You might want to fix the link in your post. The URL displayed is correct but clicking on the link will load a different URL which will download Rev B instead of 1.7.0.2 (fixed in the quoted text). I downloaded the old version by mistake and I imagine I am not the only one.
Thanks for all your hard work -- it is much appreciated.
Whoops thanks for the catch, link fixed in the post.
And from that are achieving the following EP values:
EP value DPS total DPS difference
baseline 6862.06
ap 1.00 0.65 6992.28 130.22
crit rating 1.74 1.14 6896.11 34.05
hit rating 1.63 1.06 6893.89 31.83
expertise rating 0.00 0.00 0.00 -6862.06
haste rating 1.33 0.86 6888.00 25.95
armor penetration rating 1.51 0.98 6891.47 29.41
spellpower 0.96 0.62 6955.80 93.74
mh dps 5.11 3.33 6887.02 24.96
oh dps 2.79 1.81 6875.66 13.61
strength 1.10
agility 1.69
intelligence* 1.39
spirit 0.00 0.00 0.00 -6862.06
mana 0.00 0.00 0.00 -6862.06
mp5 0.00 0.00 0.00 -6862.06
* part of the EP value is missing because of skipped tests.
( Pawn: v1: "EnhSim": Ap=1.00, CritRating=1.74, HitRating=1.63, HasteRating=1.33, ExpertiseRating=0.00, ArmorPenetration=1.51, SpellDamage=0.96, MainHandDps=5.11, OffHandDps=2.79, Strength=1.10, Agility=1.69, Intellect=1.39, Spirit=0.00, Mana=0.00, Mp5=0.00 )
http://www.lootrank.com/wow/wr.asp?Cla=64&Str=1.10&mh=1.33&dps=5.11&mp5=0.00&mcr=1.74&odps=2.79&spd=0.96&Agi=1.69&mhit=1.63&Int=1.39&Spi=0.00&map=1.00&arp=1.51&Exp=0.00
elapsed simulation time: 10000.00h
elapsed real time: 49.42s
simulation speed: 728435x
And achieving the following dps:
EnhSim version 1.7.0.2 by Verges: scionweb@gmail.com
................................................................................
MH:
miss 242983 9.57%
dodge 0 0.00%
glancing 609388 24.00%
hit 466990 18.39%
crit 1220035 48.04%
MH Windfury:
procs/hits 15.48%
procs/swings 14.19%
miss 0 0.00%
dodge 0 0.00%
hit 407157 49.02%
crit 423479 50.98%
MH Storsmtrike:
miss 0 0.00%
dodge 0 0.00%
hit 190113 49.16%
crit 196600 50.84%
OH:
miss 243897 9.60%
dodge 0 0.00%
glancing 610396 24.04%
hit 464712 18.30%
crit 1220391 48.06%
OH Flametongue:
miss 0 0.00%
hit 1887846 59.94%
crit 1261777 40.06%
OH Stormstrike:
miss 0 0.00%
dodge 0 0.00%
hit 189447 48.99%
crit 197266 51.01%
Earth Shock:
miss 0 0.00%
hit 306932 60.08%
crit 203912 39.92%
Lava Lash:
miss 0 0.00%
dodge 0 0.00%
hit 228191 48.82%
crit 239220 51.18%
Lightning Bolt:
miss 0 0.00%
hit 364795 60.00%
crit 243217 40.00%
Lightning Shield:
miss 0 0.00%
hit 438739 100.00%
crit 0 0.00%
Spirit Wolves:
miss 0 0.00%
dodge 117156 6.51%
glancing 431638 23.99%
hit 1041948 57.92%
crit 208136 11.57%
Maelstrom Weapon:
PPM 57.78
flurry uptime 95.19%
unleashed rage uptime 0.00%
elemental devastation uptime 78.17%
elemental oath uptime 0.00%
mh enchant uptime 41.48%
oh enchant uptime 37.34%
trinket1 uptime 18.08%
trinket2 uptime 20.61%
totem uptime 64.45%
set bonus uptime 0.00%
metagem uptime 0.00%
gloves enchant uptime 0.00%
cloak enchant uptime 0.00%
DPS PPM MPS
white 2127.44 31.01% 84.65
windfury 842.02 12.27% 6.92
flametongue 549.45 8.01% 52.49
stormstrike 422.09 6.15% 6.45 0.00 0.00%
lava lash 348.81 5.08% 7.79 21.46 10.54%
magma totem 503.14 7.33% 2.66 49.55 24.34%
earth shock 581.53 8.48% 8.51 55.02 27.03%
lightning bolt 977.86 14.25% 10.13 74.25 36.47%
lightning shield 218.99 3.19% 0.81 0.00 0.00%
spirit wolves 289.99 4.23% 0.40 3.31 1.62%
DPS 6861.32
MPS 203.58
MP2min 24430.01
MRPS 0.00
Out of mana time 0.00%
elapsed simulation time: 1000.00h
elapsed real time: 4.98s
simulation speed: 722311x
I am particularly interested in my haste value, which seems abnormally low. My intuition tells me that the actual value of haste is much higher. When I manipulate my config favoring haste over crit (applying haste enchant and swapping 2x ap/crit gems for ap/haste gems = + 1.55% haste and -0.34% crit/24.2 AP) I experience a dps increase- this is despite the EP value of haste being significantly lower than crit.
EnhSim version 1.7.0.2 by Verges: scionweb@gmail.com
................................................................................
MH:
miss 246769 9.59%
dodge 0 0.00%
glancing 617354 24.00%
hit 480686 18.68%
crit 1227942 47.73%
MH Windfury:
procs/hits 15.29%
procs/swings 14.01%
miss 0 0.00%
dodge 0 0.00%
hit 409001 49.32%
crit 420305 50.68%
MH Storsmtrike:
miss 0 0.00%
dodge 0 0.00%
hit 190404 49.27%
crit 196040 50.73%
OH:
miss 246170 9.57%
dodge 0 0.00%
glancing 618046 24.02%
hit 480001 18.66%
crit 1228534 47.75%
OH Flametongue:
miss 0 0.00%
hit 1912178 60.13%
crit 1268101 39.87%
OH Stormstrike:
miss 0 0.00%
dodge 0 0.00%
hit 190634 49.33%
crit 195810 50.67%
Earth Shock:
miss 0 0.00%
hit 307476 60.23%
crit 203020 39.77%
Lava Lash:
miss 0 0.00%
dodge 0 0.00%
hit 229925 49.21%
crit 237329 50.79%
Lightning Bolt:
miss 0 0.00%
hit 368112 60.12%
crit 244203 39.88%
Lightning Shield:
miss 0 0.00%
hit 441784 100.00%
crit 0 0.00%
Spirit Wolves:
miss 0 0.00%
dodge 116738 6.49%
glancing 432992 24.08%
hit 1040844 57.89%
crit 207494 11.54%
Maelstrom Weapon:
PPM 58.24
flurry uptime 94.98%
unleashed rage uptime 0.00%
elemental devastation uptime 78.18%
elemental oath uptime 0.00%
mh enchant uptime 41.62%
oh enchant uptime 37.52%
trinket1 uptime 18.14%
trinket2 uptime 20.68%
totem uptime 64.41%
set bonus uptime 0.00%
metagem uptime 0.00%
gloves enchant uptime 0.00%
cloak enchant uptime 0.00%
DPS PPM MPS
white 2143.96 31.17% 85.76
windfury 836.51 12.16% 6.91
flametongue 553.92 8.05% 53.00
stormstrike 420.27 6.11% 6.44 0.00 0.00%
lava lash 346.88 5.04% 7.79 21.45 10.52%
magma totem 502.01 7.3% 2.66 49.51 24.27%
earth shock 580.42 8.44% 8.51 54.98 26.95%
lightning bolt 983.65 14.3% 10.21 74.77 36.65%
lightning shield 220.40 3.2% 0.82 0.00 0.00%
spirit wolves 289.21 4.21% 0.40 3.31 1.62%
DPS 6877.22
MPS 204.02
MP2min 24482.53
MRPS 0.00
Out of mana time 0.00%
elapsed simulation time: 1000.00h
elapsed real time: 5.01s
simulation speed: 717846x
I am aware that haste peaks and troughs in relative value, but have been unable to find specific data supporting this. Is someone able to direct me to it?
More specifically:
1) Is my config incorrect?
2) Why am I getting more dps out of haste than crit when the EP value of haste is lower than crit?
3) Have other shaman at my gear level done extensive haste EP value testing? What values are your arriving at? A brief look at the top 20 shaman for XT (wowmeteronline) indicates a focus on crit over haste- what are the rest of you seeing?
Please don't use the same numeric version style on non-official releases of the sim. Tag it with something that will make it clear it is a fork version(even if it's meant to be a short-lived fork). Like: EhnSim 1.7.0-Verges1, 1.7.0-Verges2 etc. Explains the source of the work, author of the fork and version of the fork.
Baby, you can hold my balls.
10:10 < buu_> Raut: You are a hero of the internet.
The reason that your haste EP rating has fallen dramatically below hit/crit can be attributed too simply having 'too much' haste!
With 13.76% unbuffed haste, 20% WF, 25% flurry and 3% raid haste, your 2.6 weapons are attacking at 1.48 seconds each - just below the 1.5 speed 'threshold'. (Note, I am not saying you are doing less damage due to the 'wf pit' - you are still doing more overall damage due to more white attacks). When your hasted attack speeds fall between 1.4 and 1.49 seconds, the EP value of haste plummets.
If you are bothered by this, you can drop a few of your haste gems in favor of crit or hit. If you get exactly 12% haste unbuffed, that would put your hasted attack speed at 1.503 seconds. That's 302 haste rating in case you were wondering. Unfortunately you would have to swap out a piece of gear that has haste for one that doesn't - 23 haste and 2 8 haste gems isn't enough to get you back down to 302.
As a point of interest, if you are lucky enough to have the [Golden Saronite Dragon] or [The Masticator], you can afford to have up to 16.5% haste unbuffed (416 haste rating) and still stay above a 1.5 second attack speed threshold.
There is no true haste cap. As your weapon speed drops closer to a value that divides the 3s Windfury internal cooldown, the EP of an additional point in haste increases; the next potential proc after the cooldown is sooner. A 1.51s swing speed has a potential proc 0.02s after the cooldown, a 1.49s swing speed has a potential proc 1.47s after the cooldown. But as you add even more haste, the time before a potential proc falls again. (A 1.01s swing speed has a potential proc 0.03s after the cooldown.)
All that Windfury stuff aside, additional haste will always increase white weapon damage, Flametongue Weapon damage, Lightning Shield procs, and Maelstrom Weapon stacks.
Several people have plotted a graph of haste EP vs. current haste value. This ends up being a jagged graph. The general consensus at the time was to take a broader perspective when using EP values from enhsim, particularly for haste, hit, and expertise. For haste, use an EP value similar to that for agi or crit. Don't try to reduce haste on your gear to stay above the windfury threshold, but rather recognize that your haste EP will continue to increase past the threshold as your swing speed gets even faster.
Even reducing armor pen down to 0 there seems to be a round a 30 dps discrepency between the two simulators. Was anything other than the armor pen calculations changed for this new version?
There is no true haste cap. As your weapon speed drops closer to a value that divides the 3s Windfury internal cooldown, the EP of an additional point in haste increases; the next potential proc after the cooldown is sooner. A 1.51s swing speed has a potential proc 0.02s after the cooldown, a 1.49s swing speed has a potential proc 1.47s after the cooldown. But as you add even more haste, the time before a potential proc falls again. (A 1.01s swing speed has a potential proc 0.03s after the cooldown.)
All that Windfury stuff aside, additional haste will always increase white weapon damage, Flametongue Weapon damage, Lightning Shield procs, and Maelstrom Weapon stacks.
Several people have plotted a graph of haste EP vs. current haste value. This ends up being a jagged graph. The general consensus at the time was to take a broader perspective when using EP values from enhsim, particularly for haste, hit, and expertise. For haste, use an EP value similar to that for agi or crit. Don't try to reduce haste on your gear to stay above the windfury threshold, but rather recognize that your haste EP will continue to increase past the threshold as your swing speed gets even faster.
I did not say there was a haste cap. I simply said that with a 2.6 speed weapon, having more than 12% haste will push your attacks below 1.5 seconds between swings - resulting in higher overall damage but a dramatically lower EP value for additional haste. Additionally, I told him what to do if it bothers him.
However, I disagree with your statement that you should always use the same EP value for haste as we do with agi/crit/hit. Once you are below 1.5 and above 1.4 second speeds, your EP value for haste is much lower, meaning that all things being equal, your points are better spent upgrading your other statistics as they will provide the biggest 'bang' for the buck and a higher overall DPS. That is, after all, exactly what the simulator is telling you. Unlike hit, which maintains a fairly high rating beyond even the spell hit cap, haste is not always the best choice. It simply has multiple 'soft caps'.
I did not say there was a haste cap. I simply said that with a 2.6 speed weapon, having more than 12% haste will push your attacks below 1.5 seconds between swings - resulting in higher overall damage but a dramatically lower EP value for additional haste. Additionally, I told him what to do if it bothers him.
However, I disagree with your statement that you should always use the same EP value for haste as we do with agi/crit/hit. Once you are below 1.5 and above 1.4 second speeds, your EP value for haste is much lower, meaning that all things being equal, your points are better spent upgrading your other statistics as they will provide the biggest 'bang' for the buck and a higher overall DPS. That is, after all, exactly what the simulator is telling you. Unlike hit, which maintains a fairly high rating beyond even the spell hit cap, haste is not always the best choice. It simply has multiple 'soft caps'.
I'm finding completely different results. I'm currently running with 16.33% melee haste and the sim often comes up with EP values for haste well above 2, meanwhile my hit is only 1% over the spell hit and the EP value is around 1.5. Haste is a very touchy stat though. At one point in time, running the EP calculator with an ep_haste value of 30 gave me a negative value for haste, but both 29 and 31 gave me values around 1.4, and the further I deviated from 30, the higher the value climbed. Unfortunately, with the 3 second cd windfury debunked, we have even less of an idea how haste effects it.
With hit, the lower my hit is (without going below the spell hit cap of course), the lower the EP value I get for it.
I did not say there was a haste cap. I simply said that with a 2.6 speed weapon, having more than 12% haste will push your attacks below 1.5 seconds between swings - resulting in higher overall damage but a dramatically lower EP value for additional haste. Additionally, I told him what to do if it bothers him.
However, I disagree with your statement that you should always use the same EP value for haste as we do with agi/crit/hit. Once you are below 1.5 and above 1.4 second speeds, your EP value for haste is much lower, meaning that all things being equal, your points are better spent upgrading your other statistics as they will provide the biggest 'bang' for the buck and a higher overall DPS. That is, after all, exactly what the simulator is telling you. Unlike hit, which maintains a fairly high rating beyond even the spell hit cap, haste is not always the best choice. It simply has multiple 'soft caps'.
Flurry is 25% not 30%. So you can actually get 17.3% haste before you're under 1.5
For the 100th time in this thread: the 1.5 speed dip is an artifact of the simulator. It has next to zero impact on your gameplay.
Actually it shouldn't even be part of the sim anymore since Tukez recoded the sim to have a variable cooldown on Windfury to correspond to what is actually seen in game. Why there is such variability with haste EP values now really has me a bit confused. However running the sim with only 1000 hours is just asking for inaccuracy anyway.
1) Is my config incorrect?
2) Why am I getting more dps out of haste than crit when the EP value of haste is lower than crit?
3) Have other shaman at my gear level done extensive haste EP value testing? What values are your arriving at? A brief look at the top 20 shaman for XT (wowmeteronline) indicates a focus on crit over haste- what are the rest of you seeing?
Looks like it could potentially be a sampling issue. 0.34% crit is the reduction you should lose on spells with 16 less crit rating. Dropping 22 agility should lower your physical crit by a further 0.26%. However your crit loss seems much less than expected:
Actually it shouldn't even be part of the sim anymore since Tukez recoded the sim to have a variable cooldown on Windfury to correspond to what is actually seen in game. Why there is such variability with haste EP values now really has me a bit confused. However running the sim with only 1000 hours is just asking for inaccuracy anyway.
Yeah, the whole haste thing has me baffled as well. I ran it for 100000 hours one night and still came up with a negative haste for a 30 haste change, and positives for 31. I'm also getting it while being under the 1.5 speed fully hasted weapons and well over 1.0 speed.
Is that not correct? Windfury is 20%, Flurry is 25%, Swift Retribution is 3%. He has 13.76% haste unbuffed. So 1.1376 * 1.2 * 1.25 * 1.03 = 1.7576. 2.6 / 1.7576 = 1.479 attack speed with a 2.6 speed weapon. 12% haste would net you a 1.503 attack speed. If something is wrong with my math, please let me know!
Was anything other than the armor pen calculations changed for this new version?
Yes the Flametongue coefficients were fixed as well. Look back the thread a couple of pages to get the list of changes.
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.
Actually it shouldn't even be part of the sim anymore since Tukez recoded the sim to have a variable cooldown on Windfury to correspond to what is actually seen in game. Why there is such variability with haste EP values now really has me a bit confused. However running the sim with only 1000 hours is just asking for inaccuracy anyway.
The normalization only guarantees that you will get a positive haste value, not that the haste value will be consistent or progress linearly. As long as you aren't seeing negative values, everything is ok.
Is that not correct? Windfury is 20%, Flurry is 25%, Swift Retribution is 3%. He has 13.76% haste unbuffed. So 1.1376 * 1.2 * 1.25 * 1.03 = 1.7576. 2.6 / 1.7576 = 1.479 attack speed with a 2.6 speed weapon. 12% haste would net you a 1.503 attack speed. If something is wrong with my math, please let me know!
It doesn't matter. Windfury's cooldown is not 3 seconds. It is 3 seconds +/- 0.2 seconds with a bellcurve type distribution between those points. So worrying about hasting to specifically 1.5 speed is completely pointless.
It doesn't matter. Windfury's cooldown is not 3 seconds. It is 3 seconds +/- 0.2 seconds with a bellcurve type distribution between those points. So worrying about hasting to specifically 1.5 speed is completely pointless.
But aren't there still points where certain speeds net the most windfury procs? That seems to be the case from [Enh] Windfury Weapon Cooldown . This thread never conclusively determines a confidence interval or what is determined client side and server side. If you want to test what the actual windfury cooldown is, I would say that testing on a private server would be necessary to eliminate almost all lag. I believe that the .bat that is displayed when running one on your own comp displays damage done with probably almost perfect time stamps (though i could be wrong about this one.) It seems beneficial to stay near these cds but not on those cooldowns for the best results. Whether or not this is actually a dps increase/decrease would be another problem i suppose.
But aren't there still points where certain speeds net the most windfury procs? That seems to be the case from [Enh] Windfury Weapon Cooldown . This thread never conclusively determines a confidence interval or what is determined client side and server side. If you want to test what the actual windfury cooldown is, I would say that testing on a private server would be necessary to eliminate almost all lag. I believe that the .bat that is displayed when running one on your own comp displays damage done with probably almost perfect time stamps (though i could be wrong about this one.) It seems beneficial to stay near these cds but not on those cooldowns for the best results. Whether or not this is actually a dps increase/decrease would be another problem i suppose.
I believe the conclusion was that the server will hold procs that fall near the end of the CD window and fire them when the CD expires. While its true lag can not be discounted, no other proc/buff behaved this way with this sort of delay or separation from the actual hit. The only reasonable conclusion is that the server holds procs. If this is the case, then you will never get a negative EP for haste. You may not see a large proc increase per point at certain levels of haste, but you should not see a reduction.
I already posted this on the EnhSim wiki page, but just to make sure everyone who might benefit from it sees it...
I've just released version 1.0 of the newly-built iEnhSim, with or without source, at ienhsim - Google Code
iEnhSim is a Cocoa-based Mac GUI for EnhSim, something that I felt was sorely missing out there.
Hammer away, find bugs, please — The Google Code page has a nice bug tracker in there, so anything you find just submit a bug for it and I'll work on fixing them in the next release.
I've also added a category in the bug tracker for feature requests, so feel free to use it for that purpose as well.
Please note that this will run on 10.5 and later only, and is Intel-platform only (as I do not have a PPC machine to test against — anyone want to lend me one? =P)
1.6.9 - Game experience inconsistent with simulation
----------------------------------------------------------------
I'm not sure this is the right thread, but here goes:
I am getting simulation results which are inconsistent with the in-game testing. Details:
I am using EnhSim 1.6.9 (1.7.0 crashes due to uninitialized pointers). I am simulating two sets of gear, which I'll call Old set and New set: Old set - 4/5 valorous earthshatter (missing head item) New set - replaces earthshatter shoulder with valorous worldbreaker, earthshatter glove with valorous worldbreaker and earthshatter legs with [Leggings of the Tireless Sentry] . The exact details of the 2 sets are at the end of this message. In terms of set bonuses I am trading the 2piece and 4piece earthshatter bonuses in Old set for the 2piece worldbreaker bonus in the New set.
Anyway, I simmed the 2 sets and here are the results:
Old set - 4778 dps
New set - 4904 dps
(I tested with the 2 Rawr configs below). Obviously, New set is significantly better, right?
Except in game it's not. I went to 25man Emalon with my new gear. I am used to getting within the first 3-4 dps, and now I was somewhere in the middle of Recount. Then I went to the training dummy and did 3 x 5-minute fights with Old set on, and 3 x 5-minute fights with New set on. I did not use hero, the wolves nor any fire totem. Results:
Old set - 3173 average dps
New set - 3103 average dps
Old set is better!!!
I am further puzzled because my gems are calibrated around New set, which should make Old set even worse (e.g. I'm not expertise capped in it).
Please help me understand this. Thanks a bundle!
Attachments
------------------ OLD SET (put this in old.xml and open in Rawr):
3x 5minute combat durations in-game is so laughably random that you may as well have just made up numbers for comparison. Remember in the sim that you're running thousands (at least, you'd better be) of hours of combat to remove all the differences in crit rates and MW procs.
If you want the sim to reflect in-game a little more, turn up the min and max lag. Too often I see people posting here with 0 or 100ms on the lag configurations. That isn't realistic at all. Even if your actual network lag is 100ms or less, there is still a human reaction time to seeing an ability ready for use and then actually hitting the button. Most people probably ought to run their lag in the sim at anywhere from 200-500ms.
1.6.9 - Game experience inconsistent with simulation
----------------------------------------------------------------
I'm not sure this is the right thread, but here goes:
I am getting simulation results which are inconsistent with the in-game testing. Details:
It's not clear that your results are inconsistent with the sim. Our dps can vary wildly from fight to fight. It appears that you only did 4 runs that were 5 minutes long each.
Just as an experiment I ran enhsim with my stats with combat length set to 5 min and simulation time set to 20 minutes turning off all raid buffs except self buffs. Each time I ran the experiment I got different dps results, ranging from 4250 dps to 4465 dps. That's without changing anything between runs.
Try Verges's version of the sim, and if you're still convinced that the sim is wrong, you need to isolate what mechanic is wrong yourself, or provide hours and hours of testing that still show the inconsistency.
As an aside, it would be very helpful if the sim reported statistical measures in addition to average dps, so that people know what range of dps they should expect from their setup, and know the confidence of the average dps.
My friend asked me this question, and being the curious guy I am, I decided to do some research on it: Boss spell penetration. Now according to a wide variety of websites out there, NPCs have a 2% base resistance to spells per level above you that can not be overcome ( See: Spell penetration - WoWWiki - Your guide to the World of Warcraft ). Additionally, the simulator leaves spell resistance figures set at 0. Isn't this a rather important oversight? I saw spell resistance numbers all the time in a raiding situation in TBC.
If this is the case, what numbers should we be using for spell resistance? Would the standard 8*Lvl work?
I realize this has been more than likely discussed in passing before but I'm curious as to one's thoughts on the matter.
Why do some enhance shamans use the agility meta gem over the straight crit meta gem? Crit should have a higher EP value than agility always, should it not? It benefits spells as well as melee, meaning it helps to benefit Elemental Devastation which in turn boosts your melee crit. So why would one choose the agility meta over the crit meta?