Current State of Mages: Statistical Round-Up

[Haruichi]

I’ve compiled a few stats from Warcraft Logs that are indicative of the current state of mages.

Stats are drawn from 24 parsed dps specs, 80th percentile, over two weeks of raids, all regions.

Mythic

All bosses
Fire Mage: 22nd
Frost Mage: 17th
Arcane Mage: 21st

Scorpyron
Fire Mage: 2nd
Frost Mage: 16th
Arcane Mage: 8th

Cronomatic Anomaly
Fire Mage: 22nd
Frost Mage: 15th
Arcane Mage: 13th

Trilliax
Fire Mage: 20th
Frost Mage: 14th
Arcane Mage: 19th

Spellblade Aluriel
Fire Mage: 11th
Frost Mage: 13th
Arcane Mage: 5th

Tichondrius
Fire Mage: 1st
Frost Mage: 19th
Arcane Mage: 12th

Star Augur Etraeus
Fire Mage: 12th
Frost Mage: 5th
Arcane Mage: 18th

Krosus
Fire Mage: 21st
Frost Mage: 10th
Arcane Mage: 17th

High Botanist Tel’arn
Fire Mage: 23rd
Frost Mage: 13th
Arcane Mage: 22nd

Grand Magistrix Elisande
Fire Mage: 20th
Frost Mage: 17th
Arcane Mage: 23rd

Gul’dan
Not enough parses

Heroic

All bosses
Fire Mage: 19th
Frost Mage: 16th
Arcane Mage: 14th

Scorpyron
Fire Mage: 2nd
Frost Mage: 19th
Arcane Mage: 12th

Cronomatic Anomaly
Fire Mage: 16th
Frost Mage: 13th
Arcane Mage: 9th

Trilliax
Fire Mage: 19th
Frost Mage: 12th
Arcane Mage: 15th

Spellblade Aluriel
Fire Mage: 14th
Frost Mage: 21st
Arcane Mage: 1st

Tichondrius
Fire Mage: 3rd
Frost Mage: 18th
Arcane Mage: 1st

Star Augur Etraeus
Fire Mage: 19th
Frost Mage: 5th
Arcane Mage: 14th

Krosus
Fire Mage: 23rd
Frost Mage: 12th
Arcane Mage: 15th

High Botanist Tel’arn
Fire Mage: 16th
Frost Mage: 17th
Arcane Mage: 15th

Grand Magistrix Elisande
Fire Mage: 17th
Frost Mage: 5th
Arcane Mage: 24th

Gul’dan
Fire Mage: 22nd
Frost Mage: 13th
Arcane Mage: 23rd

Normal

All bosses
Fire Mage: 12th
Frost Mage: 13th
Arcane Mage: 17th

Scorpyron
Fire Mage: 2nd
Frost Mage: 23rd
Arcane Mage: 12th

Cronomatic Anomaly
Fire Mage: 14th
Frost Mage: 11th
Arcane Mage: 13th

Trilliax
Fire Mage: 13th
Frost Mage: 9th
Arcane Mage: 14th

Spellblade Aluriel
Fire Mage: 7th
Frost Mage: 21st
Arcane Mage: 3rd

Tichondrius
Fire Mage: 2nd
Frost Mage: 20th
Arcane Mage: 3rd

Star Augur Etraeus
Fire Mage: 15th
Frost Mage: 8th
Arcane Mage: 16th

Krosus
Fire Mage: 16th
Frost Mage: 10th
Arcane Mage: 15th

High Botanist Tel’arn
Fire Mage: 14th
Frost Mage: 13th
Arcane Mage: 18th

Grand Magistrix Elisande
Fire Mage: 13th
Frost Mage: 5th
Arcane Mage: 23rd

Gul’dan
Fire Mage: 17th
Frost Mage: 10th
Arcane Mage: 24th

Some take-aways

• 29 encounter parses, 3 aggregate parses and 87 spec specific parses spanning 3 difficulty levels.

• In all of the aggregate “All Bosses” parses, all 3 mage specs are in the bottom half with one exception: Normal difficulty where Fire Mages came in at 12th out of 24.

• No other class has all their parses in the bottom half across all bosses and all difficulty levels.

• Mages are the only class that does not have a Top 10 place across all bosses and difficulty levels.

• In 8 of the 29 encounter parses, mages have a Top 4 parse (Top 4 singled out since there are 4 pure dps specs).

• In 12 of the 29 encounter parses, mages have a Bottom 4 parse.

• Mages have 3 1st places: Mythic Tichondrius (Fire), Heroic Tichondrius (Arcane) and Heroic Spellblade Aluriel (Arcane).

• Mages have 2 24th places: Heroic Grand Magistrix Elisande (Arcane) and Normal Gul’dan (Arcane).

• The most common mage rank in encounters is 13th (9 spec parses) followed by 12th, 14th, 15th and 23rd (6 spec parses).

• Of the 87 mage spec parses, 10 are in the top 3. They occur on Tichondrius, Scorpyron and Spellblade Aluriel.

• Of the 87 mage spec parses, 11 are in the bottom 3. They occur on Scorpyron, Chronomatic Anomaly, High Botanist Tel’arn, Krosus, Grand Magistrix Elisande and Gul’dan.

• Mages have 31 spec parses in the top half and 56 spec parses in the bottom half of the encounter parses.

• The encounter parses are distributed in the following way:

• Top 25%(1st-6th place): 15 spec parses;

• 25%-50% (7th-12th place): 16 spec parses;

• 50%-75% (13th–18th place): 34 spec parses

• Bottom 25% (19th – 24th place): 22 spec parses

[Makz]

I wish we still had hybrid tax concept. Yes, it is an outdated and "unfair" concept, but if they cant guarantee single purpose classes are top, it is just sad. Why would you roll a mage right now - yes some of the utility is fun like scorch sprint, cauterize, but you can barely make it to top10.

[Shitbird]

Yet somehow blizzard thinks mages are in a good spot

[Causese]

well you need to take progress/bosskill-oriented playing into account, can't just look at overall dps.

chronomatic anomaly mythic:
as arcane you can delay your cooldowns until first add spawns (+ time runs faster there) to almost kill it just by yourself. It causes your group to have more uptime on boss and to stack faster at boss for power overwhelming.
example: https://www.warcraftlogs.com/reports/1h ... arget=67.1" target="_blank

guldan mythic:
arcane is really good at dealing with eyes of guldan (given that you know how to properly delay cds and save mana for it) , that's most likely why from scratched used an arcane mage for guldan mythic firstkill

then there's bosses where bossdps should be valued higher where frost shines at times.
there are more examples but I'm too laze to write/look it up

the only annoying thing imo is that we don't have all 3 speccs available (unless we put tons of effort into it)

there are so many things you need to take into account to figure out if a specc is good, like

- damage done during pure dps-check encounters
- damage done in critical boss phases (damage dealt to solarist @botanist mythic for example)
- damage done to adds that truly need to die asap (tichondrius watcher-add or what its called and guldan eyes..)

[Haruichi]

Your post is interesting in a number of ways.

there are so many things you need to take into account to figure out if a specc is good

If you re-read my post, you'll notice that I've included no analytical conclusions about the state of mages. I've simply presented numbers. You indicate that those numbers make mages look bad which, I guess, is telling in and of itself.

chronomatic anomaly mythic:
as arcane you can delay your cooldowns until first add spawns (+ time runs faster there) to almost kill it just by yourself. It causes your group to have more uptime on boss and to stack faster at boss for power overwhelming.
example: https://www.warcraftlogs.com/reports/1h" target="_blank ... arget=67.1

You present anecdotal evidence to counter statistical evidence. That rarely works out well. That being said: What's your ambition here? To demonstrate that Arcane Mages are able to do more damage to the first add that spawns if they save their cooldowns for it while the rest of the raid has blown their cooldowns on the pull?

guldan mythic:
arcane is really good at dealing with eyes of guldan (given that you know how to properly delay cds and save mana for it) , that's most likely why from scratched used an arcane mage for guldan mythic firstkill

Again, I'm unsure what you're trying to accomplish. You single out a situation where you think Arcane Mages are good. That's fine. Even the statistical data will account for that since there are a few fights where mages are placed 1st. However, the point of presenting a vast amount of statistical data is to paint a broadswept picture that, for better and for worse, is unaffected by singular events like you describe. It's a good thing that Arcane Mages are good at dealing with Empowered Eyes and adds spawning on CA, however, on a larger scale that is the collective Nighthold raiding scene, the presented stats are indicative of mage performances.

- damage done during pure dps-check encounters

Are mages doing particularly well in pure dps-check encounters in Nighthold?

- damage done in critical boss phases (damage dealt to solarist @botanist mythic for example)
- damage done to adds that truly need to die asap (tichondrius watcher-add or what its called and guldan eyes..)

I get that, especially Arcane Mages, handles priority adds and burst phases well. Is your point, to force it, that if mages had an ability dealing floppityjillion damage on a 3 minute cooldown so that they could pretty much single handedly kill off that Add X that needs to die really fast - but on the encounter as a whole, and other encounters as a whole, mages would always be average or below average - then that would be satisfactory?

[Causese]

well as you can see people already interpret it as mage being bad

message was just "try to give less fucks about logs based on overall dmg"

my post was also a reply to Shitbird, should've used quote function

all 3 mage speccs obviously meet the requirement for dps-check encounters but only frost has a chance to get a top parse in pure singletarget dps encounters.

https://gyazo.com/72b0db4209d3d47306fa88d27a33c0c0" target="_blank
you just need that rng and think fast once you have it

[Andrestes]

message was just "try to give less fucks about logs based on overall dmg"

lol +1

(not to trivialize the OP. it's always tough looking at damage meters and not being on top. especially when you have these warcraftlog stats that pretty much say it's easier to pull better numbers with most other classes)

[Haruichi]

message was just "try to give less fucks about logs based on overall dmg"

This, to me, is actually a very worhtwhile discussion. And the following is not so much that I disagree with you, but rather to instigate a discussion:

How much damage you do on a boss fight is a benchmark. For better and for worse, and a benchmark often used to analyze the topography and performance levels of a class compared to other classes. You present the argument to worry less about that particular benchmark. Why is that not a meaningful benchmark? I'll outline some opposing views:

1. Even if it sometimes hides singular events, like handling a certain type of add in a certain encounter, it's still a reasonably good KPI. After all, we're a dps class and a large part of our job is to do as much damage as we can.

2. For better and for worse, it's an established method of evaluating how you perform. When you set up Skada or look at logs, you can, of course, go into how certain adds or phases were handled. That will, however, only serve as small segments in the backdrop of overall damage done to a boss. I.e. it's hard to argue that you've done well on a fight if you've done solid damage to an add, but came in 15th on the boss fight as a whole, while the opposite is rarely the case.

3. Competitiveness. It's often mistaken for vanity. And for good reason: It's about defeating the boss encounter and not about who's top of the meter. That said, people play their class for a number of reasons. And, I suspect - can't claim to know for sure, of course - that most dps players, especially those who play their class in a way that warrants reading stuff on Altered Time and similar high end class ressources, are competitive. Part of why they play their class is actually to squeeze as much dps out of it as they possibly can.

As such, provided encounter mechanics are properly dealt with, damage done becomes a hallmark of you having played your class well; a source of gratification that caters to that competitiveness and, in some ways, is a part of why you pay money each month to play this game. And I posit that the same gratification isn't always achieved by killing off a certain add quickly, as an example.

Now, I acutally do agree with you that much more emphasis could be put on singular events within the respective encounters. However, I think it's reasonable to acknowledge that "overall damage" has considerable merit as a benchmark; even if it's doesn't tell the whole story.

[TheGreyPaladin]

I agree this is a worthwhile discussion. While meters are not all that matters I guarantee a lot of people look at them and judge your performance. Most players are not going to say oh its easy to rack up DPS as a frost DK and hard as a mage they are just going to see your lower numbers and think you are bad. Loot council is not going to favor you.
I cant remember in the last decade mages being this low in terms of performance and rankings. (Even during the no mages all warlocks time) Its VERY disheartening to do your absolute best and put in tons of effort only to be lower middle of the pack on most fights. And as far as I can tell this is where we will stay there are no threads of incoming buffs for mages for 7.2 so I see us staying towards the lower half. I am fire ATM since I have one of the (2) BiS (belt) and in no way shape of form am I motivated to switch to frost, regear, reget legendaries to again be middle lower of the pack, although maybey a little higher. ;(

[Causese]

imagine a 4min30 fight

- arcane has 1min30 cooldowns that last 13s and the boss will die just the moment your 3rd cooldowns are available. at 4:30 your sustained dps has dropped so hard that u don't want to look in the mirror anymore

- boomkin has 3min cooldowns that last 30s? -> at 3min30 the boomie will start dropping in sustain

- fire mage has 2min cds. so at 4min you will use mirror images and at 4min10 you combust (since u want to stack pyretic incantation prior to combustion?), at 4:20 you're done bursting and maybe you get a lucky wrist procc in the last 10s so that your sustain doesn't drop

guess who will be on top?

that's one reason for logs/dps meters "not being a meaningful benchmark".
you'd have to find a fight duration that suits everyone which isn't possible I guess?

and the fight duration is also variable, some top parses can only be made in bad/good groups

[Haruichi]

imagine a 4min30 fight...

guess who will be on top?

that's one reason for logs/dps meters "not being a meaningful benchmark".

First, you've switched argument so that now it's logs and dps meters as tools that don't provide meaningful benchmarks. What was discussed before was overall damage. And again you try to disregard statistical evidence with an anecdote. This is an imagined anecdote of a 4.5 minute fight, but still an anecdote. Just to try it out, I looked at the first log on Warcraft Logs I could find. It was a Heroic Trilliax kill that lasted 3.30. That should fit Arcane Mages much better since they can squeeze in and finish off with a third burn phase and, thus, your anecdotal evidence is countered - with another anecdote. It's nonsensical.

To that end, you provide the very reason why logs and overall damage *are* meaningful. Because anecdotes will only get you a short way when the purpose is to make interclass comparison. The point of considering logs from thousand of parses is decidedly that variables like fight length are, to a point, ironed out. If you include enough parses - like every parse, in every region, for a duration of 14 days - your obtain such a massive amount of data that it will, for example, include plenty of examples of both fight lenghts that suits X class and spec and fight lengths that don't.

[infero91]

I absolutely love this post. The information is all there and your responses are a joy to read. Thank you.

[Drozzy]

Just food for thought. An interesting comparison that could be made would be considering the statistical averages (raw dps or placements compared to all dps specs on a fight) of the 3 mage specs at different percentiles. Could provide some useful data for players who are looking to see some of the pros and cons of each spec. It may provide some insight into things such as: the reliance on gear/legendaries that a spec has and the difficulty it takes to play a spec (and potentially finding percentiles where player skill stops having a large impact). I've looked into some of it myself in small doses, but it could be very useful data for players who are considering switching specs (I have run into many players recently who are looking to learn other specs for one reason or another).

[Katsumi]

There you go.

What you see here is the data from warcraftlogs.com, extracted on 10.03.2017 with the following settings:

- Player Damage Statistics For
- Star Augur Etraeus
- Heroic
- For all DPS
- By item level
- by percentile
- Over A Range Of 2 Weeks
- Current Standings

Mage DPS by percentile ilvl 895 - Augur Heroic

The DPS difference between Frost and Fire/Arcane becomes relatively larger at the right end of the distribution. Not only is frost stronger on this fight at all percentiles, the higher parses are also relatively further away from the other two specs.

Next, we have the same data with higher item levels.

Mage DPS by percentile ilvl 900 - Augur Heroic

Not much of a difference. There's a dent in the curve again around the 90th percentile.

Mage DPS by percentile ilvl 905 - Augur Heroic

The gap seems to be widening a little for even higher item levels.

The more important first conclusion, however, is this: With decent gear, an average frost mage will beat a top tier fire or arcane mage simply because of the imbalance between the specs. This effect becomes more pronounced the higher the item level is, of course assuming everything else to remain the same (we'll relax that in a bit)

I'm limited to three attachments per post, sorry for the spam.

[Katsumi]

However, if you decide to switch there's usually the problem of not having the approriate gear.

Mage DPS by ilvl at pp50 - Augur Heroic

At the median, dropping several item levels would still be worthwhile.

Investigating all item levels, does the difficulty of the encounter have an impact?

Mage DPS by percentile all ilvl - Augur Heroic

Mage DPS by percentile all ilvl - Augur Mythic

Apparently so. Aside from the pronounced gaps between the specs at all percentiles, there's also a very clear spike for the highest percentile. For the highest parses, where skill and luck comes together, frost not only outperforms fire and arcane, it does so by a disproportionately large margin.

[Katsumi]

As a side note, things are a bit different for more AOE-oriented fights.

Mage DPS by percentile all ilvl - Spellblade Heroic

Arcane seems to go the distance on Spellblade.

Mage DPS by percentile all ilvl - Botanist Heroic

Things seem surprisingly close on Botanist.

Mage DPS by percentile all ilvl - Skorpyron Heroic

As a fire mage, I really like the last picture. Admittedly, most of those parses are very likely deliberate chaos pulls.


To conclude:

1. Frost trashes the other specs on single target fights and it does so consistently.
2. Switching from fire or arcane to frost has a higher DPS payoff than trying to get better at playing fire.
3. Dropping several item levels in the process can be worthwhile, sim yourself.
4. There is some evidence for disproportionately large rewards of playing frost at the highest level.
5. For AOE-oriented fights other than Skorpyron, the specs are somewhat closer in damage across all percentiles.

Sometimes I wish I was equally productive at the things that I should be doing... I'm excited to hear your thoughts on this.


Thanks,

Katsu.

[Daidai]

fire is traded as weak in ST-fights, but shines on AOE. This topic dominates the comm.
It is clearly visible that also arcane is in a weak situtation. And further its AOE potential is extremly situational.

Arcane is not the best of both worlds as widely accepted. Overall arcane`s situation is even worse than fire.

[Dyalia]

Thanks,

Katsu.

Why whould you thank us? We should be thanking you

To the topic and to roung that up:

Fire:
+good for sustained aoe
-not so good ST

Arcane:
+good for priority adds
+less gear/legendary dependant
-not so good ST

Frost:
+very good ST
-not so good if there are RP/phases with no dmg events such as elisande/guldan/tich, because of IV uptime

All specs equally good on cleave fights.

[Mage]

Someone with gfx skill should make the 1st post into an infographic image.

[Katsumi]

Data extracted from warcraftlogs.com on 13.03.2017 with the following settings:

- Player Damage Statistics For
- Each boss
- Heroic (and Mythic)
- For all DPS
- For all item Levels
- By percentile, pp50 (and pp95)
- Over A Range Of 2 Weeks
- Current Standings

DPS rankings, pp50, all ilvl, heroic

DPS rankings, pp50, all ilvl, mythic

Keep in mind that the botanist fight changes dramatically on mythic.