Sim DPS vs. Real DPS

[Athonel]

I am curious if other rogues are seeing something like this happen.

According to simcraft and AMR, my dps on Heroic Gruul should be around 35k-36k. Last night, I pulled 41k. My friend asked me a good question.

"If sims are supposed to represent perfect dps, how are you beating it?"

I don't have an answer. Does anyone care to weigh in?

*I'm not trying to brag about my dps, I'm genuinely curious*

Please note, I got the 715 ring AFTER the raid.

https://www.warcraftlogs.com/reports...8&type=summary

http://www.wowprogress.com/character/us/malorne/Clarith

http://us.battle.net/wow/en/characte...arith/advanced

[B L A D E]

You certain that you didn't change anything in terms of spell usage or any other settings?

[PessimiStick]

Sims attempt to predict perfect average DPS. There's a huge difference between that and any given fight. Rogues aren't even nearly the worst offender on that front. Ret in Highmaul could swing 30%+ of their average in either direction based on RNG.

[Athonel]

You certain that you didn't change anything in terms of spell usage or any other settings?

I didn't adjust anything. I just used the standard simcraft settings for mut rogue with my gear.

- - - Updated - - -

Sims attempt to predict perfect average DPS. There's a huge difference between that and any given fight. Rogues aren't even nearly the worst offender on that front. Ret in Highmaul could swing 30%+ of their average in either direction based on RNG.

So what percentile would you say sims are predicting? Would procs and buffs (ring, trinkets etc.) have a significant effect?

[Kyarxadron]

Its all about fight lenght.

[Athonel]

Its all about fight lenght.

So I should adjust fight length in sim to match my gruul kill and then go from there?

[Kyarxadron]

So I should adjust fight length in sim to match my gruul kill and then go from there?

Exactly. //

[Athonel]

Okay, I changed fight duration to match that of my kill time, and it brought the sim up to about 38k. Can I assume that the sim is not second potting under optimal conditions?

[PessimiStick]

Depends entirely on the APL you're using. The Sim pots when the APL tells it to.

[unbound]

Okay, I changed fight duration to match that of my kill time, and it brought the sim up to about 38k. Can I assume that the sim is not second potting under optimal conditions?

One of the limitations with the sims is that they are based on average percentages for RNG procs. If you happened to have a battle with good RNG procs, you will be higher than sims (potentially quite a bit depending on class, spec and talents). Conversely, you can run into battles where you will be a good deal lower than sims for the same reason (lousy RNG).

[S7orm]

4:20 fight, more heroism/cooldown uptime. Also RNG can lower/increase the DPS since sims calculate maximum average DPS

[Razzdu]

Probably the fight length. Look here:
https://www.warcraftlogs.com/reports...-done&source=2

See that big spike at the end? That's going to make your overall dps increase greatly because there's no corresponding valley on dps after it. It's just how dps is figured: damage done / time.

Here's a basic illustration, assume that each number is one second:

10+1+1+1+1+1+10+1+1+1+1+1+10+1+1+1+1+1=45
45 damage / 18 seconds = 2.5 dps

Now, shorten the fight length so that the third ten (the cooldown burst) is the last event:

10+1+1+1+1+1+10+1+1+1+1+1+10=40
40 damage / 13 seconds = 3.07 dps

Make sense?

Edit: way late, sorry.

[Kyzawolf]

I pretty consistently beat simmed dps, I use that as a benchmark for where I should be AT LEAST.

[Deleted]

With your current gear simulationcraft for a 250s patchwerk fight yields the following:

Sample Data	Clarith_Active Damage Per Second (Effective)
Count	9999
Mean	40006.64
Minimum	36096.95
Maximum	46868.84
Spread ( max - min )	10771.89
Range [ ( max - min ) / 2 * 100% ]	13.46%
Standard Deviation	1176.3615
5th Percentile	38389.28
95th Percentile	42215.74

This means that the average DPS over 9999 simulated fights was 40k, however there was at least one fight with 36k and one with almost 47K.
And about 90% of the fights were between 38k and 42k.

The log also shows that all you major damage spells had an up to 10% higher crit rate than the average 30% the simulation shows.

I think the sim also pulled bloodlust at a different time which may cause some differences as well, however your DPS can be explained by RNG alone.

[Athonel]

Some really good answers here. Thanks guys!

[Deleted]

EDITED: Similar Replys above

[Athonel]

With your current gear simulationcraft for a 250s patchwerk fight yields the following:

Sample Data	Clarith_Active Damage Per Second (Effective)
Count	9999
Mean	40006.64
Minimum	36096.95
Maximum	46868.84
Spread ( max - min )	10771.89
Range [ ( max - min ) / 2 * 100% ]	13.46%
Standard Deviation	1176.3615
5th Percentile	38389.28
95th Percentile	42215.74

This means that the average DPS over 9999 simulated fights was 40k, however there was at least one fight with 36k and one with almost 47K.
And about 90% of the fights were between 38k and 42k.

The log also shows that all you major damage spells had an up to 10% higher crit rate than the average 30% the simulation shows.

I think the sim also pulled bloodlust at a different time which may cause some differences as well, however your DPS can be explained by RNG alone.

Thank you for that. I wasn't looking closely enough at the sim results. This makes a lot more sense.

[bredar]

Its all about fight lenght.

Yes this. I believe Sims are over a certain period of time where Gruul is pretty short for Sub to burst higher

[Athonel]

Yes this. I believe Sims are over a certain period of time where Gruul is pretty short for Sub to burst higher

I'm not sub.

[Exyz]

I believe this was mostly iterated already, but RNG is definitely the culprit here. Assassination as a spec has a huge variance in damage, id suggest looking at the average number of dispatch procs and extra combo point crit procs indicated by your sim and compare them to what you got on that fight.