Goroth simulations match real logs

Hello, human friend!

I’m glad you stopped by. Your gear is very important to Mr. Robot (did you know he polishes it every time you log out? I bet you thought it was always shiny and spotless because of some magic going on in Azeroth, didn’t you?)

Anyway! To give you the best gearing advice, you have to trust the simulation engine behind it.

How do you know Mr. Robot’s simulator is giving you accurate results?

Great question! We can compare simulations directly to logs, since we use real boss fights. Real player + real boss = verifiable results.

When you can simulate a top player and the report matches the log – you know the system is spot on.

I am proud to tell you that simulations match top logs for all DPS specs!

Comparing simulations to logs allows us to confirm (or fix) things like…

• Spell abilities – is the damage correct and are they affected by the right buffs?
• Rotations – do our rotations reflect what top players are doing?
• Special items – is the damage and proc rate matching on special items?
• Boss strategies – are some people using a different strategy or positioning? If so, we make modified scripts.

Each log is imported into our simulator, exactly matching the fight length, talents, gear, and buffs that the character had on that fight.

If we find something that doesn’t match, we investigate and fix it. We’ve evaluated thousands of logs per spec (and counting), and I wanted to share a sample of the test results with you. Trust, but verify, right?!

Comparing sims to logs – the setup

I look at the following:

• 95-98th percentile (this weeds out most of the super lucky streaks, but still captures top players)
• A sampling of different legendaries & trinkets
• Different talent builds
• Various iLevels that are appropriate for the script (in this case, Heroic Goroth)

A note about our boss scripts:

We model boss fights very accurately. You have a chance to get targeted with abilities – sometimes you get targeted, sometimes you don’t, just like in a real fight. We model movement and rotations use abilities appropriate for movement. The Goroth script is based on the actual mechanics and timing of the real fight – here is a description of the script.

I also have several boss scripts to match how people play. 

• Default – Assumes you move well and assumes ranged pre-position behind pillars, and that melee have to run behind the pillars and lose range of the boss.
• Move It! – This is for ranged – assuming you move a little less perfectly and don’t pre-position for the pillars (requiring you to spend time moving).
• Move It More – This also assumes you either lose target on the boss or forget to press your buttons, when moving to hide behind the pillars.
• Pro-Pillars: This is for melee – some teams are positioning Goroth near a pillar such that melee aren’t losing range of the boss during the hide mechanic.

To use the “Move It,” “Move It More,” or “Pro-Pillars” scripts, click on the link. When you are logged in, click the copy button. Give it a name, save it. Then you can select it in the “Boss Script” drop down menu on the simulator setup page.

To compare your own character:

1. Select your character in the log (like the damage-done view). Grab that URL.
2. Head over to the simulator and click the gray ‘import from log’ button at the center-top of the page.
3. Select your region and realm, and paste in the URL. Your talents, gear, and fight length are imported.
4. Be sure to select your consumables in the ‘more options’ section on the right. If you have other people with Infernal Cinders, you can set that in ‘more options’ as well. Then simulate!

Death Knight – Frost

Fablers (rank 95%) did 906k in the log and 904k in the simulations, for a difference of 0.2%

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar. Here are some of the top spells (logs vs sims):
  • Breath of Sindragosa: 108 hits vs 104
  • Obliterate: 78 vs 78
  • Howling Blast: 46 vs 47
  • Remorseless Winter: 9 uses vs 10, and 156 ticks vs 158
  • Crystalline Swords: 77 vs 78
•  Does the damage match? Yes! I compared average hit and crits on each ability too. Here’s a sample of a couple spells (logs vs sims):
  • Breath of Sindragosa: 44m vs 40m. The log got 5 more hits and a slightly higher crit rate, accounting for 2 million damage. Buff procs account for the rest.
  • Obliterate: 23.9m vs 23.7m (main hand) and 17.1m vs 16.6m (offhand)

A few more logs vs sims

• Asuhdude (rank 94%): 881k in the log / 876k in sims, difference of 0.6%
• Dessimus (rank 94%): 882k in the log / 894k in sims, difference of 1.3%
• Rtdee (rank 95%): 895k in the log / 905k in sims, difference of 1.1%

Tips for comparing to logs

• Extra movement away from the boss really impacts Breath of Sindragosa. The default script assumes you are out of range of the boss during Infernal Burning (hide behind a pillar) for 4-6 seconds (randomized). If you are not out that long and/or your team positions so you don’t lose range often, use the “Pro-Pillars” script.
• Obliterate: the logs list Crystalline Swords on separate lines, whereas it is group in sims (though each is listed out).
• Sindragosa’s Fury is listed as such on the logs, and “Frost Breath” on the sim report.

Death Knight – Unholy

Aceofblades (rank 96%) did 952k in the log and 956k in simulations, for a difference of 0.4%

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar. Here are some of the top spells (logs vs sims):
  • Virulent Plague: 131 vs 132 for the plague and 38 vs 39 for Eruption
  • Death Coil: 47 vs 47
  • Clawing Shadows: 45 vs 45
  • Festering Wound: 85 vs 79
  • Army of the Dead melee: 527 vs 511
  • Sludge Belcher melee (listed as Ghoul on sims): 135 vs 123
• Does the damage match? Yes! I compare the average hit and crit for each spell too. Here’s just a couple examples from this player (logs vs sims):
  • Virulent Plague: 24.5m vs 25m
  • Virulent Eruption: 8.74m vs 9.3m
  • Clawing Shadows: 24m vs 26m

A few more logs vs sims

• Sherz (rank 95%): 961k in the log / 953k in simulations, difference of 0.8%
• Lostsaint (rank 97%): 994k in the log / 987k in simulations, difference of 0.7%
• Coloringbook (rank 98%): 1010k in the log / 1004k in simulations, difference of 0.6%

Tips for comparing to logs

• Virulent Plague is grouped on Warcraft Logs, and grouped under “Outbreak” on simulation reports
• Sludge Belcher is reported as Ghoul in our simulations. It is the belcher, but implemented it this way for simplicity.
• Pestilence is grouped under Army of the Dead on logs, and it is not on the sim reports (it’s on it’s own line, at the bottom of the report).
• Pets: when you can’t target the boss (out of range), your pets can still attack in game, but not in our simulator. This is on our to-do list, and has a super, super small impact on your overall DPS. You can see when I compared melee hits from pets for the example above, the simulator has a few less hits – this is why.  We like to be open and honest, so I wanted to disclose this.

Demon Hunter – Havoc

Druid – Moonkin

Starcaller (rank 97%) did 925k in the log and 916k in simulations, for a difference of 0.97%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar:
  • Starsurge: 37 in log and 39 in sims
  • Lunar strike: 36 in log and 36 in sims
  • Solar wrath: 37 in log and 44 in sims – this number will be higher in the simulator if you aren’t really diligent about moving perfectly during global cooldowns.
  • Moonfire: 17 in log and 10 in sims. This will be higher if not moving perfectly.
  • Sunfire: 13 in log and 12 in sims.
  • Starfall: 9 in log and 8 in sims.
• Does the damage match? Yes! I compared average hit and crits on each ability.

A few more logs vs sims

• Shifhappens (rank 95%): 918k in the log / 934k in simulations,  difference of 1.7%
• Pombo (rank 97%): 982k in the log / 973k in simulations, for a difference of 0.9%
• Kinzun (rank 98%): 970k in the log / 966k in simulations, based on the “Move It!” script since he had extra movement)

Tips for comparing logs and sims

• Starsurge and Goldrinn’s Fang are separate lines in logs, and grouped together (but listed separately) on the simulator.
• You will see more moonfire and sunfire uses when you have to move more – since those are instant cast and can be used while moving.

Druid – Feral

Crittenkitten (rank 95%) did 885k in the log and 899k in simulations, for a difference of 1.5%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells
  • Shred: 74 in the log and 76 in sims
  • Rip: 14 in the log and 12 in sims
  • Rake: 114 hits in the log and 112 in sims
  • Ferocious Bite: 13 in the log and 12 in sims
  • Auto Attack: 281 in the log and 292 in sims
• Does the damage match: Yes! I compared the average hit and crit for each ability. A couple examples:
  • Shred: 42m total damage in log, 45m in sims. The sims got 2 extra hits, accounting for roughly 1m extra damage. Different buff uptimes account for the rest of the difference.
  • Rake ticks: 36m total damage in the log, 38m in sims. Different buff uptimes account for the difference.

A few more logs vs sims

• Itha (rank 96%): 907k in the log / 894k in sims, difference of 1.4%
• Whoopie (rank 97%) 922k in the log / 930k in sims, difference of 0.8%
• Opakanda (rank 96%) 893k in the log / 905k in sims, difference of 1.3%

Tips for comparing logs and sims

• Rake: The initial hit and the bleed ticks are on separate lines in the logs, and grouped (but listed separately) in the sims.
• Thrash and the bleed ticks are listed as 1 item (not separated) on the logs – but you can see the bleed hit number if the “Hits” column in parenthesis. They are grouped, but listed separately, on the sims.

Hunter – Beast Mastery

Chzburgers (rank 98%) did 1,040k in the log and 1,032k in sims, for a difference of 0.7%

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims)
  • Kill command: 48 vs 53
  • Pet melee: 136 vs 138. Hati melee is 124 vs 126
  • Dire Beast: Stomps are 38 vs 35. Auto attacks are 198 vs 196.
  • Cobra Shot: 61 vs 56
  • A Murder of Crows: 4 vs 3.6
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples:
  • Kill command main pet: 35.4m logs vs 39.2m sims. 5 extra uses from the simulator add up to an extra 3.7m damage.
  • Cobra Shot: 38.8m logs vs 34.3m sims. 5 extra uses in the logs adds up to 3.2m damage.

A few more logs vs sims

• Warbak (rank 98%): 928k in the log / 939k in sims, difference of 1.2%
• Andy (rank 95%): 990k in the log / 980k in sims, difference of 1%
• Fail (rank 96%): 993k in the log / 1005k in sims, difference of 1.2%

Tips for comparing logs and sims

• On Warcraft Logs – expand everything!
• On logs, pets are grouped under each ability. On our sims, we group the abilities under the pets.

Hunter – Marksmanship

Articuno (rank 95%) did 992k in the log and 986k in simulations, for a difference of 0.6%

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Aimed Shot: 69 vs 68
  • Marked Shot: 23 vs 21
  • Windburst: 9 vs 10
  • Arcane Shot: 73 vs 92 – it looks like the player was moving between Aimed Shots and not casting Arcane Shot as diligently as the simulator.
  • Auto Shot: 107 vs 111
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Aimed Shot: 157m vs 148m. 1 extra hit from the log adds 1.3 million dmg. This player got lucky with a 10% higher crit on logs than average – adding about 10 million damage to this log. Nice, right?!
  • Arcane Shot: 19.4m vs 23m. The extra shots in the sims account for the higher damage.

A few more logs vs sims

• Evilmon (rank 95%): 996k in the log / 1006k in simulations, difference of 1%
• Bananatrap (rank 98%):1041k in the log / 1034k in simulations,  difference of 0.7%
• Wompus (rank 96^): 1033k in the log / 1052k in simulations, 1.8%

Tips for comparing logs to sims

• Each Aimed Shot does such a massive amount of damage, that any unnecessary movement can cause you to lose a shot here and there. If each aimed shot averages 1 million damage, that is a 5k DPS loss for a 3 minute fight (per lost shot). Comparing your log to the sims will help you see if you are missing shots.

Hunter – Survival

So few people play this spec I had to expand the testing parameters to players above the 90th percentile. The 3rd ranked player in 915 gear is at the 92nd percentile, for example.

Moonjumper (rank 92%) did 1032k in the log and 1031 in simulations, for a difference of 0.05%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Mongoose Bite: 78 vs 77
  • Flanking Strike: 21 vs 20
  • Lacerate: 11 vs 10
  • Pet melee: 177 vs 168
  • Serpent Sting: 64 vs 65
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Mongoose Bite: 79m vs 73m.  Getting 4 more crits in the log (from luck, yay!) accounts for 4 million of the extra damage. 1 extra hit account for another 1 million of the difference.
  • Flanking Strike: 12.7m vs 11.4m. 1 extra hit and the higher than average crit in the log account for the damage difference.

A few more logs vs sims

• Mirred (rank 92%): 1011 in logs / 1008 in simulations, difference of 0.3%
• Deathstride (rank 90%): 995k in the log / 1012 in simulations, difference of 1.7%

Tips for comparing logs to sims

• Flanking strike has you and your pet’s damage grouped under it in the logs. On the simulation report, your pet’s flanking strike is grouped with the pet, and your’s is a separate line.
• Pets: when you can’t target the boss (out of range), your pets can still attack in game, but not in our simulator. This is on our to-do list, and has a super, super small impact on your overall DPS. You can see when I compared melee hits from pets for the example above, the simulator has a few less hits – this is why.  We like to be open and honest, so I wanted to disclose this.

Mage – Arcane

Happywheel (rank 98%) did 1015 in the log and 1006 in simulations, for a difference of 1%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Arcane Missiles: 224 vs 237 – the player had a little more movement than the script, losing a few hits.
  • Arcane Blast: 76 vs 75
  • Arcane Barrage: 14 vs 16.5
  • Mark of Aluneth: 4 vs 3.2
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Arcane Missiles: 95m vs 92m. This player had some buff procs line up nicely, increasing damage over the average.
  • Arcane Blast: 85m vs 79m. This player had some buff procs line up nicely. But to double check, you can go to this event view for Arcane Blast in the logs. Then expand the ‘log’ section on the simulator report and uncheck all event types except for damage and type “arcane blast” into the spell box. You can see the hits and crits match up nicely.

A few more logs vs sims

• Incendium (rank 98%): 921k in the log / 918k in simulations, difference of 0.3%
• Terencemage (rank 97%): 890k in the log / 903k in simulations, difference of 1.4%
• Arcana (rank95%): 882k in the log / 894k in simulations, difference of 1.3%

Tips for comparing logs to sims

• Arcane Blast: the total damage include Touch of the Magi in the logs, so be sure to expand the line item to see just Arcane Blast.
• With extra movement, you lose Arcane Blast uses, which will also result in dumping your Arcane Charges less often between burn phases, leading to less Arcane Barrage uses as well. You can also test your character out with the “Move It” script.

Mage – Fire

Nerisatis (rank 97%) did 888k in the log and 889k in the simulation, for a difference of 0.1%

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Pyroblast: 67 vs 67
  • Ignite: 203 vs 204
  • Fireball: 50 vs 61 – Simulations tend to have more because this is the first spell people cut out when they have to move (and simulated characters are really great at moving!)
  • Fire Blast: 28 vs 28
  • Phoenix Flames: 10 vs 11
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Pyroblast: 77m vs 75m. 1 extra hit in the log accounts for 1 million, and 1 extra crit accounts for the rest of the damage difference
  • Ignite: 51m vs 45m – Since ignite procs of other attacks, getting more damage off Combustion looks like it accounts for the difference.

A few more logs vs sims

• Avestival (rank 97%): 933k in the log / 923k in simulations, difference of 1.1%
• Merdyce (rank 98%): 907k in the log / 890k in simulations, difference of 1.8%
• Smouss (rank 96%): 907k in the log / 912k in simulations, difference of 0.5%

Tips for comparing logs to sims

• Movement can have a big impact on Fire Mages, so try the “Move It!” script or the “Move It More” script if you aren’t perfect at moving. Recall that the default script has you pre-position for Infernal Burning (hide behind pillars).
• Hot Streak: Look at total Hot Streak procs (not uptime) when comparing damage between the logs and sims. A higher uptime could actually be an indication of sub-optimal play.

Mage – Frost

Moordn (rank 98%) did 1036k in the log and 1034k in simulations, for a difference of 0.2%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Ice Lance: 56 vs 49
  • Flurry: 28 vs 28
  • Frostbolt: 77 vs 83
  • Ebonbolt: 4 vs 4
  • Ice Lance is only cast when you have Fingers of Frost or immediately after a Flurry (which you only cast with Brain Freeze). Both Fingers of Frost and Brain Frees are random procs with a flat % chance – meaning they are more prone to luck than an RPPM mechanic. Higher ranking logs tend to be on the luckier side, so you will see more Ice Lances and less Frost Bolts.
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Ice Lance: 87m vs 83m. The extra casts in the log add in some extra damage, but the lower crit rate lowers it down just a bit.
  • Flurry: 52m vs 50m

A few more logs vs sims

• Watchmeclick (rank 97%): 1020k in the log / 1034k in simulations, difference of 1.4%
• Tuskonator (rank 97%): 1017 in the log / 1028k in simulations, difference of 1.1%
• Victorvonmag (rank 98%): 1020 in the log / 1029k in simulations, difference of 0.9%

Tips to compare logs to sims

• Icicle and Frost Bolt are grouped together on the logs, so expand them to see both. They have separate lines in the simulation report.
• Look for less that 96-100% crit rates on Ice Lance in the logs – when you see it drop lower, you know you were casting a bunch of them ‘naked’ while moving (and possibly too much movement?)
• Lower Frost Bolt casts is also a sign of extra movement.

Monk – Windwalker

Paviliun (rank 96%) did 901k in the log and 906k in simulations, for a difference of 0.3%

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Rising Sun Kick: 33 vs 31
  • Fists of Fury: 11 vs 12 uses, 53 vs 56 hits
  • Strike of the Windlord: 6 vs 6
  • Crackling Jade Lightning: 3 vs 3
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Rising Sun Kick: 46.5m vs 43m (add Rising Sun Kick and Tornado Kicks together on the sim report). 2 extra casts in the log account for 3 million extra damage, plus a slightly higher crit rate.
  • Fists of Fury: 25m vs 29m (subtract Crosswinds from the total damage on the logs to get Fists of Fury damage). 1 extra casts in simulations is about 3 million of the damage difference. The extra crit rate accounts for the rest.

A few more logs vs sims

• Mynute (rank 95%): 896k in the log / 913k in simulations, difference of 1.9%
• Zeyra (rank 95%): 894k in the log / 884k in simulations, difference of 1.1%
• Leetzu (rank 95%) 884k in the log / 901k in simulations, difference of 1.9%

Tips to compare logs to sims

• This is one of the more complicated specs to compare.
• The more abilities you use while SEF is active, the lower your average damage with that ability will look. So when comparing a log to a sim, you can see some large differences in average damage if one or othe other uses significantly more of an ability during SEF. But that might not transfer to an overal DPS difference.
• Tornado Kicks is counted as Rising Sun Kick in logs (and separated out on the sim reports)
• Crosswinds is grouped under Fists of Fury in logs – you can expand it to see a line for each one. It is a separate line on the simulation report.

Paladin – Retribution

Bodi (rank 96%) did 900k in the log and 902k in simulations, for a difference of 0.2%

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Templar’s Verdict: 74 vs 73
  • Blade of Justice: 34 vs 33
  • Crusader Strike: 57 vs 56
  • Judgement: 24 vs 23
  • Melee: 84 vs 86
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Templar’s Verdict: 102m vs 100m. 1 extra cast in the log accounts for 1.4 million of the difference.
  • Blade of Justice: 39.5m vs 35.3m. 1 extra cast in the log accounts for just over 1 million of the difference. 3 extra crits in the log adds up to almost another 3 million of the difference.

A few more logs vs sims

• Austre (rank 96%): 892k in the log / 893k in simulations, difference of 0.1%
• Paladunno (rank 98%): 905k in the log / 867k in simulations, but this player had the Retribution buff up quite a bit, gaining him an extra 4% in damage. Add that to the simulation about and it brings it up to 902k. For a difference of 0.3%.
• Bruengar (rank 96%): 892k in the log / 886k in simulations, difference of 0.7%.

Tips for comparing logs to sims

• Retribution Buff: We assume your raid members aren’t dying, so you get no retribution buff in the simulation. We might add that in at a future point. For now, look at the % uptime you have on retribution in the log (buffs tab), and divide it by 5 (for the 20% dmg boost). So if you have 10% uptime, that gives you a 2% damage increase. This is rough math, but gets you close.
• Infernal Cinders: When you set up a simulation, there is an option to specify how many other people in your raid have the trinket. It is under ‘more options’ – toward the bottom of the list, under the Mythic Affix section.

Priest – Shadow

Jacoub (rank 98%) did 911k in the log and 923k in simulations, for a difference of 1.3%

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Void Bolt: 83 vs 88
  • Shadow Word: Pain: 240 hits vs 241
  • Vampiric Touch: 160 hits vs 161
  • Mind Blast: 73 vs 74
  • Mind Flay: 73 vs 85 – The player got less uses from a little extra movement
  • Mindbender: 4 vs 3.8
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Void Bolt: 43m vs 46.8m – 5 less casts in the simulations adds up to about 2.6m missing damage, and the 6% lower than average crit rate accounts for another 2m of the lower damage difference.
  • Vampiric Touch: 41m vs 40m
  • Mind Blast: 37.4m vs 38.6m

A few more logs vs sims

• Mindflay (rank 98%): 925k in the log / 936k in simulations, difference of 1.2%
• Lemeow (rank 98%): 934k in the log / 931k in simulations, difference of 0.3%
• Braenor (rank 97%): 936k in the log / 927k in simulations, difference of 0.9%

Tips for comparing logs to sims

• Void Form: the simulations get a best average – you might get lucky on near-perfect timing in the log, or you might not. Look for this when damage appears difference.
• Shadow Word Pain: Shadowy Apparitions are grouped together on the simulation reports, but separate ungrouped lines in the logs.

Rogue – Assassination

Farz (rank 95%) did 1009k in the log and 1014k in simulations, for a difference of 0.4%

• Does the rotation match? Yes – but see the note about Kingsbane below. To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Envenom: 29 vs 29
  • Deadly Poison: 250 vs 241
  • Kingsbane: 5 vs 7.2 – This is actually a significant difference. Our rotation is more liberal about Kingsbane use – we don’t necessarily hold Kingsbane to align it with Vendetta and instead go for more total uses, which usually ends up with slightly higher total damage.
  • Rupture: 10 vs 10
  • Melee: 250 vs 242
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Envenom: 40.4m vs 40m
  • Kingsbane: 20.5m vs 20.5m

A few more logs vs sims

• Flima (rank 97%): 1011k in the log / 1004k in simulations, difference of 0.7%
• Diego (rank 96%): 1014k in the log / 1018k in simulations, difference of 0.4%
• Fells (rank 97%): 1015k in the log / 1002k in simulations, difference of 1.3%

Tips for comparing logs to sims

• Deadly Poison is grouped in the simulation reports (but listed separately) and is not group in the logs.
• Kingsbane is grouped in the simulation reports, but not in the logs (look for each separate line)
• Mutilate is also grouped in the simulation reports, but not in the logs.

Rogue – Outlaw

Illusionaire (rank 95%): 961k in the log and 959k in simulations, for a difference of 0.2%

• Does the rotation match? Yes – but see the note about Ambush below! Although Outlaw Rogues are some of the hardest to analyze because of the randomness. To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Run Through: 52 vs 45 – (
  • Saber Slash: 97 vs 102
  • Blunderbuss: 13 vs 13
  • Between the Eyes: 22 vs 25
  • Pistol Shot: 37 vs 23 – The extra uses while running generated extra combo points, allowing for those few extra Run Throughs
  • Melee: 339 vs 321 – The higher uptime on Grand Melee explains the extra auto attacks
  • Our rotation doesn’t use Ambush out of Vanish with the Mantle of the Master Assassin, since it is higher damage on average, not to do so. So you might see some rotational differences among some players in logs.
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Run Through: 51m vs 51m. The log has more casts, but got unlucky with crits – getting a 34% crit rate compared to the expected average of 48%.
  • Saber Slash: 55m vs 59m. 5 extra casts in the simulator and a higher crit rate accounts for the extra damage.

A few more logs vs sims

• Fsblood (rank 95%): 923k in the log /  929k in simulations, difference of 0.6%
• Xelion (rank 96%): 931k in the log / 945k in simulations, difference of 1.5%
• Verniy (rank 95%): 978k in the log / 963k in simulations, difference of 1.5%

Tips for comparing logs to sims

• Saber Slash has 2 separate, non-grouped lines in the logs, so add them up when comparing to the simulation reports.
• This is one of the specs with the most damage fluctuation from one fight to the next, almost twice the variation of other classes. So don’t be surprised if you see wild swings above or below simulations (which show you what the average should be)

Rogue – Subtlety

Blurtarski (rank 96%) did 1010k in the log and 1019k in simulations, for a difference of 0.8%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Eviscerate: 34 vs 35
  • Shadowstrike: 35 vs 37
  • Backstab: 82 vs 76 – The player got more backstabs because of slightly higher uptime on the boss from suing Cloak of Shadows to not hide behind a pillar every 1.5 minutes.
  • Nightblade: 7 vs 8
  • Death From Above: 7 vs 7
  • Melee: 124 vs 122
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Eviscerate: 99.6m vs 98.2m. The log has a higher crit rate, accounting for the extra damage
  • Backstab: 30.7m vs 30m.
  • Nightblade: 28.5m vs 29.2m

A few more logs vs sims

• Karlennie (rank 96%): 1020k in the log / 1038k in simulations (Pro-Pillar script), difference of 1.8%
• Ladoum (rank 96%): 1026k in the log / 1040k in simulations, difference of 1.4%
• Mikasa (rank96%): 1029k in the log / 1013k in simulations, difference of 1.6%

Shaman – Enhancement

Lowland (rank 96%) did 992k in the log and 1002k in simulations, for a difference of 1%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Stormstrike: 42 vs 45
  • Windstrike: 20 vs 15
  • Lava Lash: 28 vs 28
  • Flametongue: 13 vs 13
  • Rockbiter: 31 vs 33
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Lava Lash: 41m vs 39m. The player got lucky crits (10% higher rate) accounting for the extra damage.
  • Stormstrike: 37m vs 38m

A few more logs vs sims

• Kruzxx (rank 97%): 970k in the log / 957k in simulations, difference of 1.3%
• Mcbull (rank97%): 1056k in the log / 1054k in simulations using the Pro-Pillar script to match his teams strategy, for a difference of 0.2%
• Dhalinar (rank 96%): 1030k in the log / 1029k in simulations, difference of 0.1%

Tips to compare sims to logs

• Stormstrike and Windstrike are grouped (but listed out separately) in the simulation reports, but not in the logs.
• Damage is greatly affected when not in range of the boss (like hiding behind pillars) – so a strategy where the tank gets the boss near a pillar will increase your damage. If you do that, use the Pro-Pillars script.

Shaman – Elemental

Rej (rank 98%) did 952k in the log and 950k in simulations using the Move It! script, for a difference of 0.2%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Lava Burst: 55 vs 54
  • Lightning Bolt: 44 vs 47
  • Earth Shock: 12 vs 13
  • Elemental Blast: 13 vs 14
  • Flame Shock: 8 vs 8
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Lava Burst: 52m vs 49m. 1 extra cast accounts for 1 million of the difference, and buff line ups account for the rest.
  • Lightning Bolt: 16m vs 16m

A few more logs vs sims

• Klick (rank 97%): 946k in the log / 951k in simulations, difference of 0.5%
• Orcuda (rank 98%): 947k in the log / 953k in simulations (Move It!), difference of 0.6%
• Hexanda (rank 98%): 964k in the log / 976k in simulations, difference of 1.2%

Tips to compare sims to logs

• Elemental Shamans can lose a lot of DPS from extra movement. If you find yourself moving a lot, or not being pre-positioned for Infernal Burning (hide behind pillars), try using the “Move It!” or “Move It More” scripts.
• Expand abilities like Lava Burst and Elemental Blast in the logs to see a breakdown by spell.

Warlock – Affliction

Ariadne (rank 98%) did 1.006m in the log and 1.005m in simulations, for a difference of 0.1%

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Unstable Affliction:  72 ticks vs 71 (other tracked ticks are 60 vs 59, and 44 vs 42)
  • Corruption: 152 hits vs 151
  • Agony: 167 hits vs 169
  • Doomguard: 101 Doom Bolts vs 93
  • Drain Soul – 39 vs 29 uses – The simulated player moved less, so Drain Soul didn’t get interrupted, thus using less.
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Unstable Affliction: 104.7m vs 94.2m. The player got a 10% higher than average crit chance, accounting for almost 4 million extra damage. The extra 1-2 hits on each line item account for another 2 million of the difference. A few 5-stacks plus lucky buff procs account for the rest.
  • Corruption: 26m vs 26m
  • Agony: 31.5m vs 33m

A few more logs vs sims

• Arevig (rank 98%): 1007k on the log / 989k in simulations, difference of 1.8%
• Wukar (rank 98%): 993k on the log / 1009k in simulations, difference of 1.5%
• Cradel (rank 97%): 988k on the log / 984k in simulations (Move It script), difference of 0.4%

Tips for comparing logs to sims

• Unstable Affliction – expand it to see that Compounding Horror is grouped with it on the logs, but a separate line in simulation reports.
• Drain Soul and Rend Soul are grouped in simulation reports, but not on logs
• Harvester of Souls is grouped under Corruption in the logs, but separate on simulation reports

Warlock – Demonology

Cholik (rank 96%) did 996k in the log and 1008k in simulations, for a difference of 1.2%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – below I review some of the top spells (logs vs sims):
  • Wild Imp: 2,532 total casts vs 2488
  • Call Dreadstalkers: 15 vs 17
  • Felguard: Legion Strike – 153 vs 153
  • Doomguard: Doom Bolt – 73 vs 77
  • Service Felguard: Auto Attack: 75 vs 81
  • Hand of Gul’dan: 41 vs 44
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Wild Imp Fel Firebolt: 90.4m vs 90.6m
  • Dreadstalkers melee: 21m vs 23m – The player has a lower than average crit rate and less auto attacks, accounting for the 2m difference.

A few more logs vs sims

• Fokdisheat (rank 96%): 1002k in the log /1008k in simulations, difference of 0.4%
• Morrthos (rank 96%): 1019k in the log / 1020k in simulations, difference of 0.1%
• Tjommie (rank 97%): 984k in the log / 1001k in simulations, difference of 1.7%

Tips for comparing logs to sims

• Wild Imp – Fel Firebolt is combined as one line in the simulation reports, but split between the Wild Imp and Call Dreadstalkers on logs. Add them up together to get the total damage and uses, to compare to simulations.
• Grimoire Felguard is the called Service Felgaurd in the simulation reports

Warlock – Destruction

Yadam (rank 95%) did 972k DPS in the log and 978k in simulations, for a difference of 0.6%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Chaos Bolt: 34 vs 34
  • Immolate: 115 hits vs 115 hits
  • Incinerate: 61 vs 67
  • Channel Doomfire: 12 vs 13
  • Conflagrate: 35 vs 38
  • Imp Firebolt: 118 vs 137
  • Dimensional Rift: 11 vs 11
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Chaos Bolt: 67m vs 63m. The damage difference comes from the lucky procs of Condordance of the Legionfall – the player got almost 10% more uptime due to lucky streaks 🙂
  • Immolate: 28m vs 28m
  • Incinerate: 27m vs 28m

A few more logs vs sims

• Sebbohrwl (rank 95%): 972k in the log / 972k in simulations, difference of 0%
• Yindra (rank 97%): 973k in the log / 988k in simulations, difference of 1.5%
• Emphanis (rank 96%): 980k in the log / 970k in simulations, difference of 1%

Tips for comparing logs to sims

• Immolate & the tick are grouped together on the logs, so be sure to expand it to see both.
• Grimoire Imp is called Service Imp on the simulation report.
• Dimensional Rift: Shadow Bolt, Chaos Barrage, and Chaos Bolt are grouped under Dimensional Rift in the logs. They are separate lines in the simulations.

Warrior – Arms

Kaltuthei (rank 96%) did 1034k DPS in the log and 1042k in simulations (Pro-Pillar script), for a difference of 0.8%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Mortal Strike: 47 vs 50
  • Whirlwind: 67 vs 67
  • Execute: 23 vs 27
  • Colossus Smash: 44 vs 37
  • Trauma: 114 ticks vs 114
  • Auto Attack: 81 vs 84
• Does the damage match? Yes – except for Opportunity Strikes (see note below).  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Mortal Strike: 76.6m vs 72.9m – The player had really lucky Concordance of the Legionfall procs (16% higher uptime), adding quite a bit of damage. Also, this player used all Shattered Defenses and Precise Strikes buffs on Mortal Strike during the execute phase, whereas we allow using an Execute. So we get higher execute damage and lower Mortal Strike damage. Here is a simulation that mimics what the player did, for those of you who are curious (1028k DPS).
  • Whirlwind: 53m vs 51m. The higher Concordance of the Legionfall procs account for the extra damage.
• A note for Arms: This is the only spec where we have an outstanding question: Opportunity Strikes proc rate appears to be different in game than what we have. We are working on verifying this, but it is incredibly difficult to do so in-game. Stay tuned!

A few more logs vs sims

• Heffa (rank 95%): 1027k DPS in the log / 1025k in simulations, difference of 0.2%
• Warduck (rank 96%): 1036k DPS in the log / 1016k in simulations (Pro-Pillar Script), difference of 1.9%. The lower Opportunity Strikes damage accounts for 20k DPS, which would bring the total DPS of the log and sim to match perfectly.

Tips for comparing logs to sims

• Execute uses and damage might differ depending how your spell priority. Our default rotation does more damage, but if you want to try it as described above (under the Mortal Strike line), use this rotation.
• Corrupted Blood of Zakajz on the simulation report can be comared to Battle Cry on the logs.

Warrior – Fury

Angrycatt (rank 96%) did 1056k in the log and 1056k DPS in simulations, for a difference of 0%.

• Does the rotation match? Yes! To evaluate it, we look to see if the # of casts are similar – here are some of the top spells (logs vs sims):
  • Raging Blow: 40 vs 38
  • Rampage: 22 vs 22
  • Execute: 12 vs 16 – The player capped rage during execute, including only gaining 5 rage from Battle Cry. This is why he has several less executes.
  • Auto Attack: 232 vs 209 – The player was in melee range just a little bit more than the script during Infernal Burning (hide behind pillars)
  • Bloodthirst: 43 vs 41
• Does the damage match? Yes!  I compared the average hit and crit for each ability. A couple examples (logs vs sims):
  • Raging Blow: 58.9m vs 58.1m. The player had a lower than average crit rate, but the 2 extra casts accounts for the slightly higher damage.
  • Rampage: 28.8m vs 28.6m
  • Execute: 18.2m vs 23.8m. The simulation was able to get 4 more executes, adding up to roughly 6 million damage.

A few more logs vs sims

• Exxeccutor (rank 96%): 944k DPS in the log / 945k in simulations, difference of 0.1%
• Revys (rank 97%): 1062k DPS in the log / 1063k in simulations (Pro-pillar script), difference of 0.1%
• Brayneless (rank 97%): 998k DPS in the log / 996k in simulations (Pro-pillar script), difference of 0.2%

Tips for comparing logs to sims

• Execute uses might differ due to a small rotational difference on some players. Our default spams Execute, but some players are using Raging Blow during execute. It ends up a slight DPS loss (1%), but if you want to try it, use this rotation.
• If your team keeps the boss closer to pillars during the Infernal Burning (hide behind a pillar), use the Pro-Pillar script. The default script has you out of melee range of the boss for 6 seconds during that spell, whereas the Pro-Pillar script has you out for 1-2 seconds.