Link to current update

Hey guys, I didn't really know what to call the topic, but I guess if you're reading this, I got your attention.

I like data mining and I'm currently trying to build some kind of database of Katalamize/Hyperion Runs to evaluate PvE DPS for different classes and setups. As a single person, I can only run these instances so often and I only have a small portion of the playerbase to take data from. For instance my server has many very good Gladiators, but it's lacking good Sorcerers and leather classes.

So my point is: I'd be very happy if there were people on here saving their Katalamize/Runadium-DPS sessions on ARM and willing to share these with me so I can add them to my list as to have a broader range of examples. (I really need some new hobbies while commuting to work I guess...)

Anyhow, the basic thing looks something like this (not an up-to-date-screenshot, though)





I do normalize the 'DPS' by just dividing the absolute Damage by the time the whole run took. Yes, this does include shields and adds and what-not and drops DPS compared to what ARM shows. However, I can answer how effective different classes are for certain instances overall. Assassins might be best DPS on a resting target, but in Runadium, they are not as effective. This has to be considered.

Looks something like this:

So if you could send me some stuff per PM or just in the thread... highly appreciated


PS:
And to answer the original question... For Katalamize it's Gladiators, hands down, for Runadium it's basically any ranged DPS, with the single 'best' DPS done by a DPS Cleric in the data I have.

PPS: Please add Server and faction to your informations please. It makes a difference for Rangers and SMs mostly.

I'm well aware of this and even mentioned the implications in my description. Thanks for reading.

Yes, it does benefit certain classes - that's because the instances make these classes benefit, not the other way around. This way, I can tell how effective a class can function for this particular instance. I don't care for three-minute-training-dummy damage. While good for training purposes, it's purely theoretical DPS as there's no similar real PvE encounter.

Sure, there are shields SM can DoT through, other classes can't to that. But why would I disregard the advantage this class has over a Gladiator, who will just be idling meanwhile?
Sure, certain classes take breaks to support in one way or another. However, if it's mandatory they do take these breaks, why would I act as if they never had to stop DPS'ing? (This mostly applies to DPS Clerics, SMs and Songweavers)
Also squishy classes will, on average, die sometimes and this makes their average DPS suffer. Can't disregard that.

I compared a lot of ARM DPS numbers with the ones I calculated - and nuke classes such as Sorcerers have a much higher difference between each of them. You all know why: If I only cast nukes on my sorcerer but little inbetween, of course my DPS is going to be high, but my overall Damage is going to suffer. This is why ARM DPS is utter bullnyerk in my opinion. I'd much rather have a realistic estimate for real encounters.
Also nyerking AoEs.


I'm not saying "Class A is the best PvE DPS class period", what I'm saying is "Considering both average and peak overall damage in instance B I deem class A to be the most efficient DPS class for Instance B." This does say nothing about it's effectiveness in instance C.
Assassins can easily be the best DPS for Katalamize, however the mechanics of Runadium make Assassins subpar for this instance, as such, Assassin is not the best PvE class for Runadium.
However, having an Assassin as the only close range class in Runadium makes a run through the instance insanely fast - (even though Gladiators do just as well due to AoEs).

Also, as evident by the little data I have (that's why I need more), while it still might be random, there seems to be a slight correlation between amount of Assassins and Rangers and how fast Hyperion dies. (Or rather, the other way around - having many cloth classes slows down the run on average?)

PvE DPS is not onedimensional and with a big enough sample size we can draw conclusions regarding party setup and whatnot. I can see how good cloth classes do with a Songweaver among them or with a second chanter, how does Infernal Blight affect stuff? Dual Infernal Blight?
Plus I really love datamining and making nigh pointless statistics.

@male

It's not DPS as ARM would display it. It's really the entire damage a class did throughout the whole boss fight divided by the whole time the boss fight took. I need a common denominator, as already stated, I can't disregard downtimes certain classes have because of a dungeon's mechanics. They are part of the experience and make some classes more or less optimal for the specific encounter.

Edit:

Just to show you how 'encounter specific' and 'ARM' DPS compare, I made a small list a while ago regarding the best individual DPS for each class.
This is what it looks like:

Keep in mind, this a a little bit outdated, I do have better values for certain classes by now.
Also my server (Vehalla) is considered rather weak and all my data, as of now, is from Vehalla (Elyos and Asmo), Balder (Elyos) and Kromede (Asmo). The vast majority, obviously, from Vehalla Elyos.

Edit II:

Also, I'd like other people to interpret the raw data as well, I'm biased and probably don't see all of the possible correlations.

I'm not trying to claim absolute 100% certainity. Yes, there are individual runs where certain classes, such as templar, are on top. by comparing this damage to how the other classes did, I can put this in proportion. I can tell how far from 'average' peak damage is. Yes, Skill > Gear holds true to differing degrees. A very geared Assassin can just be as bad as a god tier played Gladiator on Battle Medaillon gear.

But over many samples, all this averages out - while I simultaneously can pick out individual instances out of the norm.
Even how Adds are dealt with averages out, how many everyone dies averages out. With enough data, I can back up any claims I make by actual numbers than my gut feeling, which feels a whole ton better.

I just wanted to get the attention of some people to click on this and send me some data. At the moment, I am only comparing these two instances because I deem them to be most imporant ones. Rune Tower often gets their boss runs sold or given to twinks. Sauro I don't do often enough, Steel Wall is *really* tough to compare.
I do have enough data to make a list of 3.x DLR, but I don't really want to bother with that.

Yes, I am well aware that alliance setup has a huge influence, however by putting all of this in a list, this both averages out of time but I can still look at each individual run and say 'Oh look, this Assassin has over 9k DPS *but* he's had two chanters and two SMs to buff him. That puts stuff in proportion.

[NICK]erufuun wrote:

however the mechanics of Runadium make Assassins subpar for this instance, as such, Assassin is not the best PvE class for Runadium.

Nope.

Nope as in...?

https://dl-web.dropbox.com/get/runa%204k?_subject_uid=149363226&w=AACd8RNPjBSFpUpMSpTOafc4ZNpqbbZYY79tZFh1hE5gAQ&dl=1

This is my latest runa run I had from idk 4-5 months ago.
ARM says 4.051 DPS, with your calculations 3.116
Considering the crapy gear I had back then and single BM bow +0, now with current one I am positive I'd reach ~3.5k+

Thanks for sharing these!

Rin wrote:

Don't really understand what you're trying to figure out by overcomplicating everything

Any dpsmeter's total damage output(not DPS) is what you should look at if you wanna know which class is better for certain instances.

arm even gives priority to total damage dealt rather than to top DPS.

All in all a templar can outdps a gladiator if the gladiator is bad enough, doesn't mean templar is good pve dps, because, it always comes down to stats and who is behind those stats (player).

Alltho i come first in most katalamize runs (gladiator), I know for a fact some assassins can outdps me there and any theory just gets blown out the window when someone is godlike at his class.

I don't see how I'm overcomplicating it by itself. The graph is straight forward. It's absolute damage for each individual class. The conclusions are much more complex though.
What I am looking at is the total damage output, after all, I consider ARM DPS to be utterly pointless. But I still need to consider how fast the dungeon was cleared. If I had a single DD with 4kk damage and everyone else resting at 500k each and the run taking 19,5 minutes, I need to consider this. That's where my weird idea of 'DPS' comes from.

Yes, an individual Templar can outdps a Gladiator, if he is bad enough. But usually, he doesn't - over many runs, I'll have my fair share of god tier Templars and god tier Gladiators after all.

And don't forget, that's why I'm asking for help here, as a single person I simply can't run instances with all of the actual god tier players. My sample size using only the runs I participated in is both small and biased. (Many good gladiators, very few leather classes!)

Rin wrote:

arm shows absolute damage dealt over any period of time you need, just hit the reset damage button.

What is so hard about it?

and yes you are overcomplicating, who the hell wants to look at that clusternyerk of a document?

Uh, I don't quite understand? I'm doing it.

I reset ARM the second the Shield goes down/the Nevilim is killed and save the session as soon as the boss is dead. After that, I look at the timestamp of the 'First Hit' and the 'Last Hit' and take the time inbetween, which is how long it took to kill the boss.

Also...
Me. And I'm sure there are other people out there. Even if not. Me.
I love to be able to back up all this nyerk I'm spouting with actual data rather than with "Uh, I guess.. um.. SM is OP... I think"*


* (Which they are.)

Edit:

D'aw I see no double post. Too much cuteness

Am I being unclear?

Time certain classes have nothing to hit are considered downtimes, which are part of the experience.

It corresponds to the effectiveness of classes for certain encounters. Ranged DDs hitting Modor at range have an advantage over classes who can't do that. Hence they are better suited for the instance.

@Balls
Also, these links don't work for me.
How much do the ranged DDs help you with the Adds in Runadium? Considering that amount of DPS, I'd guess you're completely doing them solo. xD