I started testing with the purpose of either proving you wrong or figuring out where I went wrong and instead I found something really interesting and a bit strange.
Try this out.
Hop on the PTR and remove all your talents.
Go over to the Targeting Dummy with your slow 2-hander.
Start autoattack.
Cast a Lava Burst right after a swing so that it should be able to fit inside the swingtimer.
Notice that your swingtimer reset.
Spec talents so that you have at least 1 point in MW.
Try the same thing again, using Lava Burst and without any charges of MW.
Notice that the Swingtimer doesn't reset if the cast fits inside the swing time.
Doesn't matter if you have charges of MW or even what spell you use.
It's not the charges that determine if the swingtimer resets it's whether you specced the talent or not.
This is interesting. I showed in an earlier post that lightning bolt needs to have at least one stack of Maelstrom Weapon to not reset the swing timer. Because you're reporting different behavior for Lava Burst on PTR, I tested this out on live
1.81 sec cast Lava Bolt (using WoA to speed up the cast so it's less likely to clip)
3.52 sec swing time
No Maelstrom, timed to not clip:
12/18 09:41:44.492 SPELL_AURA_REMOVED,"Maelstrom Weapon",0x1,BUFF
12/18 09:41:45.501 SWING_DAMAGE
12/18 09:41:46.425 SPELL_CAST_START3,"Lava Burst"
12/18 09:41:48.243 SPELL_DAMAGE,"Lava Burst"
12/18 09:41:51.718 SWING_DAMAGE
This matches the swing timer reseting after the spell cast completes
cast_start + cast_time + swing_time = 41:46.43 + 1.81 + 3.52 = 41:51.76
1x Maelstrom, timed to not clip:
12/18 09:57:58.464 SWING_DAMAGE
12/18 09:57:59.168 SPELL_AURA_APPLIED,"Maelstrom Weapon",0x1,BUFF
12/18 09:57:59.625 SPELL_CAST_START,"Lava Burst",0x4
12/18 09:58:01.465 SPELL_DAMAGE,"Lava Burst",0x4,2037,2036,4,0,0,0,nil,nil,nil
12/18 09:58:02.253 SWING_DAMAGE
This matches the swing timer free running during the cast (difference is within my latency reported by quartz)
first_swing + swing_time = 57:58.46 + 3.52 = 58:01.98
1x Maelstrom, timed to clip:
12/18 10:05:52.958 SWING_DAMAGE
12/18 10:05:53.211 SPELL_AURA_APPLIED,"Maelstrom Weapon",0x1,BUFF
12/18 10:05:55.931 SPELL_CAST_START,"Lava Burst"
12/18 10:05:57.748 SPELL_DAMAGE,"Lava Burst"
12/18 10:05:59.766 SWING_DAMAGE
This matches the swing timer free running, and the swing getting clipped.
first_swing + 2 * swing_time = 05:52.96 + 2 * 3.52 = 06:00.00
Either PTR is behaving differently or perhaps you had MW proc when you were testing it out. Either that or one of our testing methods is invalid. I guess the next thing for me to do is test this in PTR.
edit:
By the way. This brings up a disturbing possibility for timing spell casts with the swing timer for maximum dps.
This is interesting. I showed in an earlier post that lightning bolt needs to have at least one stack of Maelstrom Weapon to not reset the swing timer. Because you're reporting different behavior for Lava Bolt on Beta, I tested this out on live
Either Beta is behaving differently or perhaps you had MW proc when you were testing it out. Either that or one of our testing methods is invalid. I guess the next thing for me to do is test this in beta.
Test it on the PTR, please, and see if it matches your live results.
Either way it is interesting that a spell that is unaffected by Maelstrom Weapon (Lava Burst) still doesn't reset the swingtimer. Could have some interesting potential for 2-hand PvP, although after last night's attempt at doing the Arenas I really don't care anymore if it does or doesn't since I don't think I will be spending much time PvPing anymore anyway.
Actually, it's even worse using the same 100% crit model as before.
I get 6181.5 dps for that weapon's dps and speed. Of course, it will be less than this because it lacks the dps stats as well.
I think the bottom line on that one is that it's bad. Just completely bad. The amount you gain in flametongue does not make up for less from stormstrike and lash.
edit: ran thru a comparison of the weapons using my stats, 100% crit rate and 0% miss/dodge/glancing, to assure full flurry uptime, and eliminate modeling "noise". Again, using the other stats of the stray.
This type of test seems kind of stupid and inaccurate for other reasons, but I can't really articulate why or anything.
Have you Sim'd Webbed Death at all out of curiosity? It actually sims higher for me then any other weapon due to it's 1.4 speed. I am over the spell hit cap, so Sinister would probably sim slightly higher if I wasn't, but the numbers I came out with with my current gear showed Webbed Death 6 dps higher. Given equal stats/dps, the 1.4 sims for me about 60 dps higher then a 2.6 would.
I have been messing around with he sim for quite some time now, but I can never seem to get EP values that make any sense. While I understand that do to the nature of the sim I will never get the exact same results, I am still confused with some of the negative results that I have gotten for the various ratings. Here is my config file. I would really appreciate if someone could point out any mistakes that I made in setting up the sim.l
Test it on the PTR, please, and see if it matches your live results.
Either way it is interesting that a spell that is unaffected by Maelstrom Weapon (Lava Burst) still doesn't reset the swingtimer. Could have some interesting potential for 2-hand PvP, although after last night's attempt at doing the Arenas I really don't care anymore if it does or doesn't since I don't think I will be spending much time PvPing anymore anyway.
PTR matches what I see live.
3.67 swing time
1.93 cast time
I have 5/5 MW
No Maelstrom weapon proc, cast timed to not clip.
12/18 22:31:35.859 SWING_DAMAGE
12/18 22:31:36.544 SPELL_CAST_START,"Lava Burst"
12/18 22:31:38.676 SPELL_DAMAGE,"Lava Burst"
12/18 22:31:42.210 SWING_DAMAGE
The timestamps match a reset of swing timer with no MW stacks
36.54 + 1.93 + 3.67 = 42.10
Have you Sim'd Webbed Death at all out of curiosity? It actually sims higher for me then any other weapon due to it's 1.4 speed. I am over the spell hit cap, so Sinister would probably sim slightly higher if I wasn't, but the numbers I came out with with my current gear showed Webbed Death 6 dps higher. Given equal stats/dps, the 1.4 sims for me about 60 dps higher then a 2.6 would.
I had some similar thoughts with[Last Laugh] a little while back in this thread, I never bothered to actually sim it since it supposedly came out to lower dps. I'll do it when I get home, though. Just in my head I'd think a 1.6 @ 171.6dps would come out higher than a 2.5 @ 156.6dps -- very likely not higher than the Sinister Revenge due to noticeable stat loss though.
I had some similar thoughts with[Last Laugh] a little while back in this thread, I never bothered to actually sim it since it supposedly came out to lower dps. I'll do it when I get home, though. Just in my head I'd think a 1.6 @ 171.6dps would come out higher than a 2.5 @ 156.6dps -- very likely not higher than the Sinister Revenge due to noticeable stat loss though.
My OH DPS EP is generally in the low 2s, 2.15 currently. It doesn't take much in terms of OH stats to make up for ~30 EP going from 156.6 to 171.6.
[Dezzimal] The first iteration of Algalon looted your raid after a kill and posted screenshots on his website.
Tukez - on the directions for the mac compiling of the source, please add that when creating the new project in Xcode, the project type should be Command Line Utility -> C++ Tool
Every time I stop, and restart combat my weapons start out of phase. It seems like enhsim should do the same, or else it will overvalue matched weapon speed configurations.
I thought I would come back to this comment because I've learned how to start fights with my weapons in phase.
With matched main hand and off hand weapon speeds, if I enable my swing a couple seconds before I'm in range of the target they start in phase, If I start the swing when i'm in melee range of the target they start out of phase.
If I start with the weapons out of phase, steping back out of melee range for a couple seconds will line up the weapons back into phase.
Because of this, I've taken to using a macro as I approach the target
I wanted to bring back up the topic of Angry Dread vs. Sinister Revenge again mostly because of whether the mechanics for Berserking are accurate in the sim or not. If they mechanics are accurate, please ignore this post, but otherwise, Angry Dread should have a higher Berserking uptime on OH, correct?
Elemental Mastery - Lightning Bolt - Die - Reincarnation - Lightning Bolt - Lightning Overload Procs - Die
I wanted to bring back up the topic of Angry Dread vs. Sinister Revenge again mostly because of whether the mechanics for Berserking are accurate in the sim or not. If they mechanics are accurate, please ignore this post, but otherwise, Angry Dread should have a higher Berserking uptime on OH, correct?
There is another factor besides weapon speed in determining zerker uptime and it would play a role here. Haste. Sinister Revenge has 24 haste rating and Angry Dread has none so that will also need to be factored in to determine which weapon would average higher uptimes of the weapon enchants.
One thing I noted using the GUI today, I was playing around with what effects different meta gems would have and noted that on changing the meta gem I was getting almost identical figures - ie: within 1 dps each run. It then struck me that all the sim was doing was accounting for the "special" on the meta gem (eg: 3% crit damage bonus) and I was expected to change the crit rating myself. This is fair enough but not intuitive.
My issue was that for the entry Chaotic Skyflare Diamond I needed to adjust the crit rating however crit is entered as a percentage but of course all I had to go on was the rating so I had to look up Whitetooth's Combat Ratings at 80, convert the rating to a percentage and then manually adjust the rating percentage. Adjusting for Relentless Earthsiege Diamond was even worse as adding 21 Agi to the GUI in the sim makes ZERO difference to the dps, I thought I was seeing things so I added 100 and then 1000 Agi and got the same results (note I didn't have Blessing of Kings ticked). So I surmised that I had to manually adjust all the things that agi affects before I could re-run the sim, a huge amount of manual effort when the GUI could perhaps cope with that.
I fully understand why this is happening as the GUI is simply creating the text file and running the sim, I can cope with doing this as I know where to get the info, however for the average user this is way beyond what they would ever expect to be able to cope with. Perhaps a warning when changing meta gems that no change in base stat (eg: agility, crit rating etc) is taken into account, or better still a tick box (stats included or not included) and if not ticked then the sim could add the 21 Agi, 21 Crit rating etc. This way your average user could use the sim and not have to manually add a lot of changes at the same time they are experimenting with changing an item in a drop down list.
The same philosophy could be used with trinkets etc that grant fixed bonuses + specials ie: tick box to ask if the user has already included the base figures in the paper doll stats column.
NB. This change would be a huge benefit to the extreme min/maxer too as it saves a massive amount of time simply ticking a box rather than having to manually adjust lots of figures every time they change around a gem/trinket/etc before rerunning the sim. A typical intense user would want to run the sim dozens of times with dozens of different settings making small tweaks here and there. To have to do manual calculations at every stage is a huge extra time sync and makes small changes prone to user error, with forgetting to revert a change or typo having a potentially major impact.
How to intense users currently cope with this? Is there some easy method I'm overlooking?
Last edited by Levva : 12/23/08 at 8:00 AM.
Author of ShockAndAweEnhancement Shaman max dps addon
Please use the EnhSim by Ziff & others to simulate what gear, priorities etc are the best dps. You can use ShockAndAwe to export your paperdoll stats to EnhSim.
How to intense users currently cope with this? Is there some easy method I'm overlooking?
I don't have an easy method. I'm doing all the error prone work of hand calculating %'s. Perhaps there is a character builder website that would give you the paper doll values for different gear/gem setups?
What would be really nice is a section that allows you to add misc stats that the sim or gui could add in automatically. The misc stats fields would allow you to add stats as they show up on item tooltips.
For example. If I want to try out different main hand weapons, I could take off my main hand, and enter in the paper doll stats I see in the old paper doll stats fields. Then enter the stats of the weapon I want to try in the misc stats fields and run the sim. For each weapon I want to try I just change the misc stats fields, and leave the paper doll alone. This would save a lot of time for min/maxers, and would also work (with a little bit of effort) for the meta gem problem described by Levva.
I don't have an easy method. I'm doing all the error prone work of hand calculating %'s. Perhaps there is a character builder website that would give you the paper doll values for different gear/gem setups?
What would be really nice is a section that allows you to add misc stats that the sim or gui could add in automatically. The misc stats fields would allow you to add stats as they show up on item tooltips.
For example. If I want to try out different main hand weapons, I could take off my main hand, and enter in the paper doll stats I see in the old paper doll stats fields. Then enter the stats of the weapon I want to try in the misc stats fields and run the sim. For each weapon I want to try I just change the misc stats fields, and leave the paper doll alone. This would save a lot of time for min/maxers, and would also work (with a little bit of effort) for the meta gem problem described by Levva.
This was why I wish there was some way to link Rawr with the Sim so that the Rawr could be used to gather the paper doll stats to output to the Sim to determine the EP values that would be fed back into the Rawr to rank the individuals items and show the dps determined from the Sim run. Then you could change items based on those rankings and re-output it back into the Sim for another determination of overall dps or to even run another reranking of items based on the current paper doll stats showed on the Rawr interface.
The simulator is giving me an EP value of 3.36 for the intellect (or around this value in the various tests i did) and i don't know if this value is exact or not. What i did wrong?
This is my configuration file:
Real simple, you are running OOM so mana pool size is the biggest determinant of how much dps you will do.
I know that isn't the case but that is the data you are feeding into the Sim.
You have it doing 6 minute combat cycles with SR set to the lowest priority which means it probably isn't actually being used at all since something else will be using your GCD and also since it can't even use SR in the first place unless you are under 700 mana which is something you set in the priority tab.
Change the combat cycles to 3 minutes since there are few if any heroic battles that last longer then that anyway.
Then change the priority list so that SR is first priority, which is fine since it won't be used unless you drop below 700 mana anyway.
or you can just uncheck the simulate mana option and go from there.
This was why I wish there was some way to link Rawr with the Sim so that the Rawr could be used to gather the paper doll stats to output to the Sim to determine the EP values that would be fed back into the Rawr to rank the individuals items and show the dps determined from the Sim run. Then you could change items based on those rankings and re-output it back into the Sim for another determination of overall dps or to even run another reranking of items based on the current paper doll stats showed on the Rawr interface.
Currently Rawr gives a reasonable enough data export when you save your character and it saves it in XML format. I am away for Xmas at parents house for four days with only a dial up internet connection so I'll be climbing the walls without anything useful to do. You have just however saved my sanity by suggesting a useful project.
I shall endeavour to write a Java (ie: platform independent) app that will take the output of a Rawr xml file and re-write it as the same format that ShockAndAwe exports which of course EnhSim already supports. This would then mean you run the app point it at your XML file and it would give you the stuff to paste into your EnhSim config file.
In four days, assuming I'm not being offered more food every 30 mins by my mother keen not to see me starve, I'd hope to have at least a version able to be tested.
BTW. Is the Rawr 2.1.4 Enhance module broken at present or even updated for 3.0.3 as I can find no way to change the Boss level from the fixed value of 70.
Edit: Hmm bugger. The Rawr save only saves item ids not any actual stats. It does however save talents. Conversely the "Copy to Clipboard" option does seem to give some useful values - very very messy. This would mean any program would need to read two different files one as a cut n paste one as a saved file. Not at all intuitive, it would be so much nicer if the people at Rawr could be persuaded to include the paper doll stats in the XML save file.
I'll have a bash at it as I need something to do but I'm not that hopeful now.
I've posted a Issue on the Rawr Issue Tracker - perhaps if people here who like the idea of enabling data swapping between EnhSim & Rawr could vote on the issue on the tracker it would bump it up the queue and perhaps get implemented soon. The issue is logged at Rawr - View Issue
Last edited by Levva : 12/24/08 at 9:46 AM.
Author of ShockAndAweEnhancement Shaman max dps addon
Please use the EnhSim by Ziff & others to simulate what gear, priorities etc are the best dps. You can use ShockAndAwe to export your paperdoll stats to EnhSim.
I was playing around with the ele part of the sim and I discovered mana problems with my shit gear and looking at the output it displayed 0 mana gained from JoW. I found that strange because JoW gives mana on spells so this seems to be a flaw in the mana calculations, WTB fix.