The pitfalls (and benefits) of DPS meters

Now THIS was a close DPS spread.

DPS meters are such a polarizing topic. Are they good? Are they evil? You'll find no shortage of support for both camps. Of course, the correct answer (as is usually the case) is that they are neither. They're definitely useful, but they are also only a small part of a larger, more important picture. Are they necessary? Nope, and if you're not playing the game at an advanced level, they're only of minor benefit. But I firmly believe that they are an essential tool for raiding DPS who want to perform at their peak capabilities.

One of the most common complaints about DPS meter addons like Recount (there are others, such as Skada, but for simplicity I'm just going to use Recount for discussion) is that players get distracted by the ever-changing meters during a fight, diverting their attention and causing them to make mistakes. Another issue is that players might become obsessed with their Recount total and therefore neglect other essential fight duties, such as target-swapping, interrupting or moving out of fire, just to keep their DPS numbers high.

And you know what? Sure, these things happen. But this post is not aimed at new or inexperienced players who would make these rookie mistakes. I'm talking about DPS players at a raiding level, where it's absolutely a given that people will properly handle fight mechanics and perform their individual duties in addition to maximizing their damage.

And really, DPS meters are of limited (if any) benefit DURING a fight. Their true value comes afterward, examining the results to see how you did, and using the data to improve. It's like a math test - simply knowing if you passed or failed doesn't really help you at all. What helps is knowing what you got right, and what you got wrong/need to improve on.

A disclaimer about special circumstances
Before we even begin to look at this topic, it's important to realize one thing. Damage meters of any kind are absolutely useless without also considering unusual circumstances. On a Patchwerk tank-and-spank fight? Great, there's nothing out of the ordinary to consider. But on most fights, you need to ask yourself: "Was my damage particularly low/high because I was doing X? Was this other player's damage particularly low/high because they were doing Y?"

A perfect example of this is Firelands' Alysrazor, where a few DPS players fly up in the air and attack Alysrazor with a mega damage buff, while the other players kill adds on the ground. Here is a simple fact - if you are on the ground, your end damage will be far, far lower than the players flying. This is just how the fight works. And consequently, it's utterly useless to compare a ground player's DPS with a flying player's. The flying player shouldn't be happy he "beat" the ground player, nor should the ground player be sad that his DPS was so much lower. Context is everything.

It's also important to realize when Recount or damage logs simply aren't accurate, due to fight mechanics hiding essential data. For example, Recount/logs for Gara'jal in Mogu'shan Vaults are woefully inaccurate due to the constant phasing. Each player's individual Recount/combat log will be accurate for their own numbers and wildly off for everyone else.

Other examples of specific fight mechanics/duties playing havoc with DPS meters include steering Lord Rhyolith, the incredible melee cleave buff on Garalon, Amber-Shaper Un'sok's vehicle transformation, fights where a player extensively kites something, fights where some players are phased or teleported to other areas, etc.

Recount and topping the meters

One of the biggest flaws with Recount lies not with the addon itself, but with how we players interpret its data. After all, the most common reaction upon looking at Recount is "How did I do?" or "Was I the highest damage?" In other words, we look at the numbers in a competitive, me-versus-the-world light.

This a tremendously flawed way of thinking.

The problem with viewing Recount totals to see how you did against the other DPS is that it's almost pointless to compare yourself to other specs, let alone classes. In my opinion, there are two major reasons why you don't want to do this.
1) Different specs/classes interact differently with different fights/mechanics
2) Different specs/classes are mathematically capable of different damage outputs; higher damage does not necessarily mean better performance

Different specs/classes interact differently
with different fights/mechanics

If everything was equal in WoW, then yes, you could compare your Combat Rogue to your friend's Elemental Shaman and see which of you did better based on whose damage was higher. But as we know, things in WoW are NOT equal, and thank goodness for this, because otherwise things would be REALLY boring.

Comparing the damage totals of two different specs or classes is rarely worthwhile, since there are so many influential factors you need to consider. Let's look at a melee vs. ranged, for example. First of all, was there special circumstances (as described above) that automatically invalidates any sort of comparison? Assuming not, was the fight friendlier for melee or ranged? (ie, did either group have to constantly move out of fire while the other group just stood still and DPS'd?) Was there a special buff that melee or ranged received? Were there new targets during the fight that melee had to spend time running to? Were there periods of time in the fight when ranged DPS could hit the boss, but melee DPS could not?

As you can see, most non-Patchwerk fights treat melee and ranged DPS differently, which makes comparisons largely a waste of time. But what about comparing two ranged? Or two melee? Hell, what about two of the same class, but with different specs? Do comparisons work then?

The answer is: it depends. It depends on the capabilities of the specs being examined, and it depends on the fight mechanics. Some specs have significant advantages on fights that allow multi-DoTting, while other specs are amazing at fights with swarms of adds due to their powerful AoE abilities. Meanwhile, other specs love multiple-target fights because they cleave as part of their normal rotation, while other specs love single-target fights because their DPS rotations require significant ramp-up time.

To summarize, comparing the damage of a class and/or spec with that of a different class/spec can be interesting, but it rarely provides information that is actually valuable or truly meaningful for analysis purposes.

Different specs/classes are mathematically
of different damage outputs

We've just looked at how varying fight mechanics and class capabilities can make a huge difference when examining damage outputs. Well, what about when these conditions are removed from the picture, and all things are equal? What happens then?

Like always, Patchwerk is a perfect example of this - there are no special melee or ranged buffs, there's no adds, there's no other targets to DoT up, etc. So if player X does more damage than player Y on Patchwerk, does this mean they played better?

Not necessarily.

One obvious difference could be gear. Better gear = better potential DPS. I don't think this really needs to be explained or examined.

But gear aside, higher damage still does not always mean better performance. Some classes and/or specs are simply capable of more DPS than others, even when played at equal levels of skill. This should come as no surprise to WoW raiders - the constant fine-tuning and tinkering with class variables is a never-ending process. Knowing this, players should take individual class/spec potential into consideration when comparing their damage numbers.

For example, let's say Player 1 does 93k on a fight, and is disappointed to see that his same-class guildie Player 2 did 98k. A full 5k DPS more! Clearly, Player 1 got outplayed, right?

Well, hold on a minute now. What if you now learned that Player 1 is an Arms Warrior, and Player 2 is a Fury DW 1H Warrior? Take a look at this simulated "maximum theoretical DPS" chart from WoW info site Noxxic:

(Please note, I am NOT saying that Noxxic's chart is necessarily the best one out there, or that these numbers are even 100% accurate. I have no idea, and honestly, the specific numbers are actually not important. The chart and DPS spread is here to make a point, using hypothetical data, and this sample range is as good as any.)

Taking a look at the chart, we see that Player 1's 93k Arms DPS was great! In fact, it was the (theoretical) maximum possible DPS he could have done. He shouldn't feel bad about doing less damage than Player 2 at all.

In fact, when we look at Player 2's DPS, we can see that his 98k Fury DW 1H DPS was actually below his spec's maximum possible output of 106k. So Player 1 actually performed at a HIGHER level than Player 2.

Now don't get me wrong, Player 2's DPS was still very solid - 92% of theoretical maximum is pretty damn good! But the point is, Player 1 performed PERFECTLY and Player 2 did not, but if you just looked at the damage meters without the context of Noxxic's chart, you'd never realize this.

Why is this important? Because both players in this example should be aware of the truth behind the numbers. Player 1 might beat himself up and slave away at target dummies for hours on end, thinking he needs to improve to reach Player 2's level...when the truth is, Player 1 doesn't NEED to practice, because he's already performing at 100% efficiency. Meanwhile, Player 2 might see that he's on top of Recount, and smugly assume that he doesn't need to practice because he's already the best. When in reality, he could still improve.

This latter mentality is a very common flaw in thinking, one that serious DPS players need to avoid like the plague. Being "on top" doesn't mean ANYTHING, because it doesn't accurately tell you how well you are performing. If someone from Paragon or Method goes and pugs Heart of Fear with a bunch of casual players who have never raided before, does the hardcore raider's position at #1 on Recount really say anything? Of course not! Using the math test analogy again, what makes more sense? Getting a 60% on your math exam and being happy because it was the best mark in the class? Or getting a 95% and not caring what your classmates got, because really, who cares?

Now obviously, it's fun to compete with your friends and laugh about beating your buddies by 17 DPS or whatever. But when it comes time for serious analysis, ignore the others and focus only on your own numbers. The only people you can truly measure your DPS against are yourself.

Or other people playing the exact class/spec/spec specialization as you, of course. Wouldn't it be nice if there were a convenient way to look at such data?

Analyzing your performance with World of Logs

Most raiding guilds upload their combat logs to World of Logs, which is an incredibly useful tool for analysis. You can monitor buff uptimes, proc rates, how well you lined up/used certain cooldowns, etc. The possibilities are endless.

However, what we're going to look at is WoL's ranking system, which doubles as a quick and simple way of checking one's performance. World of Logs keeps records of when players "rank", which is when they are among the top 200 players in the entire world for their specific class and spec, for each individual fight. For example, at the time of this writing, the world's 200th best Stone Guard Frost DK did 104,069 DPS. If I were to upload a log where my Frost DK Morgion did 104,070 DPS, then Morgion would replace the former #200 ranked player and 104,070 DPS would become the new ranking threshold.

Obviously, trying to rank is a great goal. If you're within the world's top 200 players for your class/spec on a fight, you're clearly doing things right!

But where WoL's ranking data is truly useful is its percentage-based performance scores. You see, the #200 rank is given a performance score of 100%. You have to score 100% to rank, which makes sense. And unlike Recount, WoL rankings only consider other players of your exact class and spec as competition. Remember all those variables that can skew comparisons we discussed earlier? WoL rankings take those and throws them out the window, creating a (mostly) standardized basis for comparison.

(There are some exceptions, of course. Even if you're being compared to the same class/specs, there can still be unusual circumstances that invalidate comparisons, such as flying on Alysrazor, or significantly different boss strats.)

On one of our Protectors kills, our guild's mage agonized about how close
he had come to ranking. He knew he could do better...

You might be thinking, "this is all epeen! Boo!" And sure, you could certainly interpret ranking data and being in the top 200 as such. But I don't think it's arrogance or epeen to seek an accurate summary of how well a player is doing, whether that player is you or someone else in your raid team. The closer your percentage is to 100%, the better you are performing!

How close to 100% should be expected? Well, your expectations will vary depending on your guild and level of progression. For example, I expect elite hardcore guilds would probably expect more than 100%/#200 rankings from their players. But for the rest of us normal raiders, you might be comfortable with 40%, or maybe 70%, or even 90%! It's up to you to look at that 100% ranking threshold and figure out for yourself how high you should be aiming. Because no matter what you choose, if you find your percentages lower than what you were expecting, it's time to read up on your class/spec & figure out what isn't working. he made some adjustments, practiced, and next time ended up ranking.

In addition to personal self-evaluation, performance percentages also provide valuable stats for your raid team as a whole. When a raid team is struggling to beat a boss' enrage timer, you want to make sure all your DPS are living up to their potential. A raider doing 60k which is 90% towards ranking is doing a good job at maximizing their performance, while a raider doing 60k which is 50% towards ranking has a lot of room for improvement.

Now, REMEMBER, 60k is still 60k, and this sort of intense scrutiny is only required for top-end guilds OR when your raid team is desperately trying to meet a DPS check. If bosses are dying, then oh well, it doesn't really matter that one player is at 90% and the other is at 50%. But when you're hitting the enrage timer when a boss is still at 20% health, it's time to take a close eye at your DPS team and see where there's room for possible improvement. Looking at those two 60k DPS players, if you didn't know what their potential DPS limits were, how would your raid leader know which one needed to pick their game up? But after looking at World of Logs, they can now point out that 50% percentile and ask that person to improve.

Now mind you, if everyone is at 30-40%, the whole team might want to hit the forums/blogs and do some research...or the raid leader might have to reach the always-unpleasant-but-sometimes-necessary realization that a boss' DPS check is simply out of reach for the team at this point in time. Such decisions are never fun, but it's more useful than beating your heads against a wall endlessly when there's no hope of success right now.

How to improve

Eliminating the "winning Recount" mindset is the most important first step to self-improvement. Only after you've realized what your DPS numbers mean (and what they don't mean) can you really begin to understand and examine your raid performance. If you're not happy with your DPS, it's crucial to look at the contributing factors before you decide that you're terrible or that you need tons of practice. Gear, specs, strats, who you're comparing yourself need to consider all of these things.

And you know what? If everything IS equal, if your competition is the same spec, equally geared, and just outperforming you? Great! Find them on World of Logs, look them up on the Armory to see how they're reforged, throw some logs into Comparebot, and LOOK at why they are getting better results. Don't just mindlessly smack the dummies hoping that it will magically solve everything - analyze their superior numbers, look for what you are doing differently/not as well, and LEARN from this information.

Don't just give up and think "My DPS is low because I suck!" That is useless. It's vague and doesn't give you anything to try to improve, and it doesn't let others help you since you haven't told them anything useful. Instead, you should try for "My DPS is low because my DoT uptime is lower than it should be and I didn't cast this spell enough and I only used this cooldown once the entire fight and I forgot to refresh this buff when it fell off." Now you have clear goals to work on, and your guildies and friends can offer assistance/suggestions to help you reach said goals.

(And if you ARE happy with your DPS, you should still take a close look and make sure you're doing as well as you think you are...and if there's not something that you can still do better.)

The most important tool for improvement is not your gear, a fancy 27-button mouse, or an OP flavor-of-the-month spec.

It's knowledge.

4 Responses Subscribe to comments

  1. gravatar

    Just like duck tape and chainsaws, dps meters are wonderful tools when used properly, and goddam messes when used wrong. I'd say that 90% of the use I've gotten from them was to compare how I did after changing rotation / add-on / spec / etc versus how I was doing before. And the other 10% was to determine how hard to weep this week after seeing how much better Algie did this week. but that's another story.

    I'll never forget the time I was running a Wrath heroic and one guy started tearing into another party member for his terribad dps. "We should kick him, he sucks so much!" "Dude, you're looking at the healer."

    March 29, 2013 at 2:33 PM

  2. gravatar

    You know, if you flip that theoretical chart upside down, it'd look like a PvP chart.

    Just sayin'....

    March 29, 2013 at 3:34 PM

  3. gravatar

    Great read, I couldn't agree more.

    The last guild I was in broke up as a result of all our war between the members and the officers over the statistics being produced from dps meters. It was truly an unpleasant experience and goes without saying dps meters can be THE PROBLEM instead of the solution, especially when theirs a lot of ego lurking around.

    March 30, 2013 at 8:41 PM

  4. gravatar

    This is all stuff I personally know already, but this is absolutely perfect for linking to people when they say "ZOMG WOW'S FOCUS ON DAMAGE IS STUPID."

    April 4, 2013 at 10:11 PM