Class Statistics

Toz

So, somewhat inspired by the thread on BB, I've started a project using my vast army of alts to collect the rough APS (afflictions per second) of classes. This is things that show up on hit on affliction_view - so epteth counts, but an arm broken by limb damage doesn't, for instance. So what I'm going to do is try to find a nice, pretty way to display my data, then place a spoiler beneath explaining how I calculated my answers. What I need from you guys is, there are a few classes I don't have access to/am not good enough at to tell, and I need numbers for them.

Some brief notes about my assumptions for collecting this data:

-I'm assuming no artifacts except an EQ crown (I was assured some classes cannot fight without it). No arti weapons, either. +bal or +eq where appropriate.

Calculating ACTIVE APS:

- Everything you can do actively, every round, that delivers an aff is counted. I count Syssin dissipate as 1 affliction per round, I count anything that gives an affliction even if it's a 'worthless' one. I am assuming BEST CASE for all classes. This means if you have an ability that makes you attack 10000 times a second, I'm going to assume that's always running. We are assuming best case only as well, for now - no rebounding, no shield, no stupidity, no clumsiness, nada. Just, if I sat there and let you smack me and you were going all out with every skill possible, how many times could you ACTIVELY afflict me a second? Simple calculation here - if you can only attack 2x a round, and it takes you 4s to get balance back, 2 affs a round / 4 seconds of balance = .5 Record .5 as your Active APS and then note your calculations in a spoiler.

Passive APS Calculation:

- Again we assume best case - for instance, with sentinel I assumed 'opportunity' was running, and resins were burning as well. That doubles ent attack rate and resins tick at a rate of something like 1 every 4 seconds, but you can only use 3 of them. Again, I'm after MAXIMUM BEST APS, and nothing more.

- IF it is an ent with a % chance to proc (ghast, loki hound), I let it hit me 30 times and used that for the percent. For loki hound, for instance, it hit me less than 50% of the time, but I'm lazy so I rounded up - so, 50% of the time hitting say every 4 seconds means I double 4 seconds to 8 (50 * 2 = 100), then divide 1/8. 

- For attacks that hit every round on their own (angel battle, pestilence aura, etc.), I'm looking for 1 over the tick rate. That's how many afflictions per second that particular skill is worth.

- For most classes, or at least a good portion, there'll be more than one of these. Add all the APS calculated up, then that is your total for Passive APS.

Finally, add Active and Passive APS up to give the grand total APS of the class. PM/post your results, and I'll edit into the master list. Format should be, if possible, as follows:

Active APS: x

Passive APS: y

Total APS : x+y

[spoiler]

I calculated Active APS by:

I calculated Passive APS by:

Toz

DARK CLASS APS TABLES

Atabahi:

Active APS: 0.749

Passive APS: 0.2

Total APS:  .949

[spoiler]
I calculated Active APS by:
Rend/hamstring 2 every 2.79 2/2.79 = 0.717 Bay every 31s for 1 aff
1/31 = 0.032
.717+.032 = .749 Active APS
I calculated Passive APS by:
3 howls every 15s
3/15 = .2
I calculated total APS by:
.749 + .2 = .949[/spoiler]

Bloodborn:

Cabalist:

Carnifex:

Active APS: .781/1.333 (0 soul/max soul attack speed)

Passive APS: .313

Total APS:  1.094/1.646

[spoiler]
I calculated Active APS by:
- using a 71/101/168 halberd -2.56 base, 1.5 max, 2 affs/swing 2/2.56 = .781, 2/1.5 = 1.333 I calculated Passive APS by: max speed hound = 50% chance to hit loki, every 4.41 80% soul loss, affs tick every 5 seconds 1/5 = .2 1/8.82 = .113 Total Passive APS = .313 I calculated Total APS by: .313+ (.781 or 1.333)

[/spoiler]

Indorani:

Praenomen:

Active APS: .926

Passive APS: .218

Total APS: 1.144

[spoiler]
I calculated Active APS by:
3.48 dwhisper - (3.48 * .07) (crown) = 3.24 3 affs every 3.24 (dwhisper+jab with weapon) 3/3.24 = .926 I calculated Passive APS by: 1 every 8 (yellow bile) 1/8 = .125 50% on ghast, every 5.4, 1/10.8 =.093 Passive APS = .218 I calculated Total APS by: .218+.926 = 1.144

[/spoiler]

Sciomancer:

Syssin:

Active APS: 1.215

Passive APS: .25

Total APS: 1.465

[spoiler]
I calculated Active APS by:
2.47 dstab speed+2.47 shadow bal recovery
Bedazzle is 1 aff every 2.47
2/2.47 + 1/2.47 = 1.215

I calculated Passive APS by:
Hypnosis is 1 aff every 4 seconds (3-5 estimation, 4 is average)
1/4 = .25

I calculated Total APS by:
.25 + 1.215 = 1.465
[/spoiler]

Teradrim:

Toz

LIGHT CLASS APS TABLES

Ascendril:

Bahkatu:

Daru:

Luminary:

Shaman:

Sentaari:

Sentinel:

Toz

Ok. I'm going to not bother updating it until I can figure out a way to do spoilers that doesn't look terrible. Again let it be noted that I absolutely *HATE* these forums.

Haven

There should be no spaces between the tags. They -must- be sandwiched.

[spoiler]

Lalalalala. Look at me, hidden.

[/spoiler]

versus

[spoiler] 

PEANUT BUTTER LESSER TIME 

\:D/

[/spoiler]

versus

[spoiler]In the jungle, the mighty jungle, the Lions sleep tonight![/spoiler]

Toz

@Areka saves the day by PMing me - use shift+enter instead of just enter, and it'll all go in the spoiler box.

Haven

Huh. Weird. I never had to use shift. -shrug- Whatever works!

Setne

Haven to the rescue(as usual).

Irruel

What do you hope to gain by comparing burst speed APS (of classes that do burst affs), with the general APS of classes that don't burst?

Not saying you shouldn't be measuring it, nor that it isn't relevant, but when comparing a vamp with slash/dwhisper/ghast, to a sentinel, you will see two APS figures that are probably wildly different. We then say, oh, but the sent one is a burst rate requiring cooldowns and setup - which renders the comparison obsolete.

I'd say that for burst classes you need the baseline as well as burst rates, and maybe something useful can be learned from that.

Toz

Definitely do, yes. But I want a baseline for all classes first and foremost - you could, for instance, assert that Sentinel APS vs Vamp APS is meaningless because rebounding/shield changes things (vamps can give 2 active affs even with shield/rebounding, sentinels can only give 1 active). But to begin modeling, you have to first make some assumptions and get SOME data for a baseline, then refine from there.

Hadoryu

It will seem more tedious to you, but you're better off getting logs and counting the afflictions delivered over time there. This theorycrafting has been done before, but the numbers you get are basically useless, because there are a LOT of factors going into how fast a class delivers afflictions. Even ignoring all the conditionals (burst, rebounding, setup), your affliction rates will still wildly differ against different classes. You're not going to have the same affliction rate as a Sentinel fighting an Indorani as you are as a Sentinel fighting a mage.

Simply put, these idealized scenarios are so far divorced from the reality of what happens in combat as to be practically useless as justification for any sort of mechanical change.

Ishin

I think we did something similar before, @Hadoryu, when we were both liaisons? I'm unsure if we ever finished, but I seem to recall us doing some math about relative aff rates including things like rsl/flay. If memory serves, it's how I talked folks into letting us deliver a venom with flay or something, maybe. >.>

Not that it isn't cool that Toz is posting it on the forums, I'm of the opinion that sometimes raw numbers like these can really help someone when they start trying to decide what class they want to play - They can get an idea if a class is limb-based like lycan, pure aff-based like Syssin, or somewhere in between like BB/Lumifairy.

One thing I do disagree with you about, @Hadoryu, is that I don't believe that Toz's numbers are that far divorced from reality. I think they'll prove immensely useful when compared to your idea about showing the actual log, in how the pure mechanical numbers are 'supposed'(right?) to be, theoretically by design, and in how they turn out being in practical application.

Hadoryu

Well, this sort of thing has been done a number of times, really.

Then when you start trying to draw conclusions from it, people in the 'affected' class will start raising (quite legitimately) objections to the method, as it's not accounting for X, Y, Z complications.

My point is that in actual combat those numbers are practically never accurate, even if they are accurate in a vacuum. Ostensibly little things like the fact that DSK (or whatever it is now) requires two working arms and dwhisper requires ... almost nothing, do realistically play a huge role in what the actual output of the class comes down to, as they are compounded by many, many complexities.

Sure, a comparative analysis would help bring to light these little obscuring factors and that would be beneficial in its own way. I'd just be wary of making a comparison between the classes on the basis of these particular numbers.

Toz

@Hadoryu, there's a real lack of logs at the moment. Right now, I want a simple baseline - after that, things like how many affs you can deliver against rebounding (I'll only factor in active APS for that one), how many things can hinder you, and that sort of thing - while it won't ever be truly 100% accurate, I do think it's possible to build a model of some sort that will actually be useful. Because frankly, at the moment, if a log is presented (see @Daskalos'), people go 'oh you did it wrong' or 'it's not that bad' despite not actually providing anything of substance to explain how/why the strongest defensive class in the game got rolled by BB. And when you ask for a log proving that a class is actually beatable, everyone just sort of shrugs their shoulders. So, since there's no reliable qualitative data, I'm aiming to work (tedious though it may be) on getting as much quantitative as I can.