EnhSim, DPS simulator - Page 13

Hothgor · 09/21/08, 6:34 PM

Originally Posted by Pitbuller

You tested without raid buffs. You tested in buggy PTR(this is huge). I tested with optimal condition with sim. You tested couple dozen minute. I runned sim 25000h. Level 70 ep and level 80 ep values is clearly different becouse rating drops. Both values are in this thread. You can easily say if simulator do something wrong. But I can't say how scientific your test was. At 70lvl there wasn't any optimized gear to these mechanic but at 80lvl there is. Those 80lvl ep values what I posted show clearly that melee gear is strictly better at 80lvl. I know those values aren't perfect but those are best we have now. In game testing is good for testing how mechanic work but in game test isn't very good to test dps. In best scenario someone test how mechanic work and then someone add that to sim. Then sim work just like real game but it can produce lot more numbers. If we wanna know how we do good dps we want lot of those numbers. You scored 2200dps in your test but we want to know how to do 4000dps. It's good that there is in-game testers but actual numbers isn't interesting. Maybe if you re-produce least 15minute test with all raid buffs when ptr is stable enough and all our skills work there. Then those numbers can actual tell about gearing for raiding.

I still don't understand this argument. The simulator should be able to tell me what my damage should be raid buffed or self buffed. If it only works while raid buffed then its broken and buggy. While 25,000 hour tests are wonderful to filter out noise, I am more concerned with real playing conditions...and last time I checked, there were not 25,000 hour long boss fights!

Pitbuller · 09/21/08, 6:49 PM

Sim work fine if you turn raid buffs off. But why anyone would be interesting optimize damage at solo content?
Just compare 25-man naxx gear.

Valorous Earthshatter War-Kilt
1125 Armor
+67 Agility
+66 Stamina
+66 Intellect
Red Socket
Yellow Socket
Socket Bonus: +9 Stamina
Equip: Improves hit rating by 36 (0.97% @ L80).
Equip: Improves critical strike rating by 51 (0.98% @ L80).
Equip: Increases attack power by 170..

Valorous Earthshatter Kilt
1125 Armor
+99 Stamina
+77 Intellect
Yellow Socket
Blue Socket
Socket Bonus: +6 Hit Rating
Equip: Improves hit rating by 42 (1.13% @ L80).
Equip: Improves critical strike rating by 59 (1.13% @ L80).
Equip: Increases spell power by 98.

Enhancement gear have:
+67agility
+170 ap
=~304ep

Elemental gear have:
+11 int
+6 hit
+8crit
+98 spell power
=~161ep

Understood?

Edit: Just look those gear pieces. Enhancement War-kilt give only 28spell power less than elemental kilt when you factor talents. If you say that elemental gear do same damage than enhancemen gear then you have to spec full elemental. And if you don't trust sim then you could point what is wrong there?

Hothgor · 09/21/08, 7:20 PM

Originally Posted by Pitbuller

Sim work fine if you turn raid buffs off. But why anyone would be interesting optimize damage at solo content?
Just compare 25-man naxx gear.

Valorous Earthshatter War-Kilt
1125 Armor
+67 Agility
+66 Stamina
+66 Intellect
Red Socket
Yellow Socket
Socket Bonus: +9 Stamina
Equip: Improves hit rating by 36 (0.97% @ L80).
Equip: Improves critical strike rating by 51 (0.98% @ L80).
Equip: Increases attack power by 170..

Valorous Earthshatter Kilt
1125 Armor
+99 Stamina
+77 Intellect
Yellow Socket
Blue Socket
Socket Bonus: +6 Hit Rating
Equip: Improves hit rating by 42 (1.13% @ L80).
Equip: Improves critical strike rating by 59 (1.13% @ L80).
Equip: Increases spell power by 98.

Enhancement gear have:
+67agility
+170 ap
=~304ep

Elemental gear have:
+11 int
+6 hit
+8crit
+98 spell power
=~161ep

Understood?

Because most people cant test their raid damage on the raid targeting dummy with raid buffs for extended periods of time, and boss fights in a raid vary in complexity and don't allow conductive testing. The next best thing is solo testing. Hence we tested Yo's simulator in a solo situation until we were sure it was accurate, THEN worried about raid damage.

Understood?

Levva · 09/21/08, 8:58 PM

You are both arguing different points.

Hothgor is making a very simple point : He sets up the sim to match his gear and buffs and does some tests. He runs the sim WHICH EXACTLY MATCHES HIS TEST, he gets 2200 dps say and the sim says 3200 dps. Now one of two things is happening here. The sim doesn't match the game, and he's uncovered a bug, or Hothgor sucks at Enh Shammy dps.

Either way the point he is making is very very simple. It is possible to setup the sim to simulate your exact setup and predict your dps, if you are getting substantially more or less dps in game then that suggests the sim is NOT modelling the game and needs attention.

Pitbuller's point is entirely different, he is looking to see the extreme min max effects to see what the sim can tell us about the nature of the changes Blizzard are implementing and how that points towards gear selection, gemming, glyphing, enchanting etc to maximise our dps.

BOTH arguments are equally valid. I personally favour Hothgor's approach at present for one extremely simple reason, if the sim can't accurately match what you can do in game then its internal coding is fundamentally flawed, Assuming ofc that you set up completely matching variables and follow exactly the rotation/priority you enter into the sim.

We must be able to match game values to sim values if we can't the sim won't scale and will be useless at end game raiding. ie: You need to prove its got the basics right before you stress test it with ultimate min/maxing.

PS. I have faith in Tukez & the community to be able to make the sim work. However we must be able to prove it works and its not just some hypothetical unobtainable numbers its spitting out.

Rouncer · 09/21/08, 11:04 PM

This is something that will have to be checked once the build stabilizes to make sure the Sim is reporting real data, that said at this stage of the game you really can't expect that much accuracy. Stormstrike isn't reporting in Combatlogs or on damage meters (and for parses it shows as belonging to nil) and that's just the tip of the iceberg in regards to what isn't currently working as intended in the Beta/PTR.

Also don't forget that we are a very bursty class, very RNG oriented. That's why we run the sims for as many hours as we do, because Windfury procs can vary greatly. Also until you are spell hit capped that can have a huge effect on what the meters will show and even if you are capped that normally will mean your crit will be low which means less procs, etc.

If the sim can get within +/- 10% of the value of a Beta test run on a targeting dummy with as many variables accounted for as possible then it will have proven itself reliable in my book. But we are still quite a ways away from having a stable enough mechanics for a test like that.

darkhorse · 09/22/08, 4:42 AM

Originally Posted by Levva

You are both arguing different points.

Hothgor is making a very simple point : He sets up the sim to match his gear and buffs and does some tests. He runs the sim WHICH EXACTLY MATCHES HIS TEST, he gets 2200 dps say and the sim says 3200 dps. Now one of two things is happening here. The sim doesn't match the game, and he's uncovered a bug, or Hothgor sucks at Enh Shammy dps.

BOTH arguments are equally valid. I personally favour Hothgor's approach at present for one extremely simple reason, if the sim can't accurately match what you can do in game then its internal coding is fundamentally flawed, Assuming ofc that you set up completely matching variables and follow exactly the rotation/priority you enter into the sim.

We must be able to match game values to sim values if we can't the sim won't scale and will be useless at end game raiding. ie: You need to prove its got the basics right before you stress test it with ultimate min/maxing.

I have been looking through Hothgor's posts & can't find any where he actually gives the results of a simulation. All I could find was his postings about his results of his personal beta tests? If you could point it out to me that would be good because it does change the entire equation.

Currently, it looks like hothgor has posted a series of findings made from tests on a very bugged & nebulous server. The shear fact that there are so many bugs (such as WF not working) will in most cases invalidate all findings due to inaccurate data collection. Yes his data might be accurate now but it has no real use for the future when the bugs are fixed.

You cannot make a comparison between two setups when only one of those are functional.

The other thing to remember is that you CANNOT compare the current (broken) server (with broken mechanics) to a "theoretically correct" simulation. There is NO comparison possible at the moment. They are both working on two different mechanics & basis. It is akin to a druid plugging their numbers into a priest simulator because they are both healers.

Personal Server testing will become important in the future. When the server is stable & has all the major bugs fixed, live testing with dummies will become necessary to validate the simulations. Until that stage live tests are largely a meaningless statistic as they will change each bug fix.

Hothgor · 09/22/08, 6:13 AM

Originally Posted by darkhorse

I have been looking through Hothgor's posts & can't find any where he actually gives the results of a simulation. All I could find was his postings about his results of his personal beta tests? If you could point it out to me that would be good because it does change the entire equation.

Currently, it looks like hothgor has posted a series of findings made from tests on a very bugged & nebulous server. The shear fact that there are so many bugs (such as WF not working) will in most cases invalidate all findings due to inaccurate data collection. Yes his data might be accurate now but it has no real use for the future when the bugs are fixed.

You cannot make a comparison between two setups when only one of those are functional.

The other thing to remember is that you CANNOT compare the current (broken) server (with broken mechanics) to a "theoretically correct" simulation. There is NO comparison possible at the moment. They are both working on two different mechanics & basis. It is akin to a druid plugging their numbers into a priest simulator because they are both healers.

Personal Server testing will become important in the future. When the server is stable & has all the major bugs fixed, live testing with dummies will become necessary to validate the simulations. Until that stage live tests are largely a meaningless statistic as they will change each bug fix.

You actually don't know what you are talking about. My testing is showing what different weapon/enchant combination's do in actual testing situation. The only bugs on the BETA realm have to do with Stormstrike and WF weapon. Only the higher ranks of WF weapon are currently bugged on the BETA server. And it's well known that SS does roughly 10% of our damage even on live. The simulator itself indicates that FT/FT weapon enchants do very similar damage (less than ~1% overall) to WF/WF or WF/FT enchants. Its miss leading to say that you can not test these abilities under the current circumstances!

darkhorse · 09/22/08, 6:28 AM

Originally Posted by Hothgor

You actually don't know what you are talking about. My testing is showing what different weapon/enchant combination's do in actual testing situation. The only bugs on the BETA realm have to do with Stormstrike and WF weapon. Only the higher ranks of WF weapon are currently bugged on the BETA server. And it's well known that SS does roughly 10% of our damage even on live. The simulator itself indicates that FT/FT weapon enchants do very similar damage (less than ~1% overall) to WF/WF or WF/FT enchants. Its miss leading to say that you can not test these abilities under the current circumstances!

That is my point. What they are doing is currently meaning less until blizzard fixes the bugs. Until all the bugs have been fixed there is no way to say what is empirically better because you admit that you can't test all the options.

In fact if you look at it further of course your current tests are going to lean towards spell dmg. Two of your biggest Melee attacks (Stormstrike & WF) are broken. Thus the melee component is smaller & the tests become skewed towards the spell dmg side. It is an unbalanced test.

Real life testing will become necessary when the bugs are fixed. Until that time however all anyone who professes "actual beta testing" is presenting data that is unreliable & flawed.

But you have already shown that others can't convince you, so I don't have any expectations that what I say will change your mind.

Hothgor · 09/22/08, 6:33 AM

Originally Posted by darkhorse

That is my point. What they are doing is currently meaning less until blizzard fixes the bugs. Until all the bugs have been fixed there is no way to say what is empirically better because you admit that you can't test all the options.

In fact if you look at it further of course your current tests are going to lean towards spell dmg. Two of your biggest Melee attacks (Stormstrike & WF) are broken. Thus the melee component is smaller & the tests become skewed towards the spell dmg side. It is an unbalanced test.

Real life testing will become necessary when the bugs are fixed. Until that time however all anyone who professes "actual beta testing" is presenting data that is unreliable & flawed.

But you have already shown that others can't convince you, so I don't have any expectations that what I say will change your mind.

Thats not your point at all because you dont seem to understand how things are working out.

WF/WF with the max rank WF spell is doing within ~1% damage as someone using FT/FT setup. We don't need to discuss WF here because it wont have any bearing on the test except for in the OPPOSITE direction of what you are getting at. If i were to use max rank WF/WF, my hits per 8 seconds would go up dramatically. I would be getting far MORE MW procs, meaning my spell damage portion would go up.

Lets say that for right now, I'm doing ~20% damage with FT/FT and 31% with LvB. If I were to use WF/WF, It would STILL do ~20% of damage with WF attacks, but 33-34% with LvB. See how that works? The missing SS damage IS a problem. However, it will be adding the same damage to all my different tests because I am always casting 1 SS every 9 seconds with the rotation that I am using. With it well known to do ~10% of total DPS, I can get a reasonable approximation of total DPS. However, Searing Totem is ALSO doing ~10% of my total DPS, and I haven't even been INCLUDING these two categories in my test deliberately!!!

The sad thing is, you don't even know why I'm posting these tests! If you were to read the exchanges between myself and Tukez, you would see that I am helping him to iron out the kinks in the program by giving him real world data on Lava Lash mechanics and other aspects. I'd hardly consider that 'worthless'.

And guess what! With all the information punched into the simulator as it is, its giving me a very close approximation to the damage that I am doing on Beta!

Tukez, some problems!

Lava Lash is doing way more damage on BETA than your simulator is indicating. Lava Burst has also been doing ~40% more damage than the simulator seems to be indicating. Is it possible you dont have the correct base spell damage range and damage coefficient in place? Searing totem has done ~240 dps in all 25 test I conducted tonight, but the simulator seems to believe it should be doing 300 or so. SS also seems to be a tad low, but this may be due to the fact that I'm using a spell damage weapon for the MH instead of a beefy melee dps weapon :P

darkhorse · 09/22/08, 6:51 AM

Thank you for testing lava lash. And I am saying that honestly. That is a section of our code that does work correctly so all testing of that component is totally valid.

You can defiantly test the sub components that work for validation & I do thank you for taking the initiative for helping get the lava lash section fully.

My point was the overall data is skewed when any one sub-component is broken. But this is enough back and forth

tukez · 09/22/08, 7:17 AM

Originally Posted by Hothgor

Tukez, some problems!

Lava Lash is doing way more damage on BETA than your simulator is indicating. Lava Burst has also been doing ~40% more damage than the simulator seems to be indicating. Is it possible you dont have the correct base spell damage range and damage coefficient in place? Searing totem has done ~240 dps in all 25 test I conducted tonight, but the simulator seems to believe it should be doing 300 or so. SS also seems to be a tad low, but this may be due to the fact that I'm using a spell damage weapon for the MH instead of a beefy melee dps weapon :P

Hard to say. The way of doing it should be, that people figure out how a mechanic work in the game and then I implement it. With Lava Lash, I have been mostly guessing the mechanics.

In the sim, LL is 100% or 125% off hand damage. 50% off hand penatly is applied. It works like a melee attack, but with fire damage. That means, it can miss, dodge, crit and hit. It can be partially resisted, but not fully. Critical strike damage is with melee multiplier...200% or 206%(with RED). Damage is affected by fire damage debuffs.

Make sure you are using the level 80 version of the sim.
LvB coef 57.14%
Lvb base dmg 1230

Searing coef per tick 16.8%
Searing base dmg 90

Stopokingme · 09/22/08, 7:22 AM

At the moment it's also getting the bonus crit from Elemental Fury though, not sure yet if that's working as intended or a bug.

tukez · 09/22/08, 7:30 AM

Originally Posted by Stopokingme

At the moment it's also getting the bonus crit from Elemental Fury though, not sure yet if that's working as intended or a bug.

Are you able to tell me what is the crit bonus with and without Elemental Fury?

Pitbuller · 09/22/08, 8:26 AM

Searing totem use 2.2s cast. But it act like Npc. Npc cast seems lag some unknow time. This thing aply also warlock imp and mage water elemental. Unkown part is also is this lag caused by server lag or new beta mechanic or something what is also in live. So more testing needed. How often searing totem actual fire bolts?

Aximous · 09/22/08, 8:37 AM

I think you should include [Totem of Ancestral Guidance], [Skycall Totem], [Totem of the Void] and [The Skull of Gul'dan] in the sim as they seem to get useful with MW I think.

Stopokingme · 09/22/08, 9:08 AM

Originally Posted by tukez

Are you able to tell me what is the crit bonus with and without Elemental Fury?

Only level 74 on the beta at the moment, but I can say that my LL crits hit for roughly 280% of the non crit damage. (This is with 4/5 elemental fury)

Pitbuller · 09/22/08, 9:41 AM

I made config file for 80lvl. Including all raid buffs/debuffs with updated values. Boss armor updated to 10850(same value that boss level test dummy). I used spec what give most damage with current values. Correct if something is wrong here. Stats are just educated guess what those should look when we start raiding at 80lvl. Weapons are epic item 200lvl crafted/pvp/pve. Trinkets are from 70lvl we just don't have any good data about new trinkets.
Some help needed about these:

-Base mana?(Naked mana minus naked int mana value)
-Base strenght?
-Boss armor value range?

This config yield 4179dps but I guess armor value is too low.

simulation_time                 5000
simulation_time_combatlog       30
combat_length                   6
report_count                    80
threads                         2
min_lag                         0
max_lag                         0

ep_ap                           300
ep_crit_rating                  100
ep_hit_rating                   100
ep_expertise                    10
ep_haste_rating                 200
ep_armor_penetration_rating     85
ep_spellpower                   300

mh_speed                        2.60
oh_speed                        1.4
mh_dps                          143.5
oh_dps                          143.5
mh_crit                         20.0
oh_crit                         20.0
mh_hit                          8.0
oh_hit                          8.0
mh_expertise_rating             50
oh_expertise_rating             50
ap                              2500
haste                           0
armor_penetration               0
str                             150
agi                             500
int                             500
spellpower                      750
spell_crit                      15
spell_hit                       10.00
base_mana                       2678

mh_imbue                        windfury
oh_imbue                        flametongue

mh_enchant                      mongoose
oh_enchant                      mongoose

trinket1                        berserkers_call
trinket2                        shard_of_contempt

totem                           stonebreakers_totem

set_bonus                       -

metagem                         relentless_earthstorm_diamond

glyph_major1                    -
glyph_major2                    -
glyph_major3                    -
glyph_major4                    -

glyph_minor1                    -
glyph_minor2                    -
glyph_minor3                    -

wait_ss_with_wf_cd              0.00
use_searing_totem               1
bloodlust_casters               2
sync_bloodlust_with_trinkets    0
cast_lvb_only_if_ed_left        15.0
cast_lvb_only_if_fsdots_left    4
cast_lvb_only_if_fs_active      1
wait_mw_if_lvb_cd_left          0
cast_fs_only_if_dots_left       0

rotation_priority_count         7
rotation_priority1              fs
rotation_priority2              mw_lvb
rotation_priority3              mw_lb
rotation_priority4              ss
rotation_priority5              ll
rotation_priority6              es
rotation_priority7              ls

miss                            9.00
dodge                           6.50
glancing                        25.00
armor                           10850
spell_miss                      17.00
nature_resistance               0
fire_resistance                 0
frost_resistance                0

ancestral_knowledge             3/5
improved_shields                0/3
mental_dexterity                3/3
shamanistic_focus               1/1
flurry                          5/5
elemental_weapons               3/3
unleashed_rage                  5/5
weapon_mastery                  3/3
dual_wield_specialization       3/3
mental_quickness                3/3
improved_stormstrike            2/2
static_shock                    3/3
maelstrom_weapon                5/5
convection                      0/5
concussion                      5/5
call_of_flame                   1/3
elemental_devastation           3/3
reverberation                   0/5
elemental_focus                 1/1
elemental_fury                  5/5
call_of_thunder                 0/1
elemental_precision             0/3

armor_debuff_major              3925/3925
armor_debuff_minor              1260/1260
physical_vulnerability_debuff   2.0/2.0
melee_haste_buff                20.0/20.0
melee_crit_chance_buff          5.0/5.0
attack_power_buff_flat          688/688
attack_power_buff_multiplier    0/99.7
spell_haste_buff                5.0/5.0
spell_crit_chance_buff          5.0/5.0
spell_crit_chance_debuff        10.0/10.0
spell_damage_debuff             13.0/13.0
spellpower_buff                 280/280
spell_hit_chance_debuff         3.0/3.0
haste_buff                      3.0/3.0
percentage_damage_increase      3.0/3.0
crit_chance_debuff              3.0/3.0
stat_multiplier                 10.0/10.0
stat_add_buff                   52/52
agi_and_strength_buff           178/178
intellect_buff                  60/60

flask_elixir                    elixir_of_major_agility
guardian_elixir                 elixir_of_draenic_wisdom
potion                          haste_potion
food                            spice_hot_talbuk
misc_item                       drums_of_battle

Malan · 09/22/08, 10:19 AM

Originally Posted by Pitbuller

I used spec what give most damage with current values.

I'm curious - did you run the numbers to see that 1% Int is worth more than 5% Lightning Shield damage?

Tukez - can you perhaps give an option to begin combat with a FS- LvB combo? I've been doing this while soloing and its noticeably different when you begin combat with Elemental Devastation up.

tukez · 09/22/08, 10:30 AM

Originally Posted by Malan

I'm curious - did you run the numbers to see that 1% Int is worth more than 5% Lightning Shield damage?

Tukez - can you perhaps give an option to begin combat with a FS- LvB combo? I've been doing this while soloing and its noticeably different when you begin combat with Elemental Devastation up.

That would have zero effect even with a bit longer fight...not even talking about hours. Or are you trying to optimize your leveling?

Pitbuller · 09/22/08, 10:39 AM

+5% lighning shield damage hould be 123dps * 0.05 = 6.15dps.
2% more intellect with [500(base+gear) + 60(int buff) + 30(elixir) + 52(motw)] * 1.1kings = 706int
0.02 * 706 = 14 int.
Int ep value is about 1.33ep and dps/ep value is about 0.53dps/ap.
So 2% int is 14*1.33*0.53 =~ 10dps gain.

But somehow sim give only 2dps gain. Tukez can you check why Ancestral knowledge isn't working right.(this is new problem last time I checked it worked)

Bad news is that our t7 2set bonus is worth of ~12dps.

Malan · 09/22/08, 10:41 AM

Originally Posted by Pitbuller

2% more intellect

Its not 2% Int though - in the spec you linked you have 3 points in Ancestral Knowledge, 1 point of that isn't required to reach the next tier, and 1 point = 1%. Therefore the decision is 1% Int, or 5% LS bonus. Which means from what you posted above that the point is probably better off in Improved Shields.

Originally Posted by tukez

Or are you trying to optimize your leveling?

Yarr.

Rouncer · 09/22/08, 10:48 AM

Originally Posted by tukez

Are you able to tell me what is the crit bonus with and without Elemental Fury?

250% with Elemental Fury and 200% without.

There are no spell partials for Lava Lash, it can have glancing hits if used with a weapon skill that isn't capped but it uses the melee mechanic not the spell one.

tukez · 09/22/08, 11:57 AM

Originally Posted by Rounced

250% with Elemental Fury and 200% without.

There are no spell partials for Lava Lash, it can have glancing hits if used with a weapon skill that isn't capped but it uses the melee mechanic not the spell one.

250% and 200% don't make any sense. 100% more crit damage with physical means 300% damage. That would also make sense with 4/5 EF giving 280%.

Are you sure about, that there are no partial resists? And I thought glancing hits can happen with white damage only.

Pitbuller · 09/22/08, 12:03 PM

Ancestral Knowledge
Rank 1
Increases your intellect by 2%.

If it would be only 1% then we would be very envy to paladin counter part. They get 3% per point it's also five point talent.

tukez · 09/22/08, 12:27 PM

Originally Posted by Pitbuller

+5% lighning shield damage hould be 123dps * 0.05 = 6.15dps.
2% more intellect with [500(base+gear) + 60(int buff) + 30(elixir) + 52(motw)] * 1.1kings = 706int
0.02 * 706 = 14 int.
Int ep value is about 1.33ep and dps/ep value is about 0.53dps/ap.
So 2% int is 14*1.33*0.53 =~ 10dps gain.

But somehow sim give only 2dps gain. Tukez can you check why Ancestral knowledge isn't working right.(this is new problem last time I checked it worked)

Bad news is that our t7 2set bonus is worth of ~12dps.

Might be, that you are doing it wrong. If you increase Ancestral Knowledge, you also have to increase intelligence and spell crit. I know in these cases, it is kind of annoying, but it is consistent with other inputs. In the sim you always input what you see in your char sheet.

Well I don't know how you did the test, but I wasn't able to get 10dps gain. Dps gain was under 6dps. Maybe the EP values you have don't match with so little changes. You might want to try to calculate EP values again with lower ranges.

Then someone might think, that why do you need to input Ancestral Knowledge at all? Ep value calculation needs it. Also if you get 100 int temporary buff(trinket), with full AK, you actually get 110 int.