes have started to bleed >.<
We also have no raiding Warlocks atm (apparently they were crappy at the start of Wrath and re-rolled) so if the assumption that no warlock in the raid has any effect, note that that is in effect.
Living Bomb glyph is more dps, and will be taken by most fire mages. Scorch glyph is probably still necessary if you're the only one with the 5% crit buff, but I haven't seen any numbers comparing that to spamming Scorch 5x to put up the stacks.
Living Bomb glyph is more dps, and will be taken by most fire mages. Scorch glyph is probably still necessary if you're the only one with the 5% crit buff, but I haven't seen any numbers comparing that to spamming Scorch 5x to put up the stacks.
Rawr is showing glyph of living bomb to be at 353 dps at the moment, which surpasses the fireball and molten armor glyphs by over 150 dps (in a 3 minute fight, though similar numbers are at 5 minutes as well); so it would be safe to say that if the living bomb glyph isn't changed, it will be the last glyph to replace. I'm not sure how rawr models scorch exactly, but if you set the fight length to 20 seconds and check the box for "maintain scorch" it only drops dps by 200, so it's safe to assume it either doesn't model the 5 scorches in the beginning, or doesn't scorch if it's that low a fight duration.
Rawr doesn't model ramping up of Scorch, it only models refreshes. The scorch glyph model used is very simplistic. If you have max points in improved scorch then the only difference it makes is that without glyph it assumes you cast scorch every 30 - scorch cast time seconds, and with glyph you cast it every 30 seconds. The reasoning behind it is that wihout glyph you want to be extra cautious not to miss the window and have to rescorch while with glyph you can afford to miss it by a bit.
Could you direct me to these guides you refer to for rawr. I myself have tried using the segmented cooldowns etc, and where their are two options (comparison/display) i never know which one i need to select, and it results in a program freeze, same problem with exceeding computation limits. Is their a 'rawr for dummies' guide anywhere. I think it could be an incredibly useful tool for my guild, if only i knew how to use it properly. First thing i will do is check the rawr website, but any other reading material you know of would be much apreciated.
Rawr doesn't model ramping up of Scorch, it only models refreshes. The scorch glyph model used is very simplistic. If you have max points in improved scorch then the only difference it makes is that without glyph it assumes you cast scorch every 30 - scorch cast time seconds, and with glyph you cast it every 30 seconds. The reasoning behind it is that wihout glyph you want to be extra cautious not to miss the window and have to rescorch while with glyph you can afford to miss it by a bit.
Perhaps there should be an option that allows you to specify at what time you're rescorching, because lets be fair, no one is rescorching at 30-cast time; that doesn't leave any time for latency, random pushbacks or whatever else.
Would also be very usefull to see the ramping up effect since it seems if things stay as they are now fire mages would be running without a scorch glyph even when they are the only ones applying it, ie. I'd prefer to scorch 5 times to ramp up rather then loose any of the DPS glyphs. (The best way would be to specify how many times you're scorching in the begining ie. how many mages are there etc.)
Perhaps there should be an option that allows you to specify at what time you're rescorching, because lets be fair, no one is rescorching at 30-cast time; that doesn't leave any time for latency, random pushbacks or whatever else.
Would also be very usefull to see the ramping up effect since it seems if things stay as they are now fire mages would be running without a scorch glyph even when they are the only ones applying it, ie. I'd prefer to scorch 5 times to ramp up rather then loose any of the DPS glyphs. (The best way would be to specify how many times you're scorching in the begining ie. how many mages are there etc.)
I think you're overestimating what useful information such a model can provide for you. I think a more accurate way to think about your dps is going to be the amount of time you can spend doing your max dps rotation (not including scorch) vs the time you have to spend ramping up/maintaining the crit debuff, ie 7.5 secs at the beginning followed by 1.5 secs every 30 seconds
The glyph's effect (or lack thereof) is dictated by the mechanics of the encounter and whether or not you play imperfectly. In any fight where there is target switching involved, or you're just bad at keeping up scorches/scorch too early/have an RNG resist, unglyphed rescorching is going lower your dps a lot, varying with the length of the encounter and the amount of target switching involved.
Point is, I think it'd be better to assume a 5% crit debuff and to look at your modelled dps rotation sans scorch - your goal in a raid is to maximize the time your spend doing that max dps rotation.
...unglyphed rescorching is going lower your dps a lot, varying with the length of the encounter and the amount of target switching involved.
Point is, I think it'd be better to assume a 5% crit debuff and to look at your modelled dps rotation sans scorch - your goal in a raid is to maximize the time your spend doing that max dps rotation.
That is exactly my point, saying "lower your dps a lot" does not mean anything, rawr is an analysis tool I use it to model fights and see what in ganeral would do better. What is the breakpoint of 'imperfect' play when you're better of using glyphed scorch.
And yes your job is to maximize the amount of time spent on your max dps rotation, however fights are not all patchwerk, there are fires to run from, whelps to kill or wait for boss phase changes. What I'm asking is for a way to put a number on the 'lower your dps a lot' involved in scorching/rescorching vs. the pure dps glyphs.
If I leave "3.1 mode" unchecked in the latest Rawr 2.2.0.6 beta, will the results reflect Live? Or are there some changes in there even with 3.1 mode disabled that would skew the results when compared to live servers?
If I leave "3.1 mode" unchecked in the latest Rawr 2.2.0.6 beta, will the results reflect Live? Or are there some changes in there even with 3.1 mode disabled that would skew the results when compared to live servers?
If you don't have 3.1 mode enabled then everything should be as live except scorch will be 5% crit.
2.) Now that there will be no more swords without spirit (at least i hope so), there will be no freaking paladins running around with our casterweapons anymore... Seriously i think NO mage in ANY guild with dkp system (which sucks, but lets not talk about that) got their kel-sword before ALL healpaladins in the raid got it.. why? Cause Palas only need to get plateitems and are therefore saving massive amounts of dkp compared to a mage who has to deal with several other classes on every item....
That makes no sense.
If you had more people to compete with you would not be able to spend dkp as people would most likely have more dkp or bid more dkp and to top it off paladins who are not sharing with any other classes and can take more loot therefore being forced into spending more dkp for items that aren't anywhere near an upgrade compared to a weapon.
But then again, who actually uses dkp systems in entry level content, especially considering how trivial the game is right now.
When 3.1 was introduced for testing. I heard Both sides of the story, about who was going to maintain the crit debuff. Now that I am back around after a few weeks away. What has been decided would be the main debuff class when 3.1 hits. I am assuming mages still but was hoping not to free up the glyph spot.
When 3.1 was introduced for testing. I heard Both sides of the story, about who was going to maintain the crit debuff. Now that I am back around after a few weeks away. What has been decided would be the main debuff class when 3.1 hits. I am assuming mages still but was hoping not to free up the glyph spot.
As it stands, imp scorch will most likely always be in the raid, as it's now a mandatory talent for fire and ffb specs. Of course if you have a lock in raid who has it, there's no need for mages to maintain it.
If you followed the last couple of pages, you'd see that dropping the imp. scorch glyph for LB glyph, is generally thought of as a good idea, whether you're maintaining scorch or not.
04/05/09, 1:56 PM
Similar Threads
