Alright, so I've posted several times in other class balance discussions about wanting to do normalized/controlled testing to see what people are actually doing on their class right now. As an alternative to avoid the gross inaccuracies of just "looking at a parse", it's not perfect, but it can go a long way towards removing gut feelings and buff distribution from the equation and help show where people fall out in a controlled test. First off, grab the spreadsheet I created: download link The spreadsheet is a version of an old one I made to compare items to see which would be better (the oft-discussed "Potency vs Fervor tradeoffs", generally). It should be fairly straightforward, and has some instructions in it as well. Basically, it's intended to do two things: - Normalize your Final Ability Damage against a baseline. - Adjust your DPS by that normalized modifier so we're all comparing (as much as we reasonably can) DPS output to eachother based on the same multiplicative stats. So, here are the conditions I recommend we all use, if you want to help participate in this controlled test. There are a few small flaws in this, but short of putting together a full raid and doing a dummy parse and then changing things around in a controlled way and doing it again, this is a good way to get an idea of what people can put out: 1) Fight a single, epic, immortal training dummy. 2) Do it solo. No outside influence, either buffs or debuffs. 3) Each test should last a minimum of 10 minutes, or up to as long as you want to keep going. This will allow for several uses of longer cooldown abilities, and smooth out the effects of RNG. 4) Run at least two tests and compare, to see how close they are. If they're way off, keeping running tests until it smooths out. My first few test results (included in the spreadsheet) were almost exactly the same. 5) Do not use mitigation debuffs, where feasible. Without running a full raid against a target, classes that get a lot of debuffs that align with their damage type would show skewed data. For instance, I specced my warder out of the "debuff all magic damage" that it casts. I realize that some classes can't prevent using debuffs (pets, other stuff), but to the extend that it's not tied to a huge damage spell, consider not using it. Brigands may not be able to realistically participate in this test because they have so many debuffs. 6) Do not use any Ascension abilities. We're trying to measure base class output here, not Ascension DPS. We'd have to make the tests much longer than 10 minutes each to smooth out RNG in Ascension, and we're also unable to account for combo abilities using this method. (Your basic maintained POT buff is fine to use, as it's just a basic stat increase, and that gets normalized out anyway. Note that you can use this to purposefully change to a different ascension damage type if you are a class that gets a lot of unavoidable mitigation debuffs against a specific resist and you're trying to control for that.) 7) Do not use Rising Tide or any similar pre-pull clickies. No reason to skew the first 10 seconds of a 10+ minute parse when we're trying to control variables. 8) Begin from a cold start. Everything will smooth out over the course of 10+ minutes anyway. 9) Do not use any significant buff that either grants a persistent, major multiplicative stat increase (mostly Fervor) to the group or that would normally be placed on another player instead of yourself (Bolster, etc). This means that Conjurors should not maintain their Fire Seeds, for instance. I set the "baseline raid fervor" to a level that assumes that you're getting a healthy dose of Fervor from your groupmates; if the classes that are providing those buffs (conjuror and troubadour, mainly) use them in a solo test, it'll skew the results... it's already accounted for in the normalization, so don't use it. 10) Anything that is cast or proc'd on yourself is fair game to use in combat. This means that scouts should keep their Ethereal Ravaging runes slotted, for instance. Do not try to add in "extra fervor" to your stand-still stats in the spreadsheet based on temporary sources like this. You get those procs in normal raid combat, so use them here. That's about it. Record your stand-still stats (elixirs/scrolls/etc up, but cold start with no temps running) and then run some tests. The spreadsheet will do the math for you. This is intended for all interested classes to participate, not just scouts... ...I had to choose a forum, and I chose this one. Note to potential "I'm trying to get my class buffed" dirty sandbaggers: all it takes is one other person on your class to actually try hard and post results here and you'll be exposed for gaming the system. If nobody else wants to help with controlled community testing to help inform the devs class balance efforts, hopefully at least a few other Beastlords will do this and we can compare notes on cast orders. Other classes are encouraged to do so as well. I'll ask a moderator if I can get perma-edit permissions for this thread so I can maintain a consolidated list of results for everyone. A few final notes on possible flaws in this methodology: - AbilityMod softcaps: Our Potency should be high enough that we're past the point of AbilityMod soft caps for most spells that actually matter, so I'm not too worried there. - Crit%: Anyone with decent gear should be way past the point of diminishing returns, so I don't think we need to normalize for this stat. - Scouts With Ravaging Proc: Depending how much fervor you're running in combat in raid, you might be running into the fervor hardcap, even with overcap, when Ravaging procs. This means that your DPS in this test environment is going to skew a bit higher (relative to the normalized baseline) than it would in raid combat. I can't really control for this without asking scouts to remove their rune, and that'd make us look super weak next to mages, so sorry guys... we're going to appear a bit stronger relative to mages unless someone has a bright idea on how to control for that. Just assuming it's up 25% of the time and adding in +25 fervor to your baseline is the closest we could consider, but that then doesn't account for unloading burst damage while the proc is active.
Alright, my initial results. I'll mention where I deviated from any of the above rules, and request anyone else do the same so we can see trends and adjust perhaps. Class: Beastlord Test 1, normalized DPS: 2.094B (10 mins) Test 2, normalized DPS: 2.114B (10.5mins) Deviations/Notes: None from rules as listed. No mitigation debuffs of any kind used.
I used it, as it's part of my baseline raid stats. It doesn't matter either way though... that's the beauty of normalizing stats like this.
Troub: 24,380 pot, 4150 cb, 402k mod, 60 fervor (no hb) first test, 442m dps over 11:23 second test: 456m dps over 10:52 Test 1 normalized dps: 660,669,291 Test 2: 680,602,214
So, after pondering rule #9 above, I don't think we need it. The normalization math is designed to correct for any level of over/undergearing and bring us to the same level. Go ahead and keep your group fervor buffs up
Warlock: 25966% Potency 355,000 AB Mod 4150% CB 84% Fervor (With HB) Test 1: 861,088,547 (1,073,367,011 Norm) 10:11 Time Test 2: 853,745,120 (1,064,213,257 Norm) 10:24 Time
Finally got 2 10+ min pulls. I did this on my laptop from work so between phone calls and PC alerts it was interesting. either way on the other pulls even tho less time they were similar which tells me i'm consistent. so here it is: Beastlord: 29,210 Potency 405528 AB Mod 4150 CB 100.6 Fervor Normalized multiplier of: .9958 Deviation notes: none of the rules. played at my best with a new primal order i'm working on. I had 2 rav procs come late but for the most part it was smooth with limited lag. I also removed my pets mitigation debuff and never cast mine Results: Test 1-10:58 duration: 2,393,877,468 (Normalized to 2,404,019,614 Test 2 -10:11 duration: 2,445,225,976 (Normalized to 2,455,585,670
I decided to take this a step further. To test Entropy's Normalize formula. So I turned off my HB and my Ascension Potency buff. I re-entered my new stats into the calculator and this is what I got: Beastlord: 26062 Potency 405528 AB Mod 4150 CB 70.0 Fervor New Normalized multiplier of: .7632 Deviation notes: I kept this as consistent as possible to the previous tests. no change in cast order. There were again a few times rav procced late and I adjusted accordingly but it also felt smooth, not much lag. Results: Test 1-10:18 duration: 1,762,608,607 (Normalized to 2,309,592,245) Test 2 -11:27 duration: 1,775,298,428 (Normalized to 2,326,220,049) Previous Results were: Test 1-10:58 duration: 2,393,877,468 (Normalized to 2,404,019,614 Test 2 -10:11 duration: 2,445,225,976 (Normalized to 2,455,585,670 This is a deviation of 4.8% id say that is pretty good. So this test should yield some strong results if we get enough people to do it honestly. Nice work Entropy!
Here's mine Warlock: 19570 Int 22992,5 Pot 359536 Ab Mod 4150 CB 39,6 Fervor (No HB) Test 1: 544.041.942 (Normalized: 989.697.919) (10:27 duration) Test 2: 550.145.163 (Normalized: 1.000.800.638) (10:07 duration)
Beastlord: 17851 Agi 26589 Pot 443229 Ab-mod 4150 CB 78.6 Fervor (With HB) Test 1: (10:30) 1,912,887,670 (Normalized 2,374,005,368) Test 2: (10:13) 1,890,236,057 (Normalized 2,345,893,393)
Did a second test with Left-Side prestige (first one is rightside) just for funzies. Note that my Crit Bonus is low because I normally run Right-Side spec and going Left-Side means losing my Potency to Crit Bonus conversion Warlock: 19570 Int 23830,1 Pot 359536 Ab Mod 3009,7 CB 39,6 Fervor (No HB) Test 1: 471.014.734 (Normalized: 1.134.920.910) (11:03 duration) Test 2: 483.165.070 (Normalized: 1.164.197.417) (10:41 duration)
Brigand agi: 20744 pot 25247 cb 4056 fevor 60.2 test 1: 698331396 test 2: 787176533 NOTE: i have alot of debuffs so my numbers might be a tad higher then they should be
Coercer (Geomancer - So everything physical conversion) Main Stat - 18,363 Pot - 22,007 AMod - 363,041 CB - 4150 Fervor - 59 Standing at 9m Range. Ancient Pet. Normalized DPS 1 - 860,356,206 Normalized DPS 2 - 882,466,629 Normalized DPS 1 - Without Pet - 626,707,278
Did another run at the dummy for fun (mostly bored right before servers came down): 25966% Potency 355,000 AB Mod 4150% CB 84% Fervor (With HB) 912,980,765 (1,138,069,076 Norm) 10:41 917,549,123 (1,143,763,727 Norm) 10:38 I think it is worth noting though that even though my fervor says 84% while standing there staring at the dummy, I can keep my 25% fervor proc going 99% of the fight (There are a few times when I am a second late on re-upping it while juggling other buffs). So through the fight I have 109% fervor not counting any other fervor procs. Just for the argument's sake that would change my normalized DPS numbers to: 1,001,936,412 1,006,949,884 If we consider that a "full time" self buff. (even though it does take some thought to juggle it). If we don't count it and we say I have to go without that number, I'd say that maybe I need to conduct more tests without the buff active at all? What do you guys think? My only issue with that is my cast order would change, since only 1 of the spells used to juggle the buff is worth casting for the actual DPS value of the spell.
I would say since it is not a static buff and requires player skill and attention to maintain it it should not be counted in the base stat calculations. So you would leave your fervor as 84%. It's in a way similar to our 20% base pot stack that is up 60% ish of the time.
So I got curious at your calculator, and got my 2nd Coercer out. Bare in mind this Coercer does not have it's epic yet, so realistically it should only be compared to the DPS above with no pet. Normalized DPS 1 - 772,758,076 Normalized DPS 2 - 761,983,720 Main Stat - 16,481 Pot - 19,449 AMod - 291,072 CB - 4150 Fervor - 4.5% There is also around a 500% CC difference between Main and Alt. So how much faith we can put in these overall numbers, I do not know.