Guild Wars 2

So, one might ask, what is all the hype with running with full zerkers? Why do so many people run it? Some time ago a spread sheet was created that made some assumptions. From what I understand, it was made under many assumptions, assumptions that do not accurately model the situation regarding the people that pug, people that solo, etc. Assumptions such as 25 stack of might, 25 stack of invulnerability, 100% of fury, and so forth. Under these numerous buffs, the dps of warriors was estimated, gear choices were made, decisions were made. Later, gear was introduced with precision as the primary stat and it was evaluated. (Im not entirely sure, but I think the assumptions were also re evaluated at some point. But I do know that the spread sheet still do not accurately reflect warrior dps on an individual level.) But if your crit chance is capped through external buffs, does assassin’s gear look very attractive? No of course not. Thus zerker gear remained the preferred choice.

Then, following up with an equally good question one might ask, why has zerker remained the preferred choice? Surely some pug would have caught on that zerker is an inferior choice to mixing up/full assassin’s for them, at some point right? Well, running a mix of assassins can appear to have a superficial flaw (that isnt acutally a flaw); your crits are not hitting as hard as what they were when you were running full zerker gear. So to someone just comparing their 100 blades to each other, a natural bias forms due to human nature. We can sometimes disregard/overlook flaws in favor of a preferred outcome. And since our objective is to attain as much damage per second as possible, bigger crits would, seemingly, achieve this goal, and our preferred outcome established.

But the reality of the situation is that the math simply does not support this outcome. Collectively, the smaller crits sum to a greater total number of damage points than the larger. I am ready and willing to prove this mathematically, but before you ask, consider the experiment above; would you rather hit for massive damage 1 out of 10 hits or hit for 1/10th that same massive damage every hit with some chance to crit greater than 1/10?

Ok, now that everyone is caught up, back to the conversation with Purple Miku. Yeah, I should have guessed that you running an undead potion. Arah is a very popular dungeon. You might even run with undead sigils to further boost your dps which would further deflate your stats. But considering how often you solo, surely you do not always run with three offensive skills on your bar. And even if you do, both banner and fury do not have 100% uptime, leaving a probability that both fall off at the same time. So, should I assume your run assassins on your solo runs? And if not, why?

I see my subtlety has gone unnoticed, which I am clearly not good at. I hold both [rT] and you, Purple Miku, with high regards so I will attempt to be as respectful as I can, while still addressing (what I see as possible) flaws in optimization. But first I want to catch up anyone who is new to GW2, warriors, soloing, pugging or people that run in guilds that may as well be pugs.

To those I ask, what is the most important stat that a warrior has? Which of all the stats affect the warrior dps the most? Is there even a stat that affects warrior dps the most? If you said precision, you are on spot on.

Imagine its your Bday and you are given Bday moneyz. Well, actually a choice of Bday moneyz. You are given 100$ and then must choose either; to accept another sum of money up to 149$ or multiply your current holdings by 1.5. Which do you choose? Easy, right? This is the same problem we face in GW2, except the first option is more like 20$ and the second is some constant between 1.5 and 2.5.

Precision is easily the stat that increases our dps most (until the cap is meet). Luckily for us, warriors can readily reach the crit cap, making our ability to maximize our dps much more likely. Precision allows power to become far more effective through multiplication, making warriors a very, very strong dps class. If you are part of the set of people that I mentioned above and are not running assassin’s gear, are not taking into consideration what what banners/signets you should be running or are not even mixing up zerkers with assasin’s, ponder the following.

Each time you hit some mob or boss and you DONT crit, you have just hit as hard as someone running power/vitality/toughness. Meaning, you just had a hit that COMPLETELY ignored one of your secondary stats, ferocity. And if you are wondering how often you are ignoring that secondary stat, all you need to look at is the compliment of your critical chance (1- crit chance). So if you just calculated that probability, that same probability is how effective someone running power/vitality/ferocity would have done in your shoes that same percent of the time during that last fight. I.e., your dps = (crit chance)x(you running w/crit at cap) + (1- crit chance)x(you running with power/vitality/ferocity). Crazy, huh?

Oh, but it gets better/worse. Any person that pugs, solos, runs with guilds that might play like pugs (the set of people mentioned above), and is NOT attempting to run with their crit at cap and is running in some build with forceful greatsword, is just throwing their dps in the toilet (with a few exceptions, such as the noble souls who sacrifice their dps for Phalanx Str., or you have a considerate elementalist in the pug party.) Barring these exceptions, when you DONT crit, not only did you just hit for at least 1.5 times less than you could have, you just missed one might stack (or 35 power), which just further increased how hard you could have hit on your next swing. Meaning, if you are running under the crit cap during some portion of a boss fight, you just missed out on many stacks of culmulative might during period. And if you are running 65003, your axe rotation likely started out might starved and ended might starved, which really kitten your axe dps.

But still there is more. You know how everyone states that healing power scales poorly? Yeah it does. But you know what? Take away critical chance from the dps function and guess what? Power and healing power scale in the same manner. (Side note; It would be really interesting if Healing would be allowed to crit.)

oxtred.7658:

Sharpening stone makes a difference though, since you’ll be using a X slaying in dungeons most of the time.

Unless you are running fractals and dont mind wasting lots of Gees.

Purple Miku.7032:

Is that with 3 assassin pieces? I’m assuming so, and assuming you aren’t using a utility nourishment like maintenance oil or whatever.

No and no. I am using 6 piece Assasins w/Str runes and Sup Sharp stone for utility nourishment.

Purple Miku.7032:

I’m currently:

2578 power
2311 precision
895 ferocity

70% critical chance
210% critical damage

with truffle steak nourishment (+100 power +70 precision) and without might, fury, or banners.

I just activated banner + for great justice + signet of rage and took a picture here to make it easier on myself: http://puu.sh/blEoW.png

Same build (6/5/0/0/3) and same slot skills activated but with a rearrangement of stat priorities you can yield 78 more power, 5% crit damage and 1% more crit. (No special infusions used, just versatile simple infusions.)

http://i.imgur.com/KFbkx7R.jpg

Just thought you might be interested in knowing.

I love love love this gift. I can now offer unprepared pugs in dungeons food at no expense to myself. And an emergency source of MF for world bosses if I forget/run out of food. Such a useful gift, best in game mmo present ever!!! Thanks Anet!

Lilith Ajit.6173:

Jerus.4350:

haviz.1340:

aelfwe.4239:

almost doest exist on math.

It does

There are also Limits, which is basically “almost to X but not to X” /shrug.

PS. Calculus is evil!

Limits are like precalculus brah.

Except for those ones on the Riemann sums. Man dey tuff.

Unless you make the colossal mistake of taking a topology or real analysis class, where limits take on a completely different definition. But at that point you dont care because you are too busy trying to kill yourself before the next test.

Guanglai Kangyi.4318:

I’m talking kitten because you make no sense.

Welp, fair enough.

Guanglai Kangyi.4318:

That’s not how percentages work. And you realize you’re pretty much also admitting you made up the second number.

First off please drop the attitude. I totally understand wanting people to prove their claims, especially in this game where it appears that no one can, but realize that I am getting sick of talking to people that have no basis for their pride. Talk reasonably to me and I’ll respond reasonably to you. Ignore this request and I’ll ignore you.

And yes, this is how percentages work. If you want we can take it out of percentage form and put it in a ratio. A percentage is defined as part over the whole. We are taking 1/20th of your total condition damage and converting it to actual damage. In other words you are getting 1 unit of damage per 20 units of condition damage. The same argument holds true for the power scalar I quoted earlier.

Guanglai Kangyi.4318:

oilstorm.1748:

xxHellBladexx.9163:

Which do you prefer and why. (Heavy details and explanation would be very helpful.)

There are many reasons I prefer to stay away from condi damage, but one of the biggest reasons is the scaling. Direct damage scales so much better than condi damage. Condi damage scales with condition damage at a rate of 5%, but compare that to something like axes, which scale with power at a rate of approx 27.3% (approx since anet likes to their integers) and you can see where you are getting the most bang for your buck.

A second important reason I stay away from condi damage is the sustainability. With the exception of rifle auto attack, the cd of the abilities is much longer than the actual dot. But since most of us have at least 5 in Arms, offhand sword is the other exception. Even doubling the duration of condition damage only nets you 2 more abilites to work with, one of which is a water ability, over all weapons including water. This means that the rest of the damage you are doing is direct damage while you wait for your cd to reset.

In the case of rifle auto, assuming the auto attack is once per second (?) and that ALL you are doing is auto attacking, your expected stack size will be just 6, or 7.75 in the case of 5 in Arms. And even then, the direct damage hit when applying the dot did more damage than total damage that the dot did over time. Additionally, this average stacks size can only decrease, thus lowering your dps, if you break your auto attack chain for whatever reason. (This only holds true for condi damage that has not had the duration extended past the 5 in Arms.)

I think the only plausible case of an over all damage increase in the form of a dot is in the case of offhand sword. But even then, some analysis must be done to verify that your dot is doing more damage than, say, offhand mace, which should be used to stack invuln.

Did you just make up those percentages?

Nope. One of them is even on the wiki page. http://wiki.guildwars2.com/wiki/Condition_Damage

xxHellBladexx.9163:

Which do you prefer and why. (Heavy details and explanation would be very helpful.)

There are many reasons I prefer to stay away from condi damage, but one of the biggest reasons is the scaling. Direct damage scales so much better than condi damage. Condi damage scales with condition damage at a rate of 5%, but compare that to something like axes, which scale with power at a rate of approx 27.3% (approx since anet likes their integers) and you can see where you are getting the most bang for your buck.

A second important reason I stay away from condi damage is the sustainability. With the exception of rifle auto attack, the cd of the abilities is much longer than the actual dot. But since most of us have at least 5 in Arms, offhand sword is the other exception. Even doubling the duration of condition damage only nets you 2 more abilites to work with, one of which is a water ability, over all weapons including water. This means that the rest of the damage you are doing is direct damage while you wait for your cd to reset.

In the case of rifle auto, assuming the auto attack is once per second (?) and that ALL you are doing is auto attacking, your expected stack size will be just 6, or 7.75 in the case of 5 in Arms. And even then, the direct damage hit when applying the dot did more damage than total damage that the dot did over time. Additionally, this average stacks size can only decrease, thus lowering your dps, if you break your auto attack chain for whatever reason. (This only holds true for condi damage that has not had the duration extended past the 5 in Arms.)

I think the only plausible case of an over all damage increase in the form of a dot is in the case of offhand sword. But even then, some analysis must be done to verify that your dot is doing more damage than, say, offhand mace, which should be used to stack invuln.

Dalanor.5387:

I’m 12 year old and what is this?

Walking like cakes, care to join?

deSade.9437:

Oh, please, who are you trying to fool.

You? Did I get you? Awwww, drats.

deSade.9437:

It’s like when you answer a question regarding a previous statement with " I don’t understand what’s so hard to grasp". NO, surely not annoying and [improperly used adj.] condescending at all.

Similiar to Lilith, you seem to think I have made a claim that I did not. Claim 1: I dont understand why a group of people are upset. Claim 2: I dont understand whats so hard to grasp. Two totally different claims. They are not “like” each other at all. Wanna try again?

deSade.9437:

People getting upset? Well, i… players who join xp berserker parties with cleric staff guardian surely have an amazing talent of raising blood’s pressure. Still, the ones that really make people tie themselves in knots are the cakes walking into the forum or joining parties and going “I don’t notice any difference, why would anyone bother! Why do you care, I don’t get it, I complete dungeons in 20 mins just fine”.

Was this me? Did it make you upset that I walked like a cake up in here? Do you want an apology?

deSade.9437:

It’s so irritating it’s not even fun to use it as a trolling tool, because the reasons why people care have been exposed over and over and over, even if…. even if it’s no one else’s kitten business why someone would care.

You sound mad. Are you mad?

/continues to waddle like a cake away

Lilith Ajit.6173:

Oh dear. Now someone is getting touchy.

Nah, not yet. Apparently you and I are all over the communication spectrum if you think I am already getting touchy.

Lilith Ajit.6173:

Do I really bark? I thought I was stating things rather eloquently, as I’ve been trying to improve this ol’ mind brain of mine.

Forgive me for interpreting your posts as such.

Lilith Ajit.6173:

Anyway, point is, you have not seen a staff guardian drag you down. Probably because you run a build that is just as… efficient as a staff guardian (AND THAT’S FINE). But when someone asks for a specific thing in an lfg, respect it, and we won’t have to complain about staff guards because we won’t see them.

Woah, woah. First off, please dont interpolate the efficiency of my build to be the same as that of a staff guardian. In fact, do you even know the efficiency of a staff guardian? My guess is that you dont and have simply heard that a staff guard is bad. (PS; I strongly urge to not bother responding to this obvious bait. You’ve been fairly warned.)

Second, I never even mentioned anything regarding posts in any lfg, for or against. And I actually do respect lfg posts… yet here we are.

Lilith Ajit.6173:

oilstorm.1748:

Lilith Ajit.6173:

oilstorm.1748:

Where in my post did I judge others standards? Please quote me.

significantly slowed us down.

What’s significant to others does not need to be significant to you. It is a significance that depends on a players standards within a dungeon run.

There appears to be some ambiguity in my previous post, so let me be more clear; Where in my post did I judge others standards to be unreasonable? Claiming that I dont understand why a group of people act they way they do and claiming how they are acting are unreasonable are two very different claims.

This:

. I have no idea why people on the forums get up in arms over this subject. 

Absolutely implies that you find it unreasonable. It exudes a negative connotation that implies that you feel that this “getting up in arms” about this subject is inherently unreasonable.

Getting up in arms is simply an expression to show that people are getting upset. In fact, I explicitly stated that I didnt understand why, not that its unreasonable. And if it carries a negative connotation to you… then thats your deal as I didnt express that. Please go bark up someone elses tree.

Lilith Ajit.6173:

oilstorm.1748:

Where in my post did I judge others standards? Please quote me.

significantly slowed us down.

What’s significant to others does not need to be significant to you. It is a significance that depends on a players standards within a dungeon run.

There appears to be some ambiguity in my previous post, so let me be more clear; Where in my post did I judge others standards to be unreasonable? Claiming that I dont understand why a group of people act they way they do and claiming how they are acting is unreasonable are two very different claims.

Lilith Ajit.6173:

So? Just because you can’t see it, it makes it unreasonable for someone to have higher standards than you?

Where in my post did I judge others standards? Please quote me.

VideoGamermike.5813:

To original post:
How about you realize he has a different style then you’re own. If it happens to not work in your party, fair enough, tell him the reason why and move on. Telling another player their entire fighting style is useless simply because of a choice of weapon. you handle it like a person facing differing ideals and combat, you adapt it into a team strategy.
I am glad that you’re trying trying not to be some beserker obsessed jerk, but you’re following the same mentality of “I see his build as useless, so I don’t want it”. I think through he post I’ve read in the last 4 days(2 as lurker), this game is balance enough to handle any potential combination, is why I even bother playing it. the last thing you want to do is make the environment toxic by trying to call out entire weapon sets garbage.

I agree with this mentality. I have no idea why people on the forums get up in arms over this subject. I pug all the time and I cant recall a time when a staff guard wiped us or even significantly slowed us down.

hendo.1940:

Maybe Warlord will be the second coming of Guang and prove us all wrong.

I think Guang has to die first, otherwise things might get awkward.

Cons to DWA:
~Decreases the relative effectiveness of time warp and frenzy
~In short bursts (less than 3 seconds), the trait is worthless
~Decreases attack time of skills but leaves the CDs of those skills the same, potentially leaving wonky rotations

Pros to DWA:
~Is (statistically) significant increase in DPS since it significantly decreases attack time of skills
~Increases the probability of finishing triple chop, thus increasing the probability that you can maintain your highest damage

Warlord, you are always welcome to come to whatever conclusion you want, but I think many of us are in agreement in that DWA is a contender trait for pure axe builds.

Welp… I prefer this craziness over telling everyone to stop posting… even if the conclusion is not the one I would conclude…

Dub.1273:

To make it easy for you, get your critchance to 100% and kill a normal mob in cursed shore with just axe autoattacks. Once without, once with the trait. Assuming you always crit and deal the same amount per hit, now check both in frame for frame format. If you can’t figure the kill speed difference then, a) just don’t ever come back to the forum or, if you can, b) apologize for giving false information, followed by a).

The time it takes to kill a mob is a not a proper indicator for DPS. The base number that is selected that goes through multipliers and whatnot to become your damage, is selected at random from a uniform distribution. The crit assumption is a legit assumption, since this is possible and we can control that, but we cannot safetly assume that one can “deal the same amount per hit”. Using time to kill a mob can lead to false conclusions, such as build X is superior to Y even if the builds are actually equal or Y is actually superior to X.

For everyone else in this thread, some of Warlords claims actually have merit. But for some reason, you have dismissed everything. Why? For people that want to move the theorycrafting forward, you certainly are doing a fine job of holding it back…

Edit: Dub, I am sorry I was being a kitten and should not have called you ignorant. Ive edited my post accordingly.

cranked.3812:

It’s funny that you were “so mad that you didn’t think clearly”, but yet you had the foresight and patience to wait until the the boss was at 5% health. Your excuse is worse than your behavior and you should be banned. Doesn’t matter if the guy was running cleric’s and you asked for zerk. Either gear check in the beginning and kick or don’t do it at all. YOU are the problem.

^My feelings exactly. It seems hardly fair for someone to put in the work only to be kicked at the end.

Blackjack.2083:

GW2 PvE lol…..good lord.

Agreed. The drama is mesmerizing.

Spoj The Second.7680:

Should be using axe/mace + GS or hammer. :PP

You are clearly missing the greatest opportunity a warrior has; sword/warhorn + sword/sheild. The warhorn allows for maximum cheerleading and the sword/sheild ensures that you make it there quickly and in style.

hendo.1940:

Secondly, why aren’t you using axe/wh + axe/mace for fotm? GS doesn’t do enough to be worth taking.

StaniMorell.2681:

the think is that the most classes do 6-8 k dmg per sec, if y can play, but the warri push themself to 20+ might, so fact is the most other classes with 25might make 2x more dmg as a warri. and now its going craxy, if the warrior with heavy armor, double hp and half dps going now more tank.. loool. you need only might for max 20sec. and thats all. i would vote phalanx for kick if i would ever play random. lupicus themself is a max 30 sec kill without probs. with the runes y lose 10%dmg, with power trait 30 y lose 15%dmg, with whole traits 30% or more, all in all 50% dmg…. omg! the think with eles is that they do not only stack i make with last hammer blow 17k damage , all 3 sec , the first 2 hits make 8-10k , i make 37k in 3 sec. with firegreatsword i make 3k x 80 thats =250k , with all greatsword skill on 10sec mayby so 300k , … palanx warrior ???? 30k on 10sec ….go home …. mayby wvw ^^

You sir… are amazing. How on earth do you play an MMO with so few fingers to type with? And how did you lose them? I am imagining some heroic act where you picked up live grenade near a widowed mother and attempted a miraculous throw… Or you reached in a sharks mouth, pulling out an orphan but sacrificed those beautiful fingers of yours. Heck, if you could still catch thrown objects, I would throw you a beer because you are a real american hero!

Tree.3916:

1,000 base damage + (1,000 * 1.0) = 2,000 damage
1,000 + (1000 * 1.1) = 2,100 = 4.8% dps increase
1,000 + (1000 * 1.2) = 2,200 = 4.5%
1000 + (1000 * 1.3) = 2,300 = 4.3%

This continues infinitely.

The function f(x) = 1,000 + (1,000 * X) is linear. ( Y = MX + b) The function g(x) = 1 – ( previous damage)/ (new damage) is very much not linear. In fact, if you treat power in the same manner;
1,000 + (1,000X) = 1,000(1 + X)
1,100 + (1,100X) = 1,100(1 + X) => 1 – [1,000(1 + X)]/[1,100(1 + X)] = 0.0909090…
1,200 + (1,200X) = 1,200(1 + X) => 1 – [1,100(1 + X)]/[1,200(1 + X)] = 0.083333…
1,300 + (1,300X) = 1,300(1 + X) => 1 – [1,200(1 + X)]/[1,300(1 + X)] = 0.0769230…

As you can see, you get the same results. g(x) is not only not linear, its not even the equation you want, resulting in thinking that crit damage suffers from diminishing returns and thus, is an inferior stat.

However, lets examine whats actually happening. Note; For simplicity, I’m just going to call base damage and power the same thing. Crit is roughly defined as a attack that has some probability to hit harder. And for more simplicity, we will just say that it hits 1.5 times harder. This means that if you can reach 100% crit chance, you are hitting just as hard as someone that has 1.5 times as much power as you (but no crit). Thus, if you are sitting at 100% crit chance and can get your crit damage up to 250%, you essentially are hitting just as hard as someone with two and a half times as much power (and no crit). So you can see that crit damage not only does not suffer from diminshing returns, but was (is) the second most powerful stat we have. Basically, each percent you have in crit damage is equivalent to that same number times of power you have.

1,000 + (1,000*0) = 1,000
1,000 + (1,000 * 0.5) = 1,500
1,000 + (1,000 * 1.5) = 2,500
1,000 + (1,000 * 2.5) = 3,500

Thus if you had a base damage/power of 1,000 and no crit, you would remain with 1,000 base damage/power. And with that same base damage of 1,000, you increase your base damage 2 and half times just by maxing your crit chance and gaining 100% more crit damage. Moreover (and this is the point) that max crit chance and 100% more crit damage is worth 1,500 base damage in this example.

This is why this patch hurt our dps so much.

guanlongwucaii.3162:

however, the 30/30/0/10/0 is worse than pre-existing meta 30/25/0/0/15 by a fair margin in the event that you have 0 might, so for reeaaaaallllyyy dysfunctional pugs you should still stick to the current meta build.

^This. I know you have said this twice now and I am just now responding, but I totally agree with this. However I actually think that the pugs dont have to be all that dysfunctional. In fact, I am pretty sure that 30/25/0/0/10 will beat out all other builds unless you are running a very organized and skilled group that can maximize might, fury and invuln with near 100% up time for each. (Though to be honest, I wont be doing the math for this until after finals, so right now this is just theory.)

For those curious as to how to calculate this, both might and invuln can be approximated using Poisson distributions.

Tree.3916:

If you read the assumption, I said “assume both builds do a base 10,000 dps.” That took into account the /time factor you are looking for.

Yes, I did read your assumption. Right after you accused another person of not knowing how to calculate DPS. Then I watched as you used the “Average Damage” forumla to assert your claims. Taken straight from the wiki;

Average damage = Base damage * ( 1 – Critical Chance ) + Base damage * ( Critical Chance * (Critical Damage + 1.5) )

Since you took the above formula and used “base 10,000 dps” in place of Base damage, thus erroneously replacing one variable for another, I assumed you didnt know the difference between Base damage and Base damage per second. Which are not the same thing. But for some reason, you treated them like the same thing. And since you gave no justification for this like you should have, the only conclusion left to jump to is that you are in error. (Which you are.)

Tree.3916:

To be technical, the equation I presented was “average expected damage per second.”

To be technical, no you didnt. For completeness;

Average damage = Base damage * ( 1 – Critical Chance ) + Base damage * ( Critical Chance * (Critical Damage + 1.5) )

Which I took straight from the wiki page. http://wiki.guildwars2.com/wiki/Critical_hit

Tree.3916:

Next, I chose my numbers for the sake of being round. If you change the 10,000 to 5,000 or 1,000 the conclusion is exactly the same.

Arbitrary. In the math world when we pick w/e values we want and the values dont really matter, we use the word arbitrary. And if this is in reference to your claim that critical damage is logarithmic when crit is held constant, and is thus a stat that suffers from diminishing returns, you are still wrong. In fact you can even independently verify this yourself using pen and paper assuming you know how to plot graphs in 3 dimensions. Or know how to use wolfram alpha. Spoiler alert: its just a plane. Which is linear, as I’ve already told you.

Tree.3916:

Lastly, you don’t even seem to understand what the numbers I typed express. The extra 1000 dps that build A receives is not Power. It was a 10% damage modifier, the sort you would get from Attack of Opportunity. A 10% damage modifier does indeed function in the manner I described. Critical damage functions in the manner I described. Adding 150 Ferocity to a build adds a critical damage of 10%. In that equation it would be expressed as a .1. This is how the game works.

Actually it is you that does not seem to grasp both the numbers or labeling that you are using. You are claiming that the arbitrary 1000 value picked is not power (base damage), but DPS. Even though you use the “Average Damage” formula to show your calculation. Which you justify by assumption. Should I copy paste the formula a third time?

Tree.3916:

I’ll be happy to help you clear up your confusions if you have any remaining.

You mean like the part where one of us confused the “AVERAGE DAMAGE” formula and treated as if it could be used to calculate expected damage per second. Or the part where one of us thought that critical damage behaves like an inverse exponential function?

Tree.3916:

Like I offered the previous confused fellow, I would rather answer your questions about how the game works than have to debunk your misconceptions like I just did.

Well frankly I would rather you stop pretending you understand other peoples “misconceptions” before understanding your own. You haven’t debunked any of my claims. In fact you haven’t even come close. You know what? Lets be clear what my claims are here so that you can debunk them.

Claim1: The “Average Damage” formula cannot be used to calculate DPS and can only be used to calculate… average damage.

Claim2: Critical damage does not have diminishing returns and is linear.

Tree.3916:

Let’s assume two builds, each has a 100% crit chance. both have no crit damage from gear. Both builds do a base dps of 10k. One build has a 10% dps modifier, and one build has 10% critical damage. According to you, these builds have identical DPS. Let’s see if that is correct…

Build A: (10,000 × 1.1) + ((10,000 × 1.1) *.5) = 16,500 dps
Build B: 10,000 + (10,000 * .6) = 16,000 dps

In most MMOs DPS stands for Damage Per Second and I would like to assume that is what is being discussed here (though it was never explicitly stated). And if it is… the above formula is not correct. That is the formula for expected total damage, which is no where near the same thing as DPS. If this was a DPS function, we would expect to see some sort of division in the expression (by units of time). The expected total damage formula is really a poor model to use when attempting to maximize your DPS in my opinion.

Tree.3916:

As you can see, the base damage modifier is significantly better, even when critical damage is low. But critical damage has diminishing returns. What this means is, as you add more critical damage it does not increase your DPS 1:1. at a certain point, doubling your critical damage does not approach a doubling of your DPS. Let’s run the numbers with very high critical damage.

Build A: 11,000 + (11,000*1.6) = 28,600 dps
Build B: 10,000 + (10,000*1.7) = 27,000 dps

The gap has now widened to 6%. The gap will continue to widen the more DPS modifiers you add (since they are multiplicative) and the more critical damage you add. These are the basics of DPS theorycraft in this game, I will be happy to help you with any clarifications you need.

Critical damage is not logrithmic and therefore does not have diminishing returns (given 100% critical rate). Its linear, just like the above formulas that you’ve posted. The reason critical damage looks so undesirable in your above example is that you have made it (arbitrarily) very expensive; 1000 power (base damage) for 0.1 additional critical damage. This is no where near an accurate representation for stats exchange in this game.

Honestly, talking about critical damage while holding crit chance constant really negates the potential of critical damage. Crit dmg really shines when comparing builds with low crit, high power vs high crit, low power. its in those cases when crit dmg can really show its weight.

Mystic.5934:

https://docs.google.com/file/d/0B8xrv7IFopOqVGRpQ1JmMnZuaWs/edit?pli=1
TLDR: It is faster to restart fractals than play dredge fractal only if dredge takes more than ~2.3 times the time of the other possibilities. If the others take you 30 min, dredge would need to take 69+ minutes to merit restarting the run.

This was an excellent read and I very much enjoyed it!

The only thing that I would have added is boundary situations that might appear that a fractal restart is merited. For example running Swamp, Snowblind and Volcanic. Given a decent experienced group, it is not uncommon to run these in 15 to 20min each. If these could be run in those times and in that order then this certainly would appear to break the model.

But it doesnt. In order for this to go against your model, a group would have to guarantee that the next restart would be those fractals in that order. And seeing that the odds of getting those are (1/5)*(1/5)=(1/25)=0.04 chance, it would be silly to risk restarting your fractal for that 4% or 8% chance (that is if you can include Aetherblade in your less that fifteen min completion list). And if you did restart, it would be wise to consider the time it would take to re-roll and such.

A few days ago I switched ISPs and GW2 is unplayable. Although my old internet was slow, I never DCed and only rarely had spikes. Played today for the first time and GW2 is unplayable. Ive been experiencing everything from small lag spikes to “Server is unable to connect to server” errors.

Ensign.2189:

In a classical sense, for a given population mean, the mean sample proportion will be a binomially distributed random variable. The population mean is not a random variable; it is a (unknown) constant.

Weird. You used classical inference statistics to define your population mean, then used Bayesian statistics to make your inferences. I prefer the Bayesian approach, but to each their own.

Ensign.2189:

Not at all; in fact I find that Gaussian approximations do a poor job when you’re looking at events where Pr(x) is close to 0 or 1 as the true distribution has significant skew.

Uhhh… wut? This would be significant news to the statistical/mathematical community if such a find where true. Gaussian approximations are used frequently for binomial random variables. Recall that the binomial distribution converges in distribution to the normal distribution as our proportion goes to zero and our sample size goes to infinity. A quick google such provided a much more rigorous help of this notion; http://math.stackexchange.com/questions/435574/convergence-of-binomial-to-normal

Ensign.2189:

I took a prior of a uniform distribution on [0,1] for Pr(x) and each data point is a Bernoulli trial; this is easy to compute explicitly, as the posterior distribution is simply a beta distribution. Integration over the beta distribution generates credible intervals for my belief about the true value of Pr(x).

Integration over the beta dist. would only give you the cdf of the beta. So… of course the true value of our population proportion will be there. My guess is that you were attempting to explain how you found the posterior distribution through integration, but this is a hairy process that takes a bit. Then you still have to check for bias and sufficiency. I usually just use the proportionality of the prior and the proiri. That is if I have to find a Bayes estimator; I would much rather find the MME or MLE. But regardless, why did you not just find the expected value of your found beta?

Ensign.2189:

I’m working in an inference framework and estimating model parameters not performing hypothesis tests.

If its been a while since youve studied studied statistical convergence or are not in the math feild, I could totally understand why you think that point estimates do not need power tests. Most people just ask a computer to find some 10,000 sample means and call it good, but we are not all that lucky. Power tests are essential in finding the probability of type II errors and how large our sample sizes need to be to avoid those errors.

Ensign.2189:

oilstorm.1748:

Perhaps this is common knowledge that I am unaware of, but where did you get these probabilities and much more importantly, how do you know they are right?

Probabilities are inferred from a large sample (~8000 exotic forges) of forge output, combining my own data with a couple other sample sets I have high confidence in (such as the 10,000 gold forging streamed and publicized on reddit recently).

I know they are ‘right’ in as much as a data set of that size tells you a lot about the credible range of parameter values; for example I can tell you that based upon that data set I can say with 99.9% confidence that the precursor rate is no better than 1 in 113, and with 99.996% confidence that it’s no better than 1 in 100.

As for there being 12 commons of each type, and that the forge only kicks out 7 named, non-precursor exotics of each weapon type; that should be common knowledge, and if it isn’t someone should add it to the wiki.

Interesting. So if I am understanding you correctly, you are arguing that the mean sample proportion is binomially distributed and therefore the mean population proportion is also a binomial distribution? And that since your mean sample proportion times the sample size is less than (1-mean sample proportion) times your sample size, you invoked the Central Limit Thrm, right? I completely agree with your analysis at this point and your work is (mostly) spot on. Nicely done.

However, how do you know the mean population proportion and the population variance that allows you to do a mean proportion comparison test? Similiarly, at what power where your tests carried out? I am asking because your post lacked the rigor that I am looking for (I am naturally a curious person). I know it is not common for people to understand most of the terminology, jargon and whatnot of this subject, but please feel free to be more technical.

Lastly, how do you know that the common exotics drop with equal probability?

Ensign.2189:

Essence Snow.3194:

I think what they are getting at is that it’s not pure rng. It’s modified so that certain occurrences happen more than others.

It’s pure RNG with different probabilities of events. For each weapon type, there are 12 commons (7 pearls, legionnaire, tribal, etched, ceremonial, orrian) which drop with equal probability and collectively drop around 5/6 of the time, 7 named exotics which also drop with equal probability and collectively account for a bit over 16% of drops, and then the precursors – either the one for the weapon type, or one of the two greatswords picked randomly – which pops out somewhere between 1/125 and 1/200th of the time (roughly).

Perhaps this is common knowledge that I am unaware of, but where did you get these probabilities and much more importantly, how do you know they are right?

Ashabhi.1365:

Myth #(whatever): The more you throw in, the better your chance…

Many folks have a misconception of what RNG is an how it works. RNG is getting s a specific result out of (X) possible combinations. So, if one throws in the same exact ingredients, the result is going to be the exact same ratio for each attempt.

(Much Simplified) Example: If you roll a regular 6-sided die, you have the exact same chance (1 in 6) of getting a 4 every time you roll the die. The number you rolled is not removed from the rotation. You have the potential to roll a 4 every time you roll, or never. It’s the same chance on roll #1 as it is on roll #401.

If you extrapolate to the chances of making a precursor, you have the exact same chance of getting one on your first attempt as you do on your 9000th. the odds simply don’t get better the more you try.

I mean no disrespect, but your conclusion may or may not be correct. If the rate of return of precursors/exotics is truly uniform, then you are spot on. No matter the combination of rares/exotics thrown in the forge, you have some X percent chance of receiving a precursor that stays the same.

However, if this rate of return is not uniform (meaning it does not act like a single thrown die, but perhaps like a pair of dice or something else entirely) then the above is false. For example, mentioned in the thread is the godskull exploit. Before this was patched, throwing in some combination of godskull weapons, lvl 65ish rares, (lets call this combination k) with other rares would give higher chances of exotics (and therefore precursors). Thus given a parameter of having k godskull, our chance of getting some precursor, X percent, was increased. Since it was/is possible to vary our X percent with some parameter (which there may be others out there), to me it shows that the distribution is not uniformly uniform, (boy that sounds weird) but something else like a Weibull, Gamma, or Beta, where our parameter is not based on some number, but rather some number of objects.

I do want to give credit where credit is due though, and you deserve much credit for your post. Your post was debunking the myth that the more you throw in the better your chances, which is false since the actions in the forge are (supposedly) independent of each other. Just to add what was stated, it is totally false that your next forge will be good/bad simply b/c your last n number of forges has been bad/good.

Edit: Apparently GW2 forums dont use alt codes.

Nike Porphyrogenita.8137:

oilstorm.1748:

Epic.3950:

This is bugging me for a while and I want to know if anyone has mathed it out.

No, no one has “mathed” it out. No one has come up with a correct mathematical model to calculate damage per second. The best anyone has done are just (poor) estimates that calculate the expected value of damage called “Effective Power”.

Actually, we have “mathed it out.” GK and Zehlyn have created actual DPS calculators with actually DPS rotations taken into consideration, and Effective Power is not part of them. You have an arrogant tone in your post which perhaps you should moderate in light of the fact that you aren’t as current with the state of theorycraft in the game as you think you are.

Actually, no, you haven’t “mathed out” anything, though you can pretend all you want. In fact, you haven’t even attempted to counter any of points regarding the mathematical flaws I pointed out. And no, I wont be moderating my tone, especially considering I wasn’t even talking to you, though I can certainly fill YOU in. The state of the “theorycrafting” in this community is where it has been for a long time. Inaccurate. People "guestimating"everywhere saying, “This build is the best by roughly X%” and “Your build is inferior by Y%” using napkin math, mental math or rudimentary spreadsheets. People constantly making claims that cannot be verified. You couldn’t even make a proper rebuttal to mathematical claims. You are more than welcome to try if you’d like, but something tells me that your not all that mathematically mature.

By saying “mathed it out” I am defining it as actually using precise, correct math that accurately models damage per second. As I said in my post, a model needs to be verifiable and thus far, none have been. Come back with such a model, then you can say “we have mathed it out”. (And why are you taking credit for the work of GK/Zehlyn anyways? Were you apart of that “we” that you mentioned?)

Nike Porphyrogenita.8137:

FWIW, the reason people use Effective Power is because its much much faster and simpler to calculate. It’s useful only for comparing builds with the same weapons within the same profession so it is clearly quite limited; however, once you are aware of its limitations its still pretty useful for threads such as this where misinformation by know-nothings is plentiful.

So because Effective Power is quick and simple, its justified in being used? Yes, you are correct, if we ignore all the many limitations, then yeah, your statement is trivially true. Actually, let me put that statement in layman’s terms for you; If we pretend that EF model is actually useful, then it is amazing work. You put numbers in and it spits numbers out. Your justification and use of EF is more than enough information to gauge the state of the theorycrafting of this game.

PS. I dont think you are as current of math models that are useful as you think you are.

Epic.3950:

This is bugging me for a while and I want to know if anyone has mathed it out.

No, no one has “mathed” it out. No one has come up with a correct mathematical model to calculate damage per second. The best anyone has done are just (poor) estimates that calculate the expected value of damage called “Effective Power”. All estimates based on data are also given with error estimates (for those unaware). Those that are based not on data, but on theorems, need no estimates but are required to provide parameter estimates. The so called “Effective Power” math model has neither. In fact, if I recall correctly, EF attempts to optimize your damage without even using calculus. If you have taken calculus you will know what I am talking about. Its one of the central themes of the course.

The EF model has room to be greatly improved (if not scrapped altogether). For example, did you know that the trait “Stick and Move”, if traited, is treated as if it has 100% uptime? Yet it doesn’t have 100% uptime. (If you wanted to treat it as a multiplier, you would need to find some kind of invariant measure, of which there are several ways to do so.) Another example of a weakness of the EF model that comes to mind is condition damage; its simply ignored. Its for this reason that many people think that the longbow (LB) is more effective than rifle, when in reality sometimes it is and sometimes it isnt. Even if your build is not centered (or even care about) condition damage, it doesn’t change the fact that you still do some condition damage and needs to be accounted for when calculating expected damage done. (Hint; its not as negligible as some people think. Like the invariant measure mentioned above, there are multiple ways to calculate the damage. Its a Poisson process.)

There are several ways to know if some one has “mathed” it out correctly Epic. Any model should be verifiable using different methods, meaning that you will be able to reach the same results using completely different methods. I should be able to take partial derivatives and arrive at the same conclusion as a spreadsheet if both the derivatives and spreadsheet are correct. And if either of those two are correct, I should be able to simulate those results in a program like MatLab or Maple. Just this reason alone negates the usefulness of the EF model and why it is ineffective. No error margins and no parameters mean there is little chance of correctly arriving at the same conclusions using different methods. (Note; if EF were accurate, you would know the equilibrium of power and precision since, if defining only those two, the formula for describing them is a second degree polynomial. Meaning, unless the vertex is not in our domain, there is a point at which the two stats are equally powerful. So, for most zerker armor set ups there exists some precision armor set up has the same effective stats. Just food for thought.)

For those interested in what the correct math is in understanding min/maxing your stats or DPS or any metric, here is a blog that I love.

http://www.sacredduty.net/

Its a WoW blog, but math is universal and many of the principles apply to this game.

Hey Anet,

I am extremely unsatisfied with the fractal changes. I have been playing PVE in MMOs for many years now and I have never encountered a situation like this. I can hardly believe what I am reading; you are RESETTING the progression I have made in fractals? The personal goals I have set and accomplished are now being wiped clean?

This is absolutely unacceptable. I understand the need for change, as all MMO evolve over time. I recall when WoW was removing factions/achievements/quests/storylines from their original game to make way for more content but Blizzard did not remove my faction rep, my achievement points, my quest history or completed story line to do so. So to have my PERSONAL reward level reset is extremely unnerving. Now that I know that having (what seemed like permanent) achievements being erased is on the table to be done whenever you see fit, I am questioning my loyalty to this game. I am not saying “Thats it, I’m out”, but I can’t say my resolve to remain with GW2 is as strong as before I heard about these changes. I dont enjoy having to worry if my accomplishments are safe or not in the MMOs I play.

I have no idea what will happen with these changes, good or bad, but I can tell you that not matter what the outcome, I will never attempt to get my friends or family to play GW2 again. Why on earth would I want to convince them to play something I may not play, simply because you, Anet, want to toy around with my past accomplishments? Let’s suppose the proposed changes are the most amazing refinement to a game ever and the mere mention of the name “ArenaNet” causes angels to cry forth from secret hiding spots. Who is to say that future personal level rewards wont be wiped due to unseen circumstances? Or that perhaps a zone needs to be redone, requiring all hearts and quests need to be redone for completion?

In short, these changes to FOTM are unsatisfactory and will ultimately hurt the longevity of the game we all care about. Please remedy this Anet.

Larynx.2453:

Pros: One of the highest single heals in the game when used at 3 adrenaline
Best heal to counter burst

Cons: Must be used at 3 adrenaline
Must be used actively, unlike Healing Signet
About 25% less overheal HPS than Healing Signet if used optimally

The biggest thing is the active portion. A passive signet has the benefit of never overhealing and never go to waste. It ticks during CCs, though it is significantly more vulnerable to poison. An active ability requires more CD management, and it can be countered very easily with CC. It is possible to overheal as well, though on a warrior this is less of an issue.

The only reason a warrior should switch from Healing Signet is if Healing Signet is nerfed or the meta shifts and super high burst builds become popular again. Right now damage is very sustained, and any burst a warrior encounters can be dealt with with just using CDs.

I am going to assume you are speaking only about PVP. Healing Surge contributes to DPS when used correctly in PVE allowing free eviscerates, kill shots, arcing slice or w/e burst you have with your weapon every 30 seconds.

Stop eating your bread with the butter side down!

I am and have been, sitting at 82/83 hints completed. The “Mailbox is full” under Social tab will not activate despite my best efforts. I have deleted all sorts of mail from guildies and friends, so know that I have the requisite met. This is a problem because not only am I being denied 10 achievement points, but its taking up a the top slot for the “Nearly completed” watchlist.