Guild Wars 2

Hi,

I did some serious theorycrafting and just wanted to let you know about my results and like to know whats your experiences with the damage from thief compared to warrior is.

Thief build: here
Warrior build: here

Result:
https://docs.google.com/spreadsheet/ccc?key=0AlOXvenKdNo2dEZXWEpmUXNuMEhTUWI3LWQ2UG40VVE#

Tested with:

• full berserker equip
• ascended equip with power infusions
• 25 might stacks
• 25 bloodlust charges
• Buff food: Superior sharpening stone and +100 power food
• CnD + Backstab every 6 seconds (4 initiative need 5,32 seconds to recharge and 6 seconds for 2 dagger 1 chains)

To get the times for the attack chains i recorded around 1 minute with 60 fps for each chain to see how many chains i can do in this time. I worked very accurate to count every single frame.

For the damage values i used the ingame tooltop description and added modifiers that werent applied in the tooltip by myself (i.e. the +5% dmg when wielding an axe in offhand).

I also tested:

• sword auto attack
• pistol whip
• warrior greatsword auto attack
• warrior hundred blades
• engineer bomb kit auto attacks

Top dps:

• 1. Thief CnD + Backstab – over 2 seconds
• 2. Warrior Hundred Blades – over 3.6 seconds
• 3. Warrior Axe chain – over 3.6 seconds
• 4. Thief Pistol Whip – over 2.68 seconds
• 5. Thief sword chain – over 2.49 seconds
• 6. Thief dagger chain – over 2.06 seconds
• 7. Engi bomb skill 1 – over 0.93 seconds
• 8. Warrior great sword chain – over 2.49 seconds

Go to Heart of the Mists, and use Steady weapons for reliable results.

the pvp/mists values are NOT the same as pve ones !

Quickfoot Katana.8642:

Go to Heart of the Mists, and use Steady weapons for reliable results.

cant do it for the max dps test since the skills scale different with power.
And you cant have as much power with sPvP gear.

Maybe i should add that i checked this for PvE since in sPvP there are way more factors.
there is alot of content in PvE where you can simply stand next to a boss and attack.

edit: the “ascended gear” and “buff food” should tell you thakittens PvE anyway

“CnD + Backstab every 8 seconds (6 initiative need 8 seconds to recharge)”

Your test just became invalid. 10 into shadow arts and you’ll never need to stop backstabbing (if you get 2 cnd off in a fight, having 10 in shadow arts is better than 15 in trickery initiative wise). Thief also has a lot more damage modifiers than what you stated so it seems like you’re intentionally nerfing thieves in effort to make warrior dps look better all around. They do trounce us in melee aoe though. Cnd, backstab takes 1 second as well. Cnd in pvp also does something like 50% the dmg it does in pve and wvw.

I’m pretty sure Thief doing slightly less damage than other classes is by design because of the Initiative system (though I’m not convinced it’s a good design). Interesting, nevertheless.

I’d love to see some comparisons of the ranged weapons too, especially the Rifle vs. Pistol, which is a debate that pops up fairly frequently.

i think that if you change your build to have more ini regen you can backstab more and do more damage…

randomfightfan.4091:

“CnD + Backstab every 8 seconds (6 initiative need 8 seconds to recharge)”

Your test just became invalid. 10 into shadow arts and you’ll never need to stop backstabbing (if you get 2 cnd off in a fight, having 10 in shadow arts is better than 15 in trickery initiative wise). Thief also has a lot more damage modifiers than what you stated so it seems like you’re intentionally nerfing thieves in effort to make warrior dps look better all around. They do trounce us in melee aoe though. Cnd, backstab takes 1 second as well. Cnd in pvp also does something like 50% the dmg it does in pve and wvw.

I calculated with all the damage modifiers you get (10% when condition, when over 6 initiative and 5% dagger damage).

With shadow arts you have a good point.

i reworked my calculations and will edit my post

result:
as long as the enemy is above 50% the damage its the same but the warrior can be brain afk.
below 50% health the thief deals about 20% more damage (when every backstab hits from behind)

Quickfoot Katana.8642:

Go to Heart of the Mists, and use Steady weapons for reliable results.

That won’t work because condition damage will have a far greater effect on the damage output. then with normal weapons

crouze i applaud your enthusiasm but you have a few problems in your test hypothesis. cnd+bs is relative to opponent being mobile vs rooted.

also 25 stacks of might is very hard to maintain so i know you cant do many tests and keep that up for more than a few secs…thief can only get 22 might stacks max by self and not sure about warrior.

.4% dmg difference is less than 1% dmg difference and not worth mentioning unless you mean something else like 4% or idk.

IDEAS:

do 10 auto attack chains for each with full zerk gear no buffs. if each class has the same buffs the dmg goes up accordingly. average out the dmg per hit. hits per kill. pick a MOB in WVW or PVE so both dmgs can be correlated/compared.

drop all buffs. they are not needed as if the classes have the same base stats the dmg will be increased accordingly. %’s is what matters not total dmg.

Also using traits like 20% more dmg under 50% is one that each class has but in diff ways. just keep that out of it as it is a dmg modifier and isnt directly related to BASE dmg. base dmg tells you whom really does more dmg.

You use a thief combo to mix in your results and not a warrior combo like SOR and the 2 KDs plus the 2 whirls i always see them do. since you did not do this you will have skewed results and must start over.

most importantly you must do more than theory craft. i had a guy in my guild claim i was doing 20k backstabs in my video on a naked person. so then i made another video of a naked person and i was doing 67.5k dmg. sore loser he was but anyway.

anyway point is:

STEP 1—-Ask a Question
STEP 2—-Do Background Research
STEP 3—-Construct a Hypothesis
STEP 4—-Test Your Hypothesis by Doing an Experiment
STEP 5—-Analyze Your Data and Draw a Conclusion
STEP 6—-Communicate Your Results

EDIT: also test more than once.

frans.8092:

Quickfoot Katana.8642:

Go to Heart of the Mists, and use Steady weapons for reliable results.

That won’t work because condition damage will have a far greater effect on the damage output. then with normal weapons

AND DMG is capped so for thieves its actually quite biased/unfair.

Travlane.5948:

.4% dmg difference is less than 1% dmg difference and not worth mentioning unless you mean something else like 4% or idk.

It shows you that there is no big difference. I also could write “it deals about the same amount of dps”.

Travlane.5948:

drop all buffs. they are not needed as if the classes have the same base stats the dmg will be increased accordingly. %’s is what matters not total dmg.

Weapons skill damage scales different with power. They have different modifiers.
So 2 skills that deal the same damage with 916 power can have a big difference when you have a lot of power.

Travlane.5948:

Also using traits like 20% more dmg under 50% is one that each class has but in diff ways. just keep that out of it as it is a dmg modifier and isnt directly related to BASE dmg. base dmg tells you whom really does more dmg.

These modifiers is what makes a difference. If both deal the same “base damage” then the profession with more / higher modifiers deals more damage

They really need to make PvE test dummies with the following qualities:
– Killable
– High health
– Can be crit

This would really help give better test data and allow more solid conclusions to be made.

I have a few problems with what you posted.
1. I know for a fact thief dagger auto attack takes longer than 2.06 seconds. I saw it posted somewhere before when I was looking for coefficient values for dagger attacks.
2. D/D Thieves do not use executioner, they use hidden killer.
3. Why are you only posting results and no math for people to double check? This thread is invalid until you do.
4. Warrior will always win the DPS race if it ever hits more than 1 target and all things are equal.
5. What traits are you even using? There’s a huge difference in damage output if you build full glass cannon compared to doing anything else, not listing any trait setups is disingenuous. You’re going to make some poor all soldiers warrior think he has the same DPS as a full on glass cannon thief.

I dont mean to be a Debbie Downer, but you really need to be more specific and show your work.

Dahkeus.8243:

They really need to make PvE test dummies

This would really help give better test data and allow more solid conclusions to be made.

No it wouldnt… only a fool would sit in a game and try to determine average damage by hitting stuff when the equations for damage calculations are known and would be many, many, many hours faster. Does a parser for GW2 even exist or are you going to sit there and write down every value so you can add them up and divide?

@Wish:
1) no, It takes 2.05sec use fraps and check it out
2) yeah noob thieves are using Hidden Killer, you are right. Thieves with an average crit chance of 61% are using executioner because it deals more dmg
4) You are right! There are so many boss fights in gw2 where you have to fight against two bosses without adds that die instantly when you look at them…. uhm… wait.. no…

wish.1027:

Dahkeus.8243:

They really need to make PvE test dummies

This would really help give better test data and allow more solid conclusions to be made.

No it wouldnt… only a fool would sit in a game and try to determine average damage by hitting stuff when the equations for damage calculations are known and would be many, many, many hours faster. Does a parser for GW2 even exist or are you going to sit there and write down every value so you can add them up and divide?

…I think you’re misunderstanding.

With the right dummies, all a player would need to do to test dps of different specs/classes would be to time how long it takes to kill a dummy, then compare kill times.

Anubarak.3012:

@Wish:
1) no, It takes 2.05sec use fraps and check it out

Why dont you post a video and show us. You know, prove what you claim to be correct. The burden of proof is on you as you’re the one making the claim.

Anubarak.3012:

2) yeah noob thieves are using Hidden Killer, you are right. Thieves with an average crit chance of 61% are using executioner because it deals more dmg

I’ve done the math for hidden killer vs executioner and I can tell you the point between crit chances where one is more effective than the other based on damage done from in-stealth attacks, but please, post the math in this thread and show us all how executioner is always superior at all points in time forever.

Anubarak.3012:

4) You are right! There are so many boss fights in gw2 where you have to fight against two bosses without adds that die instantly when you look at them…. uhm… wait.. no…

So the only thing you ever fight is bosses, and only 1 mob at a time at all times, because those bosses never have adds ever. Seriously? That sounds like some super pro-level PVE ownage there. I better just take my lame self back to PVP before I look like a complete fool in this thread.

Dahkeus.8243:

wish.1027:

Dahkeus.8243:

They really need to make PvE test dummies

This would really help give better test data and allow more solid conclusions to be made.

No it wouldnt… only a fool would sit in a game and try to determine average damage by hitting stuff when the equations for damage calculations are known and would be many, many, many hours faster. Does a parser for GW2 even exist or are you going to sit there and write down every value so you can add them up and divide?

…I think you’re misunderstanding.

With the right dummies, all a player would need to do to test dps of different specs/classes would be to time how long it takes to kill a dummy, then compare kill times.

Sure, as long as the dummies take an hour or two to kill. Otherwise Im going to argue small numbers and that math is better. And at that point where you fight a dummy for an hour or two, math is faster. If they die super fast, youd have to kill them for an hour or two, then average all the times together to get an idea of your real time. Because sometimes you’ll get all crits, and sometimes you’ll get no crits, and sometimes your crits will proc on the bigger hits and not the smaller ones, etc. You’d have to do it a lot so the numbers eventually balance themselves out, then add all the times together, and divide by how many times you killed it o figure out your true time. So youre still doing math anyway even with the test dummys.

wish.1027:

Anubarak.3012:

2) yeah noob thieves are using Hidden Killer, you are right. Thieves with an average crit chance of 61% are using executioner because it deals more dmg

I’ve done the math for hidden killer vs executioner and I can tell you the point between crit chances where one is more effective than the other based on damage done from in-stealth attacks, but please, post the math in this thread and show us all how executioner is always superior at all points in time forever.

Noone said, that Executioner is always better.

According to this page, the auto chain + C&D has a coefficient of 3.76 and BS has 2.4.

The critical damage modifier is 1 + critdamage x critchanche. We assume 170% critical damage. Since executioner kicks in at 50% hp, we treat it as a 10% damage boost.

Let C be your critical chanche. We want to determine the value of C, where both traits are equal.

[Executioner] 1.1 x (1 + C x 1.7) x (3.76 + 2.4) =
[Hidden killer] (1 + C x 1.7) x 3.76 + (1 + 1 × 1.7) x 2.4

http://www.wolframalpha.com/input/?i=1.1+*+%281+%2B+C+*+1.7%29+*+%283.76+%2B+2.4%29+%3D+%281+%2B+C+*+1.7%29+*+3.76+%2B+%281+%2B+1+*+1.7%29+*+2.4

C = 0.675 (=67.5% critical hit chanche)

I have more than 67.5%, for my build executioner is better.

panicbutton.1053:

I’m really interested in seeing p/p compared to these. If you have a moment can you try speccing in anything that raises pistol damage, run power/precision, with full stack of bloodlust?

Pistol skill 1 : ~0.88 seconds
Unload : ~2.28 seconds

Unload dps: 3% more then sword chain dps
pistol 1 + unload rotation: 77,4% of the sword chain damage.

bleeding damage with the 25 might stacks included. no initiative regeneration included.

might boost the damage a little with initiative on crit

Molch.2078:

According to this page, the auto chain + C&D has a coefficient of 3.76 and BS has 2.4.

The critical damage modifier is 1 + critdamage x critchanche. We assume 170% critical damage. Since executioner kicks in at 50% hp, we treat it as a 10% damage boost.

Let C be your critical chanche. We want to determine the value of C, where both traits ar equal.

[Executioner] 1.1 x (1 + C x 1.7) x (3.76 + 2.4) =
[Hidden killer] (1 + C x 1.7) x 3.76 + (1 + 1 × 1.7) x 2.4

http://www.wolframalpha.com/input/?i=1.1+*+%281+%2B+C+*+1.7%29+*+%283.76+%2B+2.4%29+%3D+%281+%2B+C+*+1.7%29+*+3.76+%2B+%281+%2B+1+*+1.7%29+*+2.4

C = 0.675 (=67.5% critical hit chanche)

I have more than 67.5%, for my build executioner is better.

when the base crit modifier is 1.5 and you have 30 in critical strikes, how can you have less then 180% crit dmg?

And the wolframalpha link does not work for me

crouze.3078:

Molch.2078:

According to this page, the auto chain + C&D has a coefficient of 3.76 and BS has 2.4.

The critical damage modifier is 1 + critdamage x critchanche. We assume 170% critical damage. Since executioner kicks in at 50% hp, we treat it as a 10% damage boost.

Let C be your critical chanche. We want to determine the value of C, where both traits ar equal.

[Executioner] 1.1 x (1 + C x 1.7) x (3.76 + 2.4) =
[Hidden killer] (1 + C x 1.7) x 3.76 + (1 + 1 × 1.7) x 2.4

http://www.wolframalpha.com/input/?i=1.1+*+%281+%2B+C+*+1.7%29+*+%283.76+%2B+2.4%29+%3D+%281+%2B+C+*+1.7%29+*+3.76+%2B+%281+%2B+1+*+1.7%29+*+2.4

C = 0.675 (=67.5% critical hit chanche)

I have more than 67.5%, for my build executioner is better.

when the base crit modifier is 1.5 and you have 30 in critical strikes, how can you have less then 180% crit dmg?

And the wolframalpha link does not work for me

Strange, it works just fine for me. But you can copy – paste the formula, but you need to replace the x with *. Or just believe it, it would be kinda stupid to lie about that, eh?^^

In this formula, the base crit modifier is 0.5. It is the additional damage the crit provides, multiplied with the probability of that bonus damage. Thats the reason there is a 1 in (1 + C x 1.7)

Edit: The fromula is accurate for 0 and 100% crit chanche (easy to proof) and the damage multiplier grows linear in C, therefore it should be correct. If you disagree, please post your formula to calculate the overall critical modifier.

ah, then its +170% crit dmg instead of 170% crit dmg

and i see no problem with your formula

Molch.2078:

The critical damage modifier is 1 + critdamage x critchanche.

I dont think this is correct, or I’m misunderstanding what you’re saying. It shoud be:
(1 * none-critchance) + ((1.5 + critdamage) * crit chance)
With the crit chances simply being looking at you crit chance and putting a decimal in front of it. Like 44% crit chance would be .44 and then none-crit chance would be .56 Crit damage would be converted to a decimal in the same way, 50 crit damage would be .5 112 crit damage would be 1.12, etc. So 81% crit damage and 44% crit chance would look like this:
(1 * .56) + ((1.5 + .81) * .44) = 1.5764
Using your equation it looks like this:
1 + .81 * .44 = 1.3564
or this maybe?
1 + (1.5+.81) * .44 = 2.0164
But ya, theyre all different numbers.

Molch.2078:

Since executioner kicks in at 50% hp, we treat it as a 10% damage boost.

Since I laugh at thief PVE how about we think about PVP, and how people have heals they use at 50% Hp and count executioner as like 6.6% damage instead. Or realize that the warrior is going to be hitting multiple targets a lot since it’s PVE and it auto-wins.

crouze.3078:

Travlane.5948:

.4% dmg difference is less than 1% dmg difference and not worth mentioning unless you mean something else like 4% or idk.

It shows you that there is no big difference. I also could write “it deals about the same amount of dps”.

Travlane.5948:

drop all buffs. they are not needed as if the classes have the same base stats the dmg will be increased accordingly. %’s is what matters not total dmg.

Weapons skill damage scales different with power. They have different modifiers.
So 2 skills that deal the same damage with 916 power can have a big difference when you have a lot of power.

Travlane.5948:

Also using traits like 20% more dmg under 50% is one that each class has but in diff ways. just keep that out of it as it is a dmg modifier and isnt directly related to BASE dmg. base dmg tells you whom really does more dmg.

These modifiers is what makes a difference. If both deal the same “base damage” then the profession with more / higher modifiers deals more damage

when comparing classes you should only compare 1h weapons vs 1hnd weapons etc and so on and so forth. as its different for each class you are comparing class vs class not weapon vs weapon. so weapon choice should be slightly irrelevant. choose the highest dmg ewapon for each class then compare it. if u choose 2 classes with the same dmg weapons and same armor etc it should be the same. but i THINK u are trying to see the class differentials in dmg department. so weapon shouldnt matter

people use formulas alot on here but theres other factors i cant mention that go into it….try find me a formula that does 67.5k dmg backstab….just for instance. i won a bet on this bc a guy in my guild’s math was not adding up :p said it was possible. i said experience in the field dictates other wise.

Molch.2078:

In this formula, the base crit modifier is 0.5.

1 + (.5+.81) * .44 = 1.5764
That works.

Compared to how I write it out
(1 * .56) + ((1.5 + .81) * .44) = 1.5764
hich I got from another game entirely and how Ive always seen it done, your way is easier but less noob friendly. lol

Molch.2078:

I have more than 67.5%, for my build executioner is better.

So basically what youre saying is that if you require 2 full chains of CnD > autoattack > BS to kill something, you would need 67.5% crit chance for executioner to win. But what about when the second BS kills the thing youre fighting if it crits, and doesnt kill it if it doesnt crit? What if you use CnD > BS, then dodge away and get no autoattacks in, then CnD again then BS?

I think youre looking at this entirely wrong. What you should be determining is how much of your damage needs to come from your ‘from stealth’ attacks compared to whatever youre fightings HP for hidden killer to win compared to executioner. Like, if you have X crit chance/damage/power, your backstabs need to total xx% of targets HP for hidden killer to win. Then people can determine themselves through their own play style which comes out ahead. Some people do a lot less attacking outside of CnD and BS compared to others who melee people down with autoattacks and HS’s. And I really believe thats what we should be talking about to determine which is better, not calculating some magic crit chance that has no baring on playstyle. I guarantee you, even with 68% crit chance you can make hidden killer win for damage output over executioner, just dont do much outside of CnD and BS and maybe even throw in a blinding powder or hide in shadows instead of a CnD.

Travlane.5948:

people use formulas alot on here but theres other factors i cant mention that go into it….try find me a formula that does 67.5k dmg backstab….just for instance. i won a bet on this bc a guy in my guild’s math was not adding up :p said it was possible. i said experience in the field dictates other wise.

Thats totally possible, you just need like 10,000 power!

wish.1027:

So basically what youre saying is that if you require 2 full chains of CnD > autoattack > BS to kill something, you would need 67.5% crit chance for executioner to win. But what about when the second BS kills the thing youre fighting if it crits, and doesnt kill it if it doesnt crit? What if you use CnD > BS, then dodge away and get no autoattacks in, then CnD again then BS?

When you fight a boss or sth. you use like 100 chains and one final BS, this won’t make a difference.

I think youre looking at this entirely wrong. What you should be determining is how much of your damage needs to come from your ‘from stealth’ attacks compared to whatever youre fightings HP for hidden killer to win compared to executioner. Like, if you have X crit chance/damage/power, your backstabs need to total xx% of targets HP for hidden killer to win. Then people can determine themselves through their own play style which comes out ahead. Some people do a lot less attacking outside of CnD and BS compared to others who melee people down with autoattacks and HS’s. And I really believe thats what we should be talking about to determine which is better, not calculating some magic crit chance that has no baring on playstyle. I guarantee you, even with 68% crit chance you can make hidden killer win for damage output over executioner, just dont do much outside of CnD and BS and maybe even throw in a blinding powder or hide in shadows instead of a CnD.

So, you’re talking about PvP, while I talk about PvE? Well, gg then

But in PvE you will complete your chain 90-100% of the time.

wish.1027:

Molch.2078:

In this formula, the base crit modifier is 0.5.

1 + (.5+.81) * .44 = 1.5764
That works.

Compared to how I write it out anyway
(1 * .56) + ((1.5 + .81) * .44) = 1.5764
Which I got from another game entirely and how Ive always seen it done, your way is easier but less noob friendly. lol

To be consistent:
Your formula is

(1 * (1 – critical chanche)) + ((1 + the thing I call critical damage) * crtical chanche) =
(1 * (1 – cc)) + ((1 + cd) * cc) =
1 – cc + cc + cd * cc =
1 + cd*cc

q.e.d.

But you are right, it’s not that easy to understand.

wish.1027:

Anubarak.3012:

@Wish:
1) no, It takes 2.05sec use fraps and check it out

Why dont you post a video and show us. You know, prove what you claim to be correct. The burden of proof is on you as you’re the one making the claim.

Actually you made the claim:

1. I know for a fact thief dagger auto attack takes longer than 2.06 seconds. I saw it posted somewhere before when I was looking for coefficient values for dagger attacks.

Molch.2078:

wish.1027:

I think youre looking at this entirely wrong. What you should be determining is how much of your damage needs to come from your ‘from stealth’ attacks compared to whatever youre fightings HP for hidden killer to win compared to executioner. Like, if you have X crit chance/damage/power, your backstabs need to total xx% of targets HP for hidden killer to win. Then people can determine themselves through their own play style which comes out ahead. Some people do a lot less attacking outside of CnD and BS compared to others who melee people down with autoattacks and HS’s. And I really believe thats what we should be talking about to determine which is better, not calculating some magic crit chance that has no baring on playstyle. I guarantee you, even with 68% crit chance you can make hidden killer win for damage output over executioner, just dont do much outside of CnD and BS and maybe even throw in a blinding powder or hide in shadows instead of a CnD.

So, you’re talking about PvP, while I talk about PvE? Well, gg then

But in PvE you will complete your chain 90-100% of the time.

In PVE I CnD trash mobs then BS them and they die. And theyre above 50% HP when I backstab them. So the HK wins there. For stationary PVE bosses where you just autoattack them and go afk, it should be pretty obvious executioner wins because you will be doing like no damage with your ‘from stealth’ attacks. In that situation Id say its probably best to just auto attack and use HS when you have full ini until the mob reaches 30%~ then just spam HS whenever you have enough ini to. Not because its the best damage output or anything, but because it doesnt matter and your watching anime or youtube or playing a different game while doing the fight and dont care, and those people with you will think your doing great because zomg heartseekerz op!

But I still say you should determine how much of your damage needs to come from your ‘from stealth’ attacks compared to the total amount of HP the mob has to determine which would be better in a given situation. I think my previous paragraphs exaggerated examples prove that. Personally, I tell people if they have 2300~ power, 50%~ crit chance and 100%~ crit damage, (this is about average for a balanced build thief) they need their from stealth attacks to do 50% of their total damage to their target while using hidden killer for it to beat executioner. And thats pretty much true for all variations around those numbers.

Nate.6974:

Actually you made the claim:

Seriously?
Did you completely glance over the first post where it gives times and no evidence to support those times? I made a counter argument, I was not the first person to give a time for the attack chain, I didnt even give a time for the attack chain, I simply said it was longer than what the OP proposed in response to his baseless guess as to how long the auto attack chain takes.

EDIT:
I assume this difference in time is because whoever thinks it takes 2 seconds is not counting from the moment they push 1 the first time until they can start a new 1 after the auto attack chain finishes, and are only looking at the animation for the chain and saying it starts when the animation starts and ends when the final hit connects. But whatever!

In PVE I CnD trash mobs then BS them and they die.
I do not min/max my build for trash in open PvE. But you’re right, HK is better here.

For stationary PVE bosses where you just autoattack them and go afk
0/10 – bad troll attempt. Please try to be serious.

But I still say you should determine how much of your damage needs to come from your ‘from stealth’ attacks compared to the total amount of HP the mob has to determine which would be better in a given situation.

Thats what I am doing. The mob has so much HP, that the number of complete AA/C&D/BS-chains is much higher than the number of interrupted chains. Then I look at the skill coeficients to determine how much damage comes from BS.

2.4/(2.4 + 3.76) = 0.39 (39%) to be accurate.

Anubarak isn’t the OP, he just commented on your claim. Anyways, not important. Although I agree, I would like to know how people are getting their values; are you clocking it from a video? Counting how many hits over X amount of time?

wish.1027:

I tell people if they have 2300~ power, 50%~ crit chance and 100%~ crit damage, (this is about average for a balanced build thief) they need their from stealth attacks to do 50% of their total damage to their target while using hidden killer for it to beat executioner.

You call it average I call it weak… My thief has a 70% base crit chance, with + 12% while behind a foe and >90% health

Give me a banner and my crit chance is about 90%… If there is a warrior I have 100% fury uptime so I reach 110% crit chance…. tell me… why do I need hidden killer?

@Nate: I used fraps and counted the time/ the frames of each attack

@Nate: i used fraps and recorded at 60 frames per second. then i look at every single frame and can tell how long it takes

wish.1027:

Molch.2078:

The critical damage modifier is 1 + critdamage x critchanche.

I dont think this is correct, or I’m misunderstanding what you’re saying. It shoud be:
(1 * none-critchance) + ((1.5 + critdamage) * crit chance)
With the crit chances simply being looking at you crit chance and putting a decimal in front of it. Like 44% crit chance would be .44 and then none-crit chance would be .56 Crit damage would be converted to a decimal in the same way, 50 crit damage would be .5 112 crit damage would be 1.12, etc. So 81% crit damage and 44% crit chance would look like this:
(1 * .56) + ((1.5 + .81) * .44) = 1.5764
Using your equation it looks like this:
1 + .81 * .44 = 1.3564
or this maybe?
1 + (1.5+.81) * .44 = 2.0164
But ya, theyre all different numbers.

Molch.2078:

Since executioner kicks in at 50% hp, we treat it as a 10% damage boost.

Since I laugh at thief PVE how about we think about PVP, and how people have heals they use at 50% Hp and count executioner as like 6.6% damage instead. Or realize that the warrior is going to be hitting multiple targets a lot since it’s PVE and it auto-wins.

Not a major problem, but executioner is only worth a 9% damage boost, assuming you’re the only person who has a damage increase sub-50%.

For half of the health of the boss, he will lose health at a rate of the character’s DPS, the final half will be 1.2x that DPS. You then need to calculate the total time necessary to take the boss from 100% to 0% at these two rates, and divide the health of the boss by that time. This will give you an accurate percentage. This is due to the fact that you will spend more time dealing damage at the lower rate of DPS than you will at the higher rate.

Here’s how the math works out:
b = Boss HP
d = Character DPS
t1 = Time for first 50%
t2 = Time for last 50%

.5b / d = t1
.5b / 1.2d = t2
b / (t1 + t2) = Average DPS

Plug any values in there and you’ll see that you get a 9.09% DPS increase from executioner.

Prodicus.6149:

Not a major problem, but executioner is only worth a 9% damage boost, assuming you’re the only person who has a damage increase sub-50%.

For half of the health of the boss, he will lose health at a rate of the character’s DPS, the final half will be 1.2x that DPS. You then need to calculate the total time necessary to take the boss from 100% to 0% at these two rates, and divide the health of the boss by that time. This will give you an accurate percentage. This is due to the fact that you will spend more time dealing damage at the lower rate of DPS than you will at the higher rate.

Here’s how the math works out:
b = Boss HP
d = Character DPS
t1 = Time for first 50%
t2 = Time for last 50%

.5b / d = t1
.5b / 1.2d = t2
b / (t1 + t2) = Average DPS

Plug any values in there and you’ll see that you get a 9.09% DPS increase from executioner.

On the contrary, this calculation is true, if you are the only player/if your whole team has + 20% sub 50%. If you are the only one in a team, b is not the boss HP, but the part of the boss hp you bring down, and this part is bigger in the second half of the fight since you will do more damage, but the rest of your team won’t.

.5b_1 / d = t1
.5b_2 / 1.2d = t2
(b_1 + b_2) / (t1 + t2) = Average DPS

b_1<b_2

tbh I’m not 100% sure about that, let me think for a moment…
… that would mean, that your damage increase depends on your teams dps. If they don’t do damage, we’re back to your calculation, and when they do infinite amount of damage, the game gets boring really really fast.

Your damage >50%: d
Your damage <50%: 1.2 * d
Teams damge: 4 * d
Boss hp: b

b_1 = 1/2 * 1/5 * b
b_2 = 1/2 * (1.2/4 + 1.2) * b = 1/2 * 0.23 * b

as you said, it takes you
.5b_1 / d = t1
.5b_2 / 1.2d = t2
seconds to deal your damage and your dps is
(b_1 + b_2) / (t1 + t2)

Looong equotation inc:

b * (1/2 * 1/5 + 1/2 * 0.23) / (1/2 * 1/5 * b / d + 1/2 * 0.23 * b / 1.2d)
———————————————————————————————-
b * 1/5 / (1/5 b / d)

=

well fk sth is wrong….

Edit, nope, I’m just too dumb to use WA correct, it’s 1.0979 aka 9.79%

Molch.2078:

Prodicus.6149:

wish.1027:

Molch.2078:

The critical damage modifier is 1 + critdamage x critchanche.

I dont think this is correct, or I’m misunderstanding what you’re saying. It shoud be:
(1 * none-critchance) + ((1.5 + critdamage) * crit chance)
With the crit chances simply being looking at you crit chance and putting a decimal in front of it. Like 44% crit chance would be .44 and then none-crit chance would be .56 Crit damage would be converted to a decimal in the same way, 50 crit damage would be .5 112 crit damage would be 1.12, etc. So 81% crit damage and 44% crit chance would look like this:
(1 * .56) + ((1.5 + .81) * .44) = 1.5764
Using your equation it looks like this:
1 + .81 * .44 = 1.3564
or this maybe?
1 + (1.5+.81) * .44 = 2.0164
But ya, theyre all different numbers.

Molch.2078:

Since executioner kicks in at 50% hp, we treat it as a 10% damage boost.

Since I laugh at thief PVE how about we think about PVP, and how people have heals they use at 50% Hp and count executioner as like 6.6% damage instead. Or realize that the warrior is going to be hitting multiple targets a lot since it’s PVE and it auto-wins.

Not a major problem, but executioner is only worth a 9% damage boost, assuming you’re the only person who has a damage increase sub-50%.

For half of the health of the boss, he will lose health at a rate of the character’s DPS, the final half will be 1.2x that DPS. You then need to calculate the total time necessary to take the boss from 100% to 0% at these two rates, and divide the health of the boss by that time. This will give you an accurate percentage. This is due to the fact that you will spend more time dealing damage at the lower rate of DPS than you will at the higher rate.

Here’s how the math works out:
b = Boss HP
d = Character DPS
t1 = Time for first 50%
t2 = Time for last 50%

.5b / d = t1
.5b / 1.2d = t2
b / (t1 + t2) = Average DPS

Plug any values in there and you’ll see that you get a 9.09% DPS increase from executioner.

On the contrary, this calculation is true, if you are the only player/if your whole team has + 20% sub 50%. If you are the only one in a team, b is not the boss HP, but the part of the boss hp you bring down, and this part is bigger in the second half of the fight since you will do more damage, but the rest of your team won’t.

.5b_1 / d = t1
.5b_2 / 1.2d = t2
(b_1 + b_2) / (t1 + t2) = Average DPS

b_1<b_2

tbh I’m not 100% sure about that, let me think for a moment…
… that would mean, that your damage increase depends on your teams dps. If they don’t do damage, we’re back to your calculation, and when they do infinite amount of damage, the game gets boring really really fast.

Haha, yeah. The damage gained by mob hp% based skills are directly proportional to the overall damage dealt by your group below that thresh hold. I.E. If your group deals a greater proportion of the total damage sub 50%, your personal damage increase will be closer to 10%. What we see here is that your DPS increase on a reasonably long fight will be between 9% and 10%. The only time it should drop below the 9% mark is if you leapfrog the 50% mark by a significant margin.

That’s why I like flat damage skills rather than % execution style ones. I’ve disliked the math necessary to be accurate since they started being talented in WoW.

Good discussion! Thank you.

Warriors still generally have more DPS when you factor in 100B + whirlwind + might spam + fury through banners. You can still factor in might on stealth, however that requires traits located elsewhere other than the “glass” trait lines.

Regardless conjure eles w/ might still have the most DPS

Molch.2078:

Thats what I am doing. The mob has so much HP, that the number of complete AA/C&D/BS-chains is much higher than the number of interrupted chains. Then I look at the skill coeficients to determine how much damage comes from BS.

2.4/(2.4 + 3.76) = 0.39 (39%) to be accurate.

I think your 39% is fine for executioner, but you have to factor in crit chance/damage for hidden killer, it’s not that simple. At no point did I see a statement from you consisting of “you need x% of your damage to come from your ‘from stealth’ attacks for hidden killer to win,” so for me, you havent really answered the question.

Anubarak.3012:

wish.1027:

I tell people if they have 2300~ power, 50%~ crit chance and 100%~ crit damage, (this is about average for a balanced build thief) they need their from stealth attacks to do 50% of their total damage to their target while using hidden killer for it to beat executioner.

You call it average I call it weak… My thief has a 70% base crit chance, with + 12% while behind a foe and >90% health

And I call you cannon fodder or PVE carebear who runs from a 1v1 when you dont get the jump on someone. But hey, thats just the difference between PVP and PVE. You cant build a GC thief in PVP and expect to survive or do much without sneaking up on people, and should your initial burst fail you have to run away to avoid dieing.

Anubarak.3012:

Give me a banner and my crit chance is about 90%… If there is a warrior I have 100% fury uptime so I reach 110% crit chance…. tell me… why do I need hidden killer?

Are you sure about that?
Theres no critical hit chance cap before 100%? And again, PVE, no one plays a GC thief in PVP and hangs out next to a warrior all the time, and if they do, those warriors certainly dont use banners.

Anubarak.3012:

@Nate: I used fraps and counted the time/ the frames of each attack

Then why is it so hard for you prove it?
Here, I’ll do it for you. As can be seen in the uploaded screenshot of movie maker, this clip is 2.07 seconds long, and starts slightly behind where it should to get an accurate time. (There should be another .05~ seconds of time for the first screen to match the last screen, but since I did this in a rush with movie maker, it proved to be difficult to line up the start and end point on exact places in the auto attack chain, and I dont care, as Im just trying to prove someone wrong, rather than prove the exact time it takes) This took me 5 minutes of effort and proves you wrong. Which just goes to show me, if you actually did this, youd have uploaded something to showcase it the first time anyone brought up that you might be wrong.

Prodicus.6149:

Not a major problem, but executioner is only worth a 9% damage boost, assuming you’re the only person who has a damage increase sub-50%.

Thats assuming whatever youre fighting doesnt heal itself back to full the first time it reaches 50% like any competent player in PVP would.

The rest of the posts about it not actually being 10% when something just lets you fight it from 100% > 0% are pretty cool though, I never though about it like that and how since you’re doing more damage with each swing the enemy dies faster so its actually worth less than 10%. Good posts

Doomdesire.9365:

Warriors still generally have more DPS when you factor in 100B + whirlwind + might spam + fury through banners. You can still factor in might on stealth, however that requires traits located elsewhere other than the “glass” trait lines.

I think if we’re really comparing warrior and thief PVE DPS, we should count the buffs warrior gives to the other 4 players around it as a damage increase it is directly responsible for, so that DPS should be added to the warriors DPS. Which would easily push it above the thief even if fighting a single enemy.

wish.1027:

Anubarak.3012:

@Wish:
1) no, It takes 2.05sec use fraps and check it out

Why dont you post a video and show us. You know, prove what you claim to be correct. The burden of proof is on you as you’re the one making the claim.

There are many. Go to any youtube video showing a full thief auto chain, take a stop watch, start it when the 1st attack lands, stop it soon as the poison strike lands.

Anubarak.3012:

2) yeah noob thieves are using Hidden Killer, you are right. Thieves with an average crit chance of 61% are using executioner because it deals more dmg

I’ve done the math for hidden killer vs executioner and I can tell you the point between crit chances where one is more effective than the other based on damage done from in-stealth attacks, but please, post the math in this thread and show us all how executioner is always superior at all points in time forever.

You didn’t do the math well enough then or you did it for something like toughness, power, crit gear. I have 80% crit chance with fury (which I have almost all the time with allies + furious). Why would I waste a 30 point trait for nothing useful when I could get a 20% dps boost?

Anubarak.3012:

4) You are right! There are so many boss fights in gw2 where you have to fight against two bosses without adds that die instantly when you look at them…. uhm… wait.. no…

So the only thing you ever fight is bosses, and only 1 mob at a time at all times, because those bosses never have adds ever. Seriously? That sounds like some super pro-level PVE ownage there. I better just take my lame self back to PVP before I look like a complete fool in this thread.

Lets see. The main dungeons I do is arah, cof, coe, ta, and fractals. There is no boss in any that require massive aoe. The only time aoe is needed, I grab sbow or s/p to blind spam for free wins. I only need to blind spam in grawl fractal but most of the time I just annihilate the shield in melee then dagger storm or cb the adds to melt them. Ta never needs aoe since we skip all the trash and most of the mobs are spread out. There are a few that might benefit from aoe but not enough to make much of a difference. Even on spider tree we just ignore the adds, projectile wall the spiders, and burst the tree. We kill it before anyone even downs. In dredge, I’m too busy ripping off all might, regen, and protection to be worrying about individual dps. Arguably, me taking off those boons means I’ve automatically done over 33% total dps vs that unit. It’s not like the mobs even matter so I don’t know why you’re even bringing them up.

If you want to bring up cof p2 and complain that lack of aoe makes thief lose in the spawn room, well you’re wrong there too sorry. No one even needs to kill a single mob if the thief is good enough. Let the mob walk to what’s his face, drop a blind field, all the assassin’s attacks just missed. Horrah. Not to mention 2 cb usually kills a group of mobs newais.

How i exactly meassured the time for the attack chains:

• 1. recording about 1 minute of doing the attack chain with steady weapons in the mists at 60 fps
• 2. counted the number of attack chains i was able to perform.
• 3. divide the number of frames by 60 to get the exact time.
• 4. divide the time by the number of chains

example for dagger:

Frames: 3700 frames
Chains performed: 30

Time = 3700 frames / 60 frames/second = 61,666 seconds
Time for 1 chain = 61,666 seconds / 30 chains = 2,0555 seconds/chain

Adding link to my calculations.
I am not 100% sure about the damage modifications.

https://docs.google.com/spreadsheet/ccc?key=0AlOXvenKdNo2dEZXWEpmUXNuMEhTUWI3LWQ2UG40VVE#

crouze.3078:

How i exactly meassured the time for the attack chains:

To bad that didnt work out correctly for you. As the screenshot of moviemaker I posted proves you wrong.

The correct way to do it is to record 30~ seconds of auto attack animation, watch it once through to make sure it doesnt lag or skip, then pick a starting point, delete everything in front of it, then find the next time that exact frame occurs, and delete everything after it, then look at the videos time length.

randomfightfan.4091:

stuff

Youre a bit late to the party.

wish.1027:

crouze.3078:

How i exactly meassured the time for the attack chains:

The correct way to do it is to record 30~ seconds of auto attack animation, watch it once through to make sure it doesnt lag or skip, then pick a starting point, delete everything in front of it, then find the next time that exact frame occurs, and delete everything after it, then look at the videos time length.

oh, why didnt i thought about it :s

i tested it with the sword auto attack and i am 0.015 seconds off so i am really close to that.

thx for this input!

@randomfightfan: Thank you^^ You wrote exactly what i thought

wish.1027:

The correct way to do it is to record 30~ seconds of auto attack animation, watch it once through to make sure it doesnt lag or skip, then pick a starting point, delete everything in front of it, then find the next time that exact frame occurs, and delete everything after it, then look at the videos time length.

Then you did it wrong, too. Look at the icon of dagger 1 in both of your pictures. It is not the same.

And just to get this straight. We’re arguing about the difference between 2.06 and 2.07 seconds, yes?

And since 0.01 seconds are the smallest step your program can measure, you have to calculate a error of + – 0.01 anyway. It could be anything from 2.065 to 2.0749999 sec.

Molch.2078:

wish.1027:

The correct way to do it is to record 30~ seconds of auto attack animation, watch it once through to make sure it doesnt lag or skip, then pick a starting point, delete everything in front of it, then find the next time that exact frame occurs, and delete everything after it, then look at the videos time length.

Then you did it wrong, too. Look at the icon of dagger 1 in both of your pictures. It is not the same.

And just to get this straight. We’re arguing about the difference between 2.06 and 2.07 seconds, yes?

And since 0.01 seconds are the smallest step your program can measure, you have to calculate a error of + – 0.01 anyway. It could be anything from 2.065 to 2.0749999 sec.

Ya, look at the #1 icon, that clearly shows that the beginning occurs after the end, and for this video to be accurate it would need at least another .05 seconds like I said in my original post about it.

My point wasnt to show the correct time and prove what it is, it was to prove others wrong, so it not being perfect is totally fine.

wish.1027:

Dahkeus.8243:

wish.1027:

Dahkeus.8243:

They really need to make PvE test dummies

This would really help give better test data and allow more solid conclusions to be made.

No it wouldnt… only a fool would sit in a game and try to determine average damage by hitting stuff when the equations for damage calculations are known and would be many, many, many hours faster. Does a parser for GW2 even exist or are you going to sit there and write down every value so you can add them up and divide?

…I think you’re misunderstanding.

With the right dummies, all a player would need to do to test dps of different specs/classes would be to time how long it takes to kill a dummy, then compare kill times.

Sure, as long as the dummies take an hour or two to kill. Otherwise Im going to argue small numbers and that math is better. And at that point where you fight a dummy for an hour or two, math is faster. If they die super fast, youd have to kill them for an hour or two, then average all the times together to get an idea of your real time. Because sometimes you’ll get all crits, and sometimes you’ll get no crits, and sometimes your crits will proc on the bigger hits and not the smaller ones, etc. You’d have to do it a lot so the numbers eventually balance themselves out, then add all the times together, and divide by how many times you killed it o figure out your true time. So youre still doing math anyway even with the test dummys.

ok, here’s the thing:
A ) you don’t need to spend hours with each build/class test to average out crits.
B ) Even if you average out kill times from multiple tests, the math is still much, much, much simpler than calculating together weapon damage, stats, chance of proc, chance of crit, etc. together.