Wednesday, March 5, 2008

Damage Meters Lie.

Anyone who has compared the results from a Damage Meter from two different players has likely noticed that they don’t tell the same story. How do you tell which one is right? You don’t. They both tell “a story” but neither tells the whole story. This post is going to cover why Damage Meters don’t tell the truth and is inspired by all the ego whores who like to output a damage report whenever they top the chart.

Combat Log
All damage meters use your Combat Log to provide the majority of the data it records. There are a few problems with this method:
  • Log Range: Your log thinks of you as the center of it’s universe. If someone is outside of the maximum range of the log, it’s as if they don’t exist to the meter. This is why two meters rarely ever show an identical amount. Some addons will extend the range (which helps), but it’s still not possible to extend it far enough to be able to capture all data, all the time. Data synching can help with this problem, but that leaves the door open for other inaccuracies. The net effect is that you never have the entire picture.

  • Death: This is related to Range, but it’s worth pointing out as it’s own thing. When you are running back as a ghost, you are out of range and collect no data from your log. More likely however, is that you won’t release. Your log will stick pick up data, but it’s centered on your corpse. Often, a fight will move away from your corpse and therefore out of your log.

Over-healing and Over-damage
Over-healing is much better understood, so I’ll start with it. The basic idea is that a player has 9000 of 10,000 HP and you crit a big heal for 3000 HP. The effective healing is only 1000 HP and you over-healed for 2000 HP. The total healing done is noted as 3000 HP, but you only really helped for 1000HP. The good news is that most meters can be set to make a comparison to the player health at the time of the heal and track the difference as effective healing. However, it’s still not an accurate picture because it doesn’t show the whole picture. For example, let’s say a Druid drops a heal over time at the 9000 mark. It gets 1 tick at +300 and then the Greater Heal trumps it for +3000. The meter will record +300 for the HoT and +700 effective healing for the Greater Heal. However, in terms of overall healing effectiveness for mana spent, you can easily argue that the big heal was wasteful.

The same logic is also applied to over-damage, but the big difference here is that no addon tracks effective damage dealt by making the comparison against the health remaining of the mob. Why is that important? Consider this – a Warrior unleashes a 4000 crit execute on a trash mob with 200 health. The over-damage is +3800 damage. Yet, the damage meter records all 4000 instead of just 200. Now let’s say the fight lasted 40 seconds and you also had a non-crit execute with +1500 damage on another mob. Your DPS for the fight would reflect 132 DPS higher than the actual DPS you contributed. When I played a Fury warrior, I was constantly frustrated that I never really knew my own personal DPS even when soloing simply due to the sheer amount of over-damage caused by my executes when the mob dropped below 20%. Think of it this way, it’s very easy to deal 8000 damage to a mob that has 5800 HP. If you killed it in 13 seconds, that is the difference between 615 DPS and 446 DPS.

Over-damage alone is reason enough to not trust damage meters.

Critter kills
Linked to Over-damage, a guaranteed crit for 1500 on a 100 HP critter is an easy way to inch your damage number up by +1400. I’m willing to bet the same guy killing all the snakes is the same one who asks for a meter report later. Trust me. He already has a meter. The people who ask for a meter report are the usually the ones who already know they are at the top of the meter.

Damage per second
Damage per second is calculated by the amount of damage you have dealt divided by the amount of time YOU have spent in combat. The problem of course, is when you end up in combat and your role is to do nothing. The mage who crits his Pryoblast to start the fight is oddly rewarded over the player who entered combat but exercised caution early on to allow the tank to build up threat.

Boss fights vs. AoE clears on non-elites
The amount of over-damage created by AoE spam can be incredible. AoE clearing is certainly useful, but as a measure of damage contribution it is pretty heavily skewed. In most cases, the same goal could have been accomplished without the AoE just over a longer period of time. You could argue that all the drinking afterwards should be included in the DPS calculation since most AoE clears include an intermission for refreshments. Big props to all the mages that have died on these pulls for the wonderful self sacrifice. You deserve to top the meter if you die.

Boss fights, on the other hand, can vary greatly depending on the encounter. For example, an endurance encounter may favor the melee types who don’t have a mana pool to worry about. Likewise, a boss who spams a bunch of AOE might keep the melee types out of range while range types take down the boss. Also consider, how much less healing was need due to that Hunter’s nature buff or Shaman’s totem?

Group Contribution vs. Damage or Healing done
The biggest single issue with any meter is that it simply reflects the output recorded by the log. It does NOT provide an accurate picture of all the things that contribute to the actual output.

For example, what effect did that Feral druid bring to the group with his Mark of the Wild and Leader of the Pack (5% to crit)? How much extra healing did the Mage bring with his Arcane Brilliance? How much extra DPS did the Warrior bring the other melee with 5x Sunder Armor and the Battle Shout (+300 AP)? What is Blessing of the Kings worth? How many extra nukes did the mage cast safely because of Blessing of Salvation. How much of the extra damage from a Rogue’s Hemo debuff went to the other melee (including the other Rogues to which they are being compared)? How much healing did that crowd control PREVENT?

None of these things is measured in a damage meter. We like damage meters because they simplify things down to this person contributed X and this person Y. Unfortunately, it’s not quite that simple. The actual contribution varies greatly based on the group configuration and encounter.

So what are damage/healing meters good for?
Personal measurement. They aren’t the bible, but you can use them as a gauge for your own personal betterment. You can also see if you are roughly in the same league with other people in your group or use it to learn how much another spec or class complements your output. It’s not gospel, but if you use it as a tool, it can be helpful. The exception is if you are using a threat meter. These are helpful as a general guideline. Again, not the gospel, but they do act as pretty immediate feedback on whether you need to slow down your threat. Along those lines, monitoring your aggro duration can be a good measurement of personal skill as well.

Using a meter correctly – Reset It Often
Your meter is really a measurement over time. If what your situation changes, then it stops being relevant or as useful. For example, if you were grinding out boars for an hour before Wailing Caverns and then didn’t reset your Damage meter after entering the instance with your group, then your meter will provide pretty meaningless information about WC. Typically speaking, I recommend resetting the meter whenever you join a group, enter an instance, gain a new buff, and finally both before and after boss fights. This last is important because trash fights are considerably different than boss encounters. On the other hand, don’t reset too often because it only really works as a tool if it is allowed to average itself out.

1 comment:

Terry said...

Very true. The only time I get dmg meter results posted in the chat channels is when the poster is the top dps listed.

You certainly can't argue that dps is an important aspect of every fight but I feel 'how' you played is more important than your dps imho. How did u react when the boss debuff'd all the aggro? How did u handle the unexpected add in the fight? etc.

Like you point out, there are simply too many parameters in any fight to have a simple log reading meter display 100% truth!