Actual v Sim DPS

[TeaLeaf]

I've been concerned a little by our lack of DPS (e.g. NewKidz can out-dps us!!) so far this tier so I thought I would play with some numbers to see if it made me feel any better - it didn't.

The chart uses SimDPS (Patchwerk fight, 450 secs, 1000 iterations) based on our current gear and compares it with the actual DPS recorded by WoL. For example, it shows that in Immersius the very top bar did 76% of the SimDPS the person's ilevel says they can do.

What the comparison does not do is take account of the extra movement we have to do, AOE v single target opportunities, whether the fight is melee or ranged friendly, the down time in the fight (waiting for new phase etc), off-spec roles (hymn/healing etc).

BUT it does allow a rough and ready comparison of how we each did compared to the average for each encounter. The theory being that it might show a trend if for example whatever the fight you are always significantly lower than the relative % DPS that others do in the fight (ie a comparison with the average %simDPS should help eliminate some, if not the bulk of any of the factors not taken into account when comparing the actual dps records with a Patchwerk SimC number).

You can see from the graphs that we have a huge variability in the degree of performance relative to the level of DPS that we might expect from the SimDPS number.    A difference is to be expected as no encounter is a perfect Patchwerk fight, but we need to look at the trends and there are some who are tending to consistently under-perform their gear in whatever type of encounter, some who have good & some bad encounters and some others who are relatively consistently out-performing.

In the interests of fairness we've tried wherever possible to use the highest dpse number available in the logs for each encounter when working out the % performance relative to the SimDPS number.  We've also removed the names from the chart (and no they are not class colour coded either).

[ATTACH=CONFIG]2411[/ATTACH]

There is no way we should be out-dps'd by a guild who are 20th in realm ilevel when we are 11th in the ilevel list and we started with significantly better Heroic gear.

We need to do a reality check on our own performance, so please check the below links and look at whether or not you are performing at the right level please.    

You'll find your log dps here:
http://www.worldoflogs.com/guilds/38782/

You'll find your SimDPS here:
http://www.wowprogress.com/simdps/eu/aszune/

Seek to understand where your actual DPS differs from SimDPS (and it always will differ, for better or for worse).

Are there specific fights where you need to work on your ability rotation/talent choices or survivability?

When you find where you are weak then go and compare yourself to other characters of other top guilds and see where their rotation or talent build differs.   Are they using different talents, different spells or different timing compared to you?
You can find other guilds' logs here:
http://www.worldoflogs.com/realms/251/

Where can you improve?   That's for you to find out please and for all of us to benefit from!

Note:
There may be reasons why on a specific fight you are lower or higher than the SimDPS expection.   We know this, we just want you to explore & come to understand why your personal performance is different and to work to improve it where there is the opportunity or where you can see from other logs that similarly geared, same class toons in other guilds are doing significantly better dps.  We know there is room for improvement though!

[Ziley]

Hmm wowprogress puts me 40-50k below what I'm doing and also thinks I'm unholy. It doesn't look like their data is quite up to date or do I need to update my profile manually?

Edit: I also think that Iron Juggernaut might be the best fight for comparing dps with other guilds (in terms of pure rotation/gemming & reforging) since it is the one that is closest to a patchwerk encounter. This is probably also one of the best fights for comparing with simmed dps. Obviously we should still look at a fight to fight basis to see whether we can do something in each fight to better ourselves.

Also I'm not sure I quite understand the first graph, is that a bar for each of our dps? Also, would it be possible for me personally to see which one of those bars is me (or can you tell me?).. Apologies if I missed something very obvious.

[Ziley]

Why is there such a small amount of raiding frost DKs on Aszune.. I can't find anyone to compare with.

[Devlan]

Í'm happy with how close to my SimDPS I am on Malkorok and Iron Juggernaut, I'm also happy that I'm the only Fire Mage from Aszune-EU to be ranked on WoL, even if it IS generally in the 175+ area, considering I am the 30th Mage on wowprogress' SimDPS for Aszune.

However, I can still see places where I can improve, even if only a little. Specifically Mage Bomb application and Invocation uptime.

[TeaLeaf]

Yes it is one bar per player per fight (where log data exists for that fight).

I can tell you which one you are if you really want but it is not for publication as we want people to do their own research on things and work it out for themselves. So if you have done your own work then ask away when I am online.

I'd also rather not let this thread descend into identification by elimination for obvious reasons!

Finally, keep in mind that the sim number is a Patchwork fight so you'll only get closer to that number (in theory) when you are doing a Patchwerk fight.

Sent from my Nexus 4 using Tapatalk 4

[Ziley]

I'm very surprised that wowprogress ranks me so highly on the simdps list (no. 3) when in comparison, I'm barely in the top 20 in terms of gear - and I don't even have the legendary cloak. I'm generally pretty happy with my performance on patchwerk type fights where my two biggest problems are AMS usage after the new glyph introduced and also the usage of Plague Leech. However, I'm still a little behind the sim-dps, a difference that doesn't seem to be caused solely by those two factors as they are very minor (or I at least screw them up way too little for it to matter that much). I'm not great at analyzing logs, but I'm gonna take a look at what could be causing this discrepancy (and also do my own sim, in case the wowprogress ones are flawed in some way).

Also TL, I don't think the graph is much use without names, but obviously we should not have the names revealed, that is - in my opinion - up to people themselves to reveal and discuss if they are interested in outside feedback. Since I don't know which one is me (although I can of course figure out my own dps in comparison to my simmed data) it's only going to give me a general idea of the raid group, which I imagine is the purpose with the graph as well.

[TeaLeaf]

Following up on this there is an interesting table at Noxxic which uses SimC Version 540-4 (2013-09-26) to show how dps scales with ilevel.   Whilst I have not been a fan of the quality of Noxxic's info in the past, in this case they are reporting someone else's info (SimC's) and I'm already very happy with the quality of SimC's open-source work, so the list is well worth a look.  

This simulation offers a 'Maximum DPS' scenario (ie a Patchwerk simulation as in the OP), but also looks at a 'Realistic DPS' scenario (which uses what is known as a 'Helter Skelter' simulation.

We use a modified version of the Helter Skelter encounter with different raid events. This event list is a work-in-progress and we’re constantly tweaking the parameters in search of a better realistic fight model.

Casting: Increased variance to the Casting condition to cover more fight types.
Movement: Increased base cooldown on Movement, but with more variance.
Stun: Increased variance to the Stun condition to cover more fight types.
Distraction: Decreases the ‘skill’ level of players to simulate being distracted.
Optimal Raid: Disabled to reflect a more likely use of raid buffs.
Tricks of the Trade: Rogues still cast this on themselves.

The basic idea of a Helter Skelter simulation is to acknowledge that most fights are not Patchwerk fights, so it includes allowances for Casting problems, Movement, Stuns, Distraction, Raid setup and personal buffs.  i.e. it more closely resembles what we should expect to see in a typical boss encounter.   You can find the full two-part notes on the model here:
http://noxxicgaming.tumblr.com/post/34715734535/warcraft-dps-rankings-fight-types
http://noxxicgaming.tumblr.com/post/37338282586/warcraft-realistic-dps-rankings-part-ii

The original web page (which includes options to switch between Maximum DPS, Realistic DPS and different ilevels) is here:
http://www.noxxic.com/wow/dps-rankings/realistic

The list of what is theoretically possible surprised me in a few places.   For example, I hadn't noticed Arcane becoming so competitive or how good Subtlety might be.   The only thing I already knew was the shadow did not scale well with ilevel and the table confirmed that in spades! :roflmao:

This is the 'Realistic (Helter Skelter) DPS' list at i553:

#1|Frost |292,494 DPS |100%
#2|Feral |292,044 DPS |99.85%
#3|Subtlety |287,037 DPS |98.13%
#4|Survival |285,199 DPS |97.51%
#5|Enhancement |279,905 DPS |95.7%
#6|Destruction |279,691 DPS |95.62%
#7|Arcane |275,826 DPS |94.3%
#8|Balance |275,788 DPS |94.29%
#9|Beast Mastery |273,703 DPS |93.58%
#10|Marksmanship |272,531 DPS |93.17%
#11|Affliction |270,295 DPS |92.41%
#12|Windwalker 1h |270,279 DPS |92.4%
#13|Combat |264,406 DPS |90.4%
#14|Assassination |263,427 DPS |90.06%
#15|Fire |261,947 DPS |89.56%
#16|Fury 2h |261,822 DPS |89.51%
#17|Elemental |261,153 DPS |89.28%
#18|Windwalker 2h |261,108 DPS |89.27%
#19|Demonology |259,098 DPS |88.58%
#20|Frost 1h |258,497 DPS |88.38%
#21|Retribution |255,452 DPS |87.34%
#22|Fury 1h |252,751 DPS |86.41%
#23|Frost 2h |247,055 DPS |84.46%
#24|Arms |241,426 DPS |82.54%
#25|Unholy |236,052 DPS |80.7%
#26|Shadow |222,114 DPS |75.94%Clearly there are other considerations here, it's not a list to say "you should do this level of DPS", after all it's an imperfect simulation, but it gives us some data to work with, some ideas to think about and might help us fine tune what we do.  

Read, digest, cogitate & then come out fighting!

I raided with someone who plays a low ranked spec, but they owned my DPS … how is that possible?

There are any number of reasons why in-game results will differ from simulated (or log based) results. A few common differences include the following.

Fight Length:
Heroic dungeons and LFR raids often have shorter fight durations which can result in much higher DPS as the value of DPS cooldowns is greatly increased in short fights.
Multiple Targets:
The current DPS rankings are strictly single-target with no opportunity to multi-DoT, cleave, AoE, etc. This has a large impact on in-game DPS.
Player Variables:
Skill, latency, gear, a lucky string of crits, etc. can all significantly impact in-game DPS. Rankings can help represent general trends, but they’ll rarely predict the outcome of a specific encounter with all its unique factors.

[Ziley]

Well, TL.. I guess we can look forward to partying it out at the bottom of the meter!
Quite surprised to see Subtlety and Frost topping, as an old-school player I'd never have expected that!

It also looks like Jas should be topping us all whenever he goes kitty, what's your explanation Jas?