EnhSim, DPS simulator - Page 184

Shammiewow · 11/25/09, 10:23 AM

Originally Posted by Levva

@Shammiewow

We really cannot help unless you post your config.txt file. BTW please use v1.9.0.0 to see if that makes a difference.

As you suggested, I tried 1.9.0.0 to see if it had the same large swings. I ran it four times at 10,000 hours and found that the results were almost without variance. The 4 runs varied by about 1 dps, or about 0.015%.

I'll be happy to post my 1.8.5.2 config if it would help you, but it seems like you've got this resolved in 1.9.0.0. Thank you!

Quaglia · 11/25/09, 11:39 AM

With this version it seems that also without using Fire nova there is a certain increase in DPS while moving MT up in the priority list.
But what it really puzzles me is that LL_QE seems to be rated quite low. I have run the simulator several times trying to put LL_QE in different places in the priority list and eventually I had the highest DPS putting it very low, just before "normal" LL.
Anybody observed the same behaviour? Before adjusting my priority accordingly, I would like to be sure that the sim is handling correctly the Totem of Quaking Earth.

My config file:

### Saved by v1.9.0.0 of EnhSimGUI

config_version                  1.9.0.0
simulation_time                 5000
simulation_time_combatlog       300
combat_length                   6.00
report_count                    80
threads                         1
combat_length_rng_factor        0.35
min_lag                         100
max_lag                         200
simulate_mana                   1
cast_sr_on_cooldown             0

ep_precision                    2
ep_base_stat                    ap
ep_ap                           80
ep_crit_rating                  40
ep_hit_rating                   40
ep_expertise                    4
ep_haste_rating                 40
ep_armor_penetration_rating     40
ep_spellpower                   46
ep_dps                          7.5
ep_mana                         600
ep_spirit                       40
ep_mp5                          40

mh_auto_attack                  1
oh_auto_attack                  1
wait_ss_with_wf_cd              0.00
cast_ll_only_if_wf_on_cd        0
bloodlust_casters               1
sync_bloodlust_with_trinkets    1
cast_lvb_only_if_ed_left        15.0
cast_lvb_only_if_fsdots_left    4
cast_lvb_only_if_fs_active      1
cast_fs_only_if_dots_left       0
cast_ls_only_if_charges_left    0
cast_sr_only_if_mana_left       1250
use_mana_potion_if_mana_left    3000

rotation_priority_count         12
rotation_priority1              SW
rotation_priority2              SR
rotation_priority3              MW5_LB
rotation_priority4              FS
rotation_priority5              SS_0
rotation_priority6              ES_SS
rotation_priority7              SS
rotation_priority8              MT
rotation_priority9              ES
rotation_priority10             LL_QE
rotation_priority11             LL
rotation_priority12             LS

miss                            8.00
dodge                           6.50
glancing                        24.00
armor                           10643
spell_miss                      17.00
nature_resistance               0
fire_resistance                 0
frost_resistance                0
arcane_resistance               0
shadow_resistance               0

armor_debuff_major              20.0/20.0
armor_debuff_minor              5.0/5.0
physical_vulnerability_debuff   4.0/4.0
melee_haste_buff                20.0/20.0
melee_crit_chance_buff          5.0/5.0
attack_power_buff_flat          687/687
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        5.0/5.0
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                   51/51
agi_and_strength_buff           178/178
intellect_buff                  60/60

replenishment                   1
water_shield                    0
mana_spring_totem               0
blessing_of_wisdom              1
judgement_of_wisdom             1
mixology                        1

flask_elixir                    flask_of_endless_rage
guardian_elixir                 -
potion                          potion_of_speed
food                            fish_feast

###############################################################################
### Everything in the section below can be replaced by information obtained ###
### from your paper doll stats or exported by the ShockAndAwe addon         ###
###############################################################################

race                            troll
mh_speed                        2.7
oh_speed                        2.6
mh_dps                          178.9
oh_dps                          196.5
mh_crit                         40.53
oh_crit                         40.53
mh_hit                          11.92
oh_hit                          11.92
mh_expertise_rating             139
oh_expertise_rating             139
ap                              4633
melee_haste                     14.35
armor_penetration               13.15
str                             147
agi                             1190
int                             665
spi                             170
spellpower                      1390
spell_crit                      29.52
spell_hit                       14.91
spell_haste                     11.04
max_mana                        14091
mp5                             0

mh_imbue                        windfury
oh_imbue                        flametongue

mh_enchant                      3789
oh_enchant                      3789

mh_weapon                       fist
oh_weapon                       axe

trinket1                        45286
trinket2                        45931

totem                           totem_of_quaking_earth

set_bonus1                      t9_battlegear_4
set_bonus2                      -
set_bonus3                      -

metagem                         41285

gloves_enchant                  -
cloak_enchant                   -

glyph_major1                    feral_spirit
glyph_major2                    flame_shock
glyph_major3                    stormstrike

glyph_minor1                    -
glyph_minor2                    -
glyph_minor3                    -

ancestral_knowledge             2/5
improved_shields                3/3
mental_dexterity                3/3
shamanistic_focus               1/1
flurry                          5/5
elemental_weapons               3/3
unleashed_rage                  3/3
weapon_mastery                  3/3
dual_wield_specialization       3/3
mental_quickness                3/3
improved_stormstrike            0/2
static_shock                    3/3
maelstrom_weapon                5/5

convection                      0/5
concussion                      5/5
call_of_flame                   3/3
elemental_devastation           3/3
reverberation                   2/5
elemental_focus                 0/1
elemental_fury                  5/5
improved_fire_nova              2/2
call_of_thunder                 0/1
unrelenting_storm               0/3
elemental_precision             0/3
lightning_mastery               0/5
elemental_oath                  0/2
lightning_overload              0/5
lava_flows                      0/3
storm_earth_and_fire            0/3
shamanism                       0/5

And the output:

EnhSim version 1.9.0.0

................................................................................

MH:
miss                1122162        9.07%
dodge               1211           0.01%
glancing            2970732        24.01%
hit                 1431237        11.57%
crit                6849315        55.35%
average ap          8107.08        
clip                0              0.00%

MH Windfury:
procs/hits          14.91%
procs/swings        13.74%
miss                0              0.00%
dodge               418            0.01%
hit                 1641767        41.71%
crit                2293517        58.27%
average ap          8107.35        

MH Storsmtrike:
miss                0              0.00%
dodge               181            0.01%
hit                 811708         41.71%
crit                1134229        58.28%
average ap          8088.63        

OH:
miss                1164534        9.09%
dodge               1283           0.01%
glancing            3073951        24.00%
hit                 1482344        11.57%
crit                7086141        55.32%
average ap          8111.32        
clip                0              0.00%

OH Flametongue:
miss                0              0.00%
hit                 8928123        56.34%
crit                6919439        43.66%
average sp          3033.46        

OH Stormstrike:
miss                0              0.00%
dodge               188            0.01%
hit                 812629         41.76%
crit                1133120        58.23%
average ap          8124.86        

Earth Shock:
miss                0              0.00%
hit                 1012709        56.42%
crit                782323         43.58%
average sp          3035.97        

Flame Shock:
miss                0              0.00%
hit                 493493         56.30%
crit                383111         43.70%
average sp          3035.90        
dots hit            2961415        56.30%
dots crit           2298209        43.70%
average dot sp      3035.90        

Lava Lash:
miss                0              0.00%
dodge               240            0.01%
hit                 942515         41.71%
crit                1316862        58.28%
average ap          8078.25        

Lightning Bolt:
miss                0              0.00%
hit                 1471830        56.32%
crit                1141371        43.68%
average sp          3040.30        

Lightning Shield:
miss                0              0.00%
hit                 3145048        100.00%
crit                0              0.00%
average sp          3041.34        

Spirit Wolves:
miss                0              0.00%
dodge               0              0.00%
glancing            2150308        23.97%
hit                 5788742        64.53%
crit                1031614        11.50%
average ap          8120.99        

Magma Totem:
miss                0              0.00%
hit                 4580652        56.35%
crit                3548033        43.65%
average sp          3038.10        

Maelstrom Weapon:
PPM                 48.58

flurry uptime                  93.73%
unleashed rage uptime          100.00%
elemental devastation uptime   79.75%
elemental oath uptime          0.00%

mh enchant uptime              41.77%
oh enchant uptime              36.94%
searing totem uptime           0.00%
magma totem uptime             88.03%

trinket1 uptime                18.52%
trinket2 uptime                20.62%
racial uptime                  6.88%
totem uptime                   91.71%
metagem uptime                 0.00%
gloves enchant uptime          0.00%
cloak enchant uptime           0.00%

                    MRPS without overregen
mana regen          0.00      0.00%
mp5                 18.10     8.80%
replenishment       42.19     20.52%
unrelenting storm   0.00      0.00%
judgement of wisdom 9.06      4.41%
shamanistic rage    136.23    66.26%
mana potion         0.00      0.00%
misc                0.00      0.00%
stormstrike         0.00      0.00%

                    DPS                 PPM       MPS
white               2531.73   31.38%    83.94
windfury            948.46    11.75%    6.56
flametongue         604.51    7.49%     52.83
stormstrike         447.12    5.54%     6.49      35.74     15.39%
lava lash           340.22    4.22%     7.53      20.75     8.94%
magma totem         585.02    7.25%     2.79      51.94     22.37%
earth shock         571.18    7.08%     5.98      38.67     16.65%
flame shock         406.24    5.03%     2.92      17.83     7.68%
lightning bolt      956.57    11.86%    8.71      63.82     27.49%
lightning shield    346.34    4.29%     1.16      0.00      0.00%
spirit wolves       331.30    4.11%     0.41      3.43      1.48%

DPS                 8068.70
MPS                 232.18
MP2min              27861.30
MRPS                209.96
Out of mana time    0.17%

elapsed simulation time: 5000.00h
elapsed real time: 52.41s
simulation speed: 343472x

Edit:
I noticed that the totem uptime was 91.71%. Is it this value realistic with the current priority?

Levva · 11/25/09, 11:47 AM

I've added the magma totem ticks option. Its very interesting, setting this to a non zero value or even just in v1.9.0.0 moving MT higher up the priority list results in a significant MT boost. This will be due to higher uptime.

V1.9.0.1 will be out soon.

Saidbeastie · 11/25/09, 2:11 PM

I dunno if this is useful-- maybe not. But the way I handle magma totem while I'm in an actual raid is such:

If all my abilities are on cooldown, and lightning shield is up with a high number of stacks, I refresh magma totem regardless of how much time it has left (using my better judgment of course; I don't see myself refreshing the totem at over 5 seconds). The way I see it, the DPS loss from a couple reset swing timers is made up for by magma totem being up nearly 100% of the time, which for me it is.

I think the way to model this in the sim would be to add another rotation item called, say, MT_<5 that could be added into the bottom of the priority queue.

I obviously don't know how effective my own method is without the simulator to tell, but I think going into 3.3 it'll be necessary to model a more robust technique with magma totem.

Rouncer · 11/25/09, 2:59 PM

Personally, I refresh it at 10 seconds or less when I have a free GCD.

The option that Levva has integrated in 1.9.0.1 will allow you to tell the sim the minimum number of ticks remaining on the current Magma Totem before it will start looking for a GCD under the priority list structure to refresh the totem. Since you like 5 seconds that gives you a choice between either 3 ticks (6 seconds) or 2 ticks (4 seconds). Probably better to stick to 3 ticks, though, as that will more likely represent your current behavior.

ziff · 11/25/09, 11:21 PM

Looking at the graphs above, it seems pretty consistent that EnhSim is hiting some Haste spikes when haste is increased. Or let me be clearer, from those graphs haste is plateauing for wide ranges where you would hope it would have some kind of positive impact on our dps. I'd be interested on seeing which damage components are the main culpret here, just so that we can verify that the code is working correctly for those components.

Leevva, could you post another graph with a breakdown of each damage component for one of the abovee simulation runs? Maybe we can see some outlier that would require some further investigation.

Levva · 11/26/09, 4:43 AM

Originally Posted by ziff

Looking at the graphs above, it seems pretty consistent that EnhSim is hiting some Haste spikes when haste is increased. Or let me be clearer, from those graphs haste is plateauing for wide ranges where you would hope it would have some kind of positive impact on our dps. I'd be interested on seeing which damage components are the main culpret here, just so that we can verify that the code is working correctly for those components.

Levva, could you post another graph with a breakdown of each damage component for one of the abovee simulation runs? Maybe we can see some outlier that would require some further investigation.

I only cut n paste the data others prepared and stuck it in Excel to produce the graph. I'm sure that Extenze could see how feasible this is to add output of different data.

Rouncer · 11/27/09, 2:29 AM

Originally Posted by Saidbeastie

If all my abilities are on cooldown, and lightning shield is up with a high number of stacks, I refresh magma totem regardless of how much time it has left (using my better judgment of course; I don't see myself refreshing the totem at over 5 seconds). The way I see it, the DPS loss from a couple reset swing timers is made up for by magma totem being up nearly 100% of the time, which for me it is.

Responding to this point at the request of another shaman.

Casting any INSTANT cast spell, as a shaman, does NOT reset your swing timer. Casting a spell with a cast time when you have any charges of Maelstrom Weapon active does NOT reset your swing timer. When there are any charges of Maelstrom Weapon active, if the cast timer intersects with the swing timer then that swing simple does not occur, ie the swing timer is clipped not reset.

Lchaim · 11/27/09, 3:10 AM

Is there any way we could get an option to cast CL instead of LB if additional targets > n? (or just > 0)
It seems that this would give more accurate data for AoE situations.

e: I specifically mean that if there are adds 25% of the fight, you would probably cast CL 25% of the fight, not 100%.

Levva · 11/27/09, 5:23 AM

Originally Posted by Lchaim

Is there any way we could get an option to cast CL instead of LB if additional targets > n? (or just > 0)
It seems that this would give more accurate data for AoE situations.

e: I specifically mean that if there are adds 25% of the fight, you would probably cast CL 25% of the fight, not 100%.

Good idea. I've added that as an issue.

For those that haven't yet seen v1.9.1.2-beta is now available at EnhSim I'd like testers to test all the changes please and report back things tested and proved working as much as reporting bugs. We'd like to get this version to the point we are confident of calling it a release not just a beta.

Rouncer · 11/27/09, 9:57 AM

Well, we can't call it a release until Blizzard pushes 3.3 anyway. Speaking of which, if anyone has access to the PTR (mine is throwing errors at me when I try to patch to the latest version) would they mind testing the interaction of Elemental Focus (the talent) and Fire Nova. Last I checked, Fire Nova could proc Clearcasting from the talent but it didn't get the mana reduction nor would it use a charge when cast while it was active.

Levva · 11/27/09, 11:04 AM

I very strongly suspect we only have until the 8th ie: 11 days so I'd like to get it ready for then as there has been a lot of changes and we need some stability checks to verify all is well.

Personally I'm also spending a lot of time on Rawr as well as EnhSim so personally I really can't put in the time to get EnhSim tested. It is better anyway if someone who isn't coding does the quality control checks. As it is a lot easier to verify your tests as you are not bound by the assumptions the programmer made and thus would test things he may not have thought of.

If some volunteers could help we can provide a full list of changes that require testing and sign off. Rouncer could you possibly co-ordinate this task given that you aren't one of the programmers?

• Jessamy · 11/29/09, 5:00 PM

My character config, enhsim v1.9.0.0-beta. 1000 hours per sim, 5 runs per increment, 1% (0.01) increment step size. 1.53 mean stdev.

Same setup, 10000 hours per sim. 0.48 mean stdev.

Extenze · 11/29/09, 7:17 PM

I ran the usual 10 sims pers RNG, and took the average of the 10 to get these values. 5k hour simulations
Edit: I could've sworn I've typed this in, but I got these numbers by running 10 sims per RNG change, and averaging them otherwise the graph would've been monstrous.

	white	WF	FT	SS	LL	MT	ES	FS	LB	LS	SW	totalDPS
0	2737.02	1026.85	612.07	437.61	342.88	521.98	584.07	317.76	962.48	316.12	334.47	8193.3
0.05	2747.28	1029.42	613.73	437.77	342.77	521.83	584.26	317.72	965.02	316.8	329.07	8205.66
0.1	2755.45	1031.6	615.11	437.97	342.74	521.92	584.53	317.64	967.22	317.39	316.03	8207.6
0.15	2758.49	1032.46	615.65	438.16	342.88	522.1	584.76	317.67	967.97	317.79	308.44	8206.36
0.2	2760.31	1033.01	615.96	438.24	343	522.23	584.84	317.68	968.44	317.93	304.6	8206.24
0.25	2761.02	1033.23	616.11	438.22	342.9	522.21	584.88	317.66	968.78	317.99	304.01	8207.01
0.3	2761.74	1033.46	616.2	438.18	342.89	522.08	584.83	317.64	968.95	318.06	308.06	8212.09
0.35	2762.26	1033.65	616.26	438.17	342.8	522.11	584.92	317.61	969.09	318.05	311.17	8216.08
0.4	2762.5	1033.81	616.28	438.18	342.79	522.01	584.91	317.62	969.21	318.09	310.3	8215.69
0.45	2762.66	1033.84	616.35	438.24	342.85	522.12	584.81	317.61	969.24	318.16	308.37	8214.24
0.5	2763.05	1033.8	616.44	438.27	342.88	522.17	584.85	317.63	969.3	318.18	306.72	8213.29
0.55	2762.91	1033.75	616.39	438.27	342.91	522.09	584.94	317.64	969.32	318.19	306.27	8212.67
0.6	2763.13	1033.99	616.43	438.25	342.81	522.11	584.9	317.63	969.41	318.2	308.14	8214.99
0.65	2763.36	1034.16	616.43	438.26	342.8	521.99	584.89	317.63	969.48	318.24	309.96	8217.19
0.7	2763.61	1034.27	616.51	438.26	342.83	521.99	584.82	317.57	969.41	318.21	309.45	8216.93
0.75	2763.78	1034.32	616.51	438.23	342.84	522.12	584.89	317.63	969.42	318.31	308.39	8216.42
0.8	2763.6	1034.04	616.48	438.27	342.86	522.13	584.77	317.63	969.42	318.32	307.36	8214.89
0.85	2763.78	1034.33	616.49	438.21	342.84	522.04	584.82	317.66	969.42	318.26	306.91	8214.77
0.9	2763.87	1034.32	616.5	438.17	342.82	522.08	584.96	317.6	969.53	318.28	308.14	8216.26
0.95	2763.78	1034.33	616.55	438.23	342.85	522.11	584.83	317.63	969.49	318.4	309.39	8217.57
1	2762.93	1033.81	616.38	438.25	342.83	522.05	584.96	317.61	969.31	318.26	309.21	8215.59

There's a graph in my spreadsheet, but I don't really see any usable detail in it.

Subtracting 0.00 rng values from the 1.0 values you can see what changes the most.

white	WF	FT	SS	LL	MT	ES	FS	LB	LS	SW	totalDPS
25.91	6.96	4.31	0.63	-0.06	0.07	0.89	-0.15	6.82	2.14	-25.26	22.28

edit2: Here's the spreadsheet I used, containing all the raw data.
http://dl.dropbox.com/u/816613/attacksOut_me_5k.ods

Levva · 11/30/09, 7:23 AM

I was investigating an issue reported by inconsistencies in Rawr.Enhance & EnhSim with [Talisman of Volatile Power] now it turns out there were bugs in both programs. It did throw up an interesting issue what precisely constitutes "spell damage" and "spell casts" as far as special effect triggers are concerned. A special effect trigger is something that initiates a proc - eg: Set bonus LL triggering Quaking Earth, or spell cast triggering trinket.

Currently EnhSim has the following definition for spell cast...

  public pShockCast,
  public pLightningBoltCast,
  public pLightningOverloadCast,
  public pChainLightningCast,
  public pLavaBurstCast,
  public pLightningShieldCast,
  public pFireNovaCast

Now I note no mention of things like magma totem drops which are spell casts, and apparently do trigger the talisman shouldn't these be included?

We also have the related Spell Damage...

  public pShockDamage,
  public pLightningBoltDamage,
  public pLightningOverloadDamage,
  public pChainLightningDamage,
  public pLavaBurstDamage,
  public pFireNovaDamage,
  public pLightningShieldDamage

The sim has no concept of separating non periodic damage, periodic damage and spell damage. I suspect we should make a separation between them. So that we can have effects proc on any spell damage (including periodic damage such as FS Dot ticks and Magma/searing totem pulses), periodic damage (FS Dot ticks and magma/searing totem) and non-periodic damage (as above).

I may be wrong but doesn't most trinkets that proc off spell damage also proc off magma totem damage & FS Dot ticks?

Whilst I'm musing on these things, where does Flametongue weapon imbue damage come into this? Should it be treated as spell damage? Does it proc things that come under the spell damage list? If so then the Sim is ignoring this source of spell damage at present.