7.5k Ways to Die: DPS your way to happiness

[sphie]

I think everyone by now has had experience with Festergut 25/10 and Rotface25/10 to realize that we are not winning. We're making progress, but we're not winning. There are a multitude of reasons why we haven't made them faceplant and collect their loot. A little introspection is needed for us to discover why we aren't winning and what we need to do to rectify it. At the most basic level, there are only 3 roles in raiding. DPS, Tanking and Healing. Any one of those could be the reason why we aren't winning (or conversely, any one of those is the reason why we do win). But that's not what this post is about. Muddling through nebulous generalizations such as "our DPS sucks" or "our healing sucks" really doesn't help anyone, and blaming those concepts as the reason we lose serves to just cause more grumbling and unhappiness. So it might be more helpful to just break down one aspect of the 3 most basic levels of raiding and try to see what we can, individually, do to improve that level. Since I am more comfortable at the DPS role, I'll break that down.

Using some napkin math that I stole from someone else's napkin, it is generally accepted that in a 25man raid, the average DPS that everyone who fulfills the DPS role in the raid must produce somewhere between 7k and 7.5k dps in order to defeat Festergut (and judging from our attempts on tuesday, we need that kind of DPS as well for Rotface). Do we meet this criteria? Let's find out! (Warning: what's going to follow is heavy log analysis which will help you understand yourself better)

Lesson 1:
Reading and understanding world of logs

Before I even get to what needs to be done to increase your DPS, the first thing that needs to be done is to understand your own DPS. You need to understand yourself before you can improve yourself. Thanks to the efforts of Brad and Andy, we have an amazing resource called world of logs that chronicles our adventures in Azeroth. This will be your primary tool in auditing your own performance. Let's get started.

If we pull up our most recent raid, and check out the damage done (open this link up in a new window) Icecrown 25 we notice a few things. The first thing to note are the top outliers, those being Sphied, Logos and Ishildur pulling out the most DPS and total damage. It seems impressive, but upon further inspection, the overall instance DPS is very deceiving when it comes to actually beating content. As much as I enjoy towering over everyone when it comes to DPS, this is just an illusion, as most of my DPS comes from trash packs, which Unholy DKs excel at killing.

Now hover your mouse over "Full Report" and click on "Bosses" (don't go to individual fights yet). Now we see that i've dropped to 2nd place, and other classes have shifted for one reason or another. Now compare the DPS only on bosses vs the DPS on the entire raid. Notice something? My DPS in the entire instance is 8.2k, but my DPS on bosses only is only 6.2k. Uh oh! That's not good. You'll notice a similar drop off on pretty much everyone. I need to hit 7.5k! 6.2k isn't going to cut it! Okay, this DPS is also an illusion. It gives us a better picture, but there are still hundreds of confounding factors that make this DPS misleading. Some fights may be better for melee, some may be better for ranged, there might have been an early death, someone might have needed to change roles for one encounter. Lots of things could have happened! Let's drill down a bit more to something that's easier to get a grasp on.

Now hover over "All boss encounters" > "Kills" > "Deathbringer Saurfang"

A ha! Perfect! A good fight to measure single target DPS! Well, not quite. This fight is a good fight to measure single target DPS from melee, but is quite a bit harder for ranged since they have to kill Blood Beasts. So for this analysis, i'm going to ignore ranged DPS for this fight and focus on our main melee DPS. Those being Sphied, Devinoch, Camyu, Dept and Daewen. Since i've already mentioned that this is a melee favored fight, it's no surprise to see the top 4 being all melee (with Daewen at 6th).

Let's go down the line:
Sphied: As far as I can tell, this is probably the best fight I can have so far in Icecrown. No movement, minimal impact of AE which might inflate my DPS, basically just a tank and spank for me with an occasional taunt on a bloodbeast that strays too far. If we look at the DPS of the top 4 melee (7.5k, 7.48k, 7.4k, 7.3k respectively) are well into the range of "Good enough DPS!". My DPS is 7.5k, but my simulations that I ran with my gear and spec tell me that my ideal DPS would be slightly over 8k, so my real world performance is pretty close to the ideal performance through the simulation. Sweet! That means if i played perfectly, lined up all of my procs, didn't have to taunt or do anything but attack saurfang for 1000 hours, I'd get to around 8k dps. Ofcourse that's not realistic, but getting close to the simmed amount is a good sign telling me that I'm performing my rotation more or less correctly. But being who I am, I'm very competetive, even against myself. I see this 7.5 and I'm not satisfied in the least. I see other Death Knights in slightly better gear who are pulling 9.5k in the same fight. What am I doing wrong? Is it a gear issue? Is it a rotation issue? Is it a gemming/enchanting issue? I'll cover that in the next post since it's getting late.

Devinoch: Cliff is in the same boat as me. Our DPS is similar, but we'll probably need to do a lot of things to try to improve our DPS. I don't know jack about Ret paladins, so there really isn't much I can say.

Camyu: Wynne would have been in the top spot by a much larger margin, but Mutilate got nerfed quite hard so she fell to 3rd. Some other factors to consider are the presence of bleeds. Since we had 3 other characters who had bleeds, she didn't have to rupture (dps loss) to keep Hunger for Blood up.

Dept: Trevor had a funny time that night. It seemed like the WoW gods wanted him to mangle to reduce his DPS. Initially we had Avva providing Trauma (+bleed damage), but he was the first one to get marked so he had to go away. So Andy picked up the slack in cat form to to be the mangle bitch. But then the 2nd Mark, ofcourse, fell on Andy so Trevor had to start mangling for himself. Reports from Wynne state that Trevor was 2nd while he had a mangle bitch, but then steadily started dropping in DPS once Avva and Andy went away. He eventually dropped to 4th place.

Daewen: I know nothing about how combat rogues work, so I'll leave this blank too.

With a little work and research, everyone's position on the DPS charts can be fairly accurately reasoned out. Next episode we'll examine gear choices (including gems/enchants) and specs.

I didn't get as far as I wanted to tonight on this post. But my ultimate goal is to help people improve their DPS from solid to top class. Everything in this series will be very theory heavy, with lots of analysis, research, simulations and real world testing. This isn't a post for the faint of heart. I'm an accountant by trade so analysis, research, simulations and testing is what I do for a living. But if you have the drive and patience to reach the top percentile of DPS in Schad, you might be well served by following along. It's actually quite fun.

Tomorrow's lesson: Sphied discovers his DPS isn't that great after comparing himself to other death knights. Depression begins. But also the road to recovery begins.

[Texaporte]

I asked Cenon to do this series since his accountant mind works in better ways than my writer mind, but keep in mind that all of this advice is intended to be constructive and friendly. We're all playing a game together, but we all have to step up our game a little bit. We have, what, 18 dps on Rotface, and our wipes came consistently at 5 minutes. If every DPS gets just an extra 100 dps out of knowing what to do and knowing their class, we will get 18 * 300 = 540,000 extra damage in that time, which would bring him 10% further without very much effort at all. If we, say, bring every dps up to about 4.5k, which is where our low end should be given our gear levels and fight mechanics that require dps gaps, then we'll get a full extra million off.

It's a bunch of pebbles that can end with a mountain getting moved, and the mountains are named Rotface and Festergut!

[Daewen]

Let me add to this a cross-post from another thread, "new 80's and gear" or something or other: Some of people's upgrades are in ToC25, we need to add that to our raid schedule, and I think it will help us meet our ICC progression to farm it.

[Texaporte]

Let me add to this a cross-post from another thread, "new 80's and gear" or something or other: Some of people's upgrades are in ToC25, we need to add that to our raid schedule, and I think it will help us meet our ICC progression to farm it.

It would definitely help, but this thread is #1 about how to improve yourself. Knowing what to do as your class and when to do it is worth way more than the 50 dps upgrade from a piece of gear.

[hung]

Just in case the point hasn't hit home yet. Here's a quick visual aid: its a comparison of the dmg done in an instance. On the left is what our logs consistently look like. On the right, what we should strive for.

Log-capture.gif (21.16 KiB) Viewed 276 times

Its an ideal. Reality is, people leave early/join as replacements, various fight mechanics that are more advantageous to different classes/specs, there's differences in gear, etc... Still, if you're barely doing more dmg than the tanks, its time to ask yourself why. Chances are, you can do something different and do more. And chances are, you can ask a guildie for advice on what that something is.

[Dengeki]

TL;DR

[Werehamster]

Too snarky; didn't invite to raid.

[Texaporte]

Too snarky; didn't invite to raid.

If I left people on the sidelines for being snarky, the raid would consist of like 8 people.

Let's not sidetrack this much, though. Cenon's putting a lot of thought into how we can help each other get better, and I don't want to take over his thread.

[sphie]

Moving on!

Lesson 1.5:
Before I can actually get to talenting, I want to make a quick detour to another part of world of logs. You can also more deeply drill down into your own logs and also drill down into the logs of other characters in other guilds. This is a self assessment step. If you don't have any other class in the guild to compare your stats to, the next best thing is to compare your own logs to those of another class in another guild on the same fight. Take note though that this method of comparison is going to be hard to trust. The group makeup and buff distribution will most likely be very different and you don't know everyone's specific circumstances in each fight. So you're just going to have to go with your gut on this one and make an assumption that in a 25man situation, it's pretty likely that all the buffs will be more or less covered and you're roughly going to have the same job.

It sucks but what can ya do. First you need a specific fight. I'll use the same fight as in my previous post, Saurfang 25 but this time i'll drill down deeper. Click on the "Damage by Spell" tab.

So just for my own masochistic curiosity, I went to the root page of world of logs, and clicked the little icon on the left (above all the icecrown encounters) that says "View the DPS Player rankings in this zone". Click on the 25m normal tab. Wow look at that, almost ALL of the top damage dealers are rogues. So i'm going to have to view more on Saurfang to find an appropriate deathknight. A few dozen pages later, I find a death knight in possibly the best gear I've ever seen. It's a pretty unfair comparison, but here he is, from the guild CUTIES ONLY, Potlol and his parse

If we compare the two parses, we see the gear makes a STAGGERING difference in the damage we do. His autoswings on average do OVER 700dmg than mine. The shadow portion of his scourgestrike does an average of over 1100dmg per swing. We have the standard death knight specs but since he has t10 he has switched over to a reaping build. It's only a movement of 3 talent points, but it causes huge differences depending on gear and rotation.

Anyhoo, so what does this actually tell us? Well, it tells us that holy crap, all that gear does make a massive difference in damage. And on some level, no matter how much I tweak my spec and smooth out my rotation, there's no way I can cover the gear gap.

This kind of analysis actually isn't quite as valuable for a Death Knight. Seeing as we're on a pretty rigid rotation, there wont be much of a variance on the ratios of ability to ability. But say you're a resto druid and you're wondering why another resto druid in your guild or another guild has a much higher HPS than you do. It's easy to check by doing this kind of comparison. Check the crit rates of your individual abilities, check how many times rejuvs/wild growths have ticked. Another important thing is checking the usage ratios between abilities. Say for example a druid is looking at another druid who has a 1:1 ratio of rejuvs to lifeblooms cast! =O That's obviously inefficient and probably not a great way to spend mana. Maybe you're wondering why your raid healing isn't as effective as someone elses raid healing. You can then check their ratio of rejuvs:wild growths. Ideally if you were just focused strictly on raid healing, your ratio of rejuvs:wildgrowths would be around 5:1. 5x rejuv 1x wild growth, and repeat. If you see that a druid's ratio of rejuvs:wildgrowth is sitting in the area of maybe 2:1 or 3:1, it's easy to infer that the druid isn't using rejuv enough. While this is valuable information, some stats can be misleading. Perhaps that same healer was not strictly focused on raid healing, but was also dropping hots on tanks or if they were in a fight where they needed to perform some other role than just blanketing the raid with heals. But it could also signal a deficiency in that healer. If that healer was SUPPOSED to have been raid healing, and was casting a lot of healing touches or rolling lifeblooms instead, you can infer that they weren't focused on their assigned role. I'm only using druids as an example because I probably know them the best out of all the healing classes that I've played.

As you can see, there are piles and piles of information that you can drag out of world of logs if you really need to. The information gleaned can serve many purposes: auditing yourself and others, and trying to find deficiencies in your own performance. What's really needed is an open mind and an openness to constructive criticism. An open dialog between our healers, between our dps, and between our tanks can serve much good. Don't be afraid to look at logs and ask questions of your guildmates, and don't get offended when you get asked a question. It serves to help the asker and the askee understand each other better.

The next two lessons will probably cover talents and gear. I haven't decided which is next yet, but I'm leaning towards talents. Here is a preview of the next discussion topics!

It's "good enough" to just find a talent spec and put points in it and never think about it again. But this thread isn't about being 'good enough', it's about trying to squeeze out as much DPS as possible. You have to understand talent interactions and why you're picking one talent over another.

It's also "good enough" to to use maxdps or best in slot lists to pick gear. But one must realize that best in slot lists are ONLY best in slot in relation to all the other pieces in that list. In isolation, best in slot pieces aren't really best in slot at all. The best item in that slot can only be determinable in relation to what else you're wearing currently.

Alright, see you next time!

[Daewen]

I was looking at our latest ICC25 just for myself (I recently respec'd a few points around, and assessment says I need to re-gem/enchant pieces), but I noticed something which should be encouraging. On Rotface try 2, until the first slime-splosion when dps scattered from the boss, the whole raid was doing ~120K dps, which is the amount to reach 36M in 5min. But after dps regrouped, it didn't get back up to that mark (I blame the slimes not being in the mood to mate), and the 2nd slime-splosion came as dps was going up again. After that, it just all trailed off.

Though overall for the fight we wouldn't beat the timers, for a time we were doing... well, the bare minimum required. So we still need to bring dps up to where steady dps matches the current burst damage.