Guild Wars 2

I did these tests for the other classes I play. So, just before I returned to Guardian for a little bit, I thought I’d do a quick rundown on weapons to see if there were any surprises compared to what I thought I knew about the class.

Test Guardian is my standard PvE build: 20/30/0/20/0. Damage+10% v. burning, GS+5%, Sword+5%, 1H+15% crit, 2H CD-20%. So, this test is more personal to my (very PvE’ish) guardian than it would be if I ran a blank trait setup.

Heavy Golem in the Mists. Kills timed with a stop watch:

               auto   skills  skills+f1
mace      10s     8s      6s
sword      7s      6s      5s
gsword    8s      6s      5s
hammer 8s      8s      7s
staff        15s     13s    12s 
scepter  11s      6s       4s

NB: My usual weapons are sword+torch/greatsword. Staff swapped for map travel only. I used to use a scepter/focus long ago. Pretty much never for mace & hammer. So, there might be better sequences giving nicer results for my less familiar weapons.

I wasn’t surprised scepter/2 skill was a heavy hitter, but I was astonished at how big it was compared to other weapons. I was also a bit surprised at our DPS. It’s basically as good as Thief with scepter being better than Thief.

I was surprised that our weapons all had such similar kill rates. They’re so equal in DPS that choice will be driven by the secondary effects, such as wide AE & closer for GS, defensives for hammer & mace, gap closer for sword (and tiny AE & defensives, too).

For comparison: Thief ~6s (2s spike), Ranger 7~9s (4~5s spike), Mesmer 15s+ (4s spike). I didn’t test Guardian spike, but I suspect 3~4s would be about as good as I could get it simply because of UI issues for making scepter pay off more versus less DPS to improve on the other weapons..

BTW, I’m not using this post as means to advocate scepter or any other particular weapon. Scepter has issues that make it difficult to use. For instance, I had to enable “fast ground targeting” to get that 4s rate. This is something I can’t really do in live play because the mouse cursor is so easy to loose — it’s too small and of a color too close to colors used in the game. Normal ground targeting adds a LOT of time to using scepter/2 and that translates to “much less DPS”.

Getting something other than a stupid PvE critter to stand still inside a scepter’s tiny AE and not laugh as it calmly sidesteps the projectiles is an exercise for any reader determined to make scepter pay off in live play. ^~

http://www.guildwars2guru.com/topic/70080-guardian-effective-dps-tests/

GS and scepter are actually miles away from the rest of the weapons, DPS-wise. Your problem is that you are just running at a golem and unloading with all your skills active, which doesn’t account for:

1) Damage over time. Symbols take at least 4 seconds, or 6 with trait, to deal full damage. If the golem dies before full damage is applied, then that’s that.

2) Skill recharges. Since all your skills are off recharge when you start and the golem will be dead before you ever use a skill a second time, long-recharge, high-damage skills like Symbol of Wrath, Symbol of Swiftness, Orb of Light detonation, etc. will heavily skew numbers.

3) Traits. You are using a non-optimal build for damage for pretty much every weapon. You don’t have any crit damage so it’s basically pointless to go 30 in Radiance since your crits deal no damage anyway. You can’t compare DPS potential unless your weapons are actually working at potential.

I also want to point out the difference in sustained DPS and burst damage. On these training golems big bursty abilities can eat through chunks of life quickly, but refresh slowly. So they may look like they kill faster on a small HP pool, but in a long duration fight it may change in effectiveness.

So as gunglai pointed out damage over time abilities don’t reach full potential and other things like long cooldown attacks.

Guanglai Kangyi.4318:

http://www.guildwars2guru.com/topic/70080-guardian-effective-dps-tests/

Are you seriously trying to argue that estimated DPS calculations using tool-tip data have more validity to game play than actual tests in game using the weapons themselves?

I love your estimates. It’s a lot of nice work. But, when the rubber meets the road, it’s always better to go with reality than to mutter “but my estimates showed I could take that corner at 72mph” as you go flying off the cliff.

You could firm up your estimates with animation data. Tool-tips aren’t accurate. They’re missing time per skill. You have to include the (unstated) animation time. The only way to get that is to measure it in game; eg. quickness would have no effect on DPS if current tool-tip data told the whole story.

But they’d still be estimates. If you truly want better weapon DPS measurements, it wouldn’t be hard to do. Mine were really quick — just simple checks on kill rates. You could easily produce more accurate results. I used a stop watch, which is prone to human error. Using a frame grabber is a LOT more reliable.

I had a small sample set — typically only doing 2-3 timings per weapon. I worked the skills myself instead of using a macro (bot). All of these factors can perhaps influence timing.

Accurate timing is critical to producing accurate DPS data.

BTW, real data needs a real build, too, not a gimick build. Coming from other classes that get crit-damage for free, the 10/30/30 build was one of the first I tried. It’s OK. It’s neither enough DPS nor enough flexibility for my taste, though. For myself, 10/30/x/10/x is mandatory. The last 20, I’ve tried all over and likely will shuffle around again and again.

I like how you are trying to argue that your kitten tests with flawed methodology are more accurate than my numbers using actual math. I’m not the one who thinks that having one skill require ground targeting makes it “difficult to use”. This isn’t Engineer grenade where you need to have the cursor on target 100% of the time to get max DPS. All you have to do is point and click. I don’t see how this is hard.

It’s not “hard”. It takes time. Every fraction of a second matters. If the mouse isn’t already in the correct position, you could loose whole seconds. This the same reason is why your estimates aren’t as useful as real tests. You’re not using accurate timing data. GIGO.

I’d be the first to agree that my tests were sloppy and that I put a lot less effort into them than you did for your estimates. The difference is that my data is real. Yours is a guess. A guess that doesn’t account for animation timing, which is VERY significant factor and wherein I think lies the discrepancy between your estimates and my test results.

Estimates are a guide towards what might happen with the real thing. They’re not a substitute. When real data contradicts the estimates in some large unexpected way, it’s time to examine what you did wrong in the estimate — not blame reality for not conforming to expectation or the person who brought reality into the discussion.

My numbers that use actual numbers and calculations are a “guess” while your random flailing is “real”. Okay.

Okuza.5210:

It’s not “hard”. It takes time. Every fraction of a second matters. If the mouse isn’t already in the correct position, you could loose whole seconds. This the same reason is why your estimates aren’t as useful as real tests. You’re not using accurate timing data. GIGO.

Whole seconds, huh.

Guanglai Kangyi.4318:

My numbers that use actual numbers and calculations are a “guess” while your random flailing is “real”. Okay.

Okuza.5210:

It’s not “hard”. It takes time. Every fraction of a second matters. If the mouse isn’t already in the correct position, you could loose whole seconds. This the same reason is why your estimates aren’t as useful as real tests. You’re not using accurate timing data. GIGO.

Whole seconds, huh.

Don’t get worked up man this is America math is not well known here. They don’t realize everything is math, math is pure fact. To argue with math is to argue with fact. They don’t realize this.

Actually Kangyi’s calcuations aren’t that useful for different reasons – in situations where you need to apply sustained DPS over several utility skill cycles, trait spec matters. I wouldn’t use greatsword without honor in real application because it is the least defensive weapon.

Putting aside the fact that most non-boss battles last for 1 cycle or less, what would be a lot more useful is DPS calcs for the most popular trait specs. I actually don’t know which situation calls for max DPS on the guardian since they’re horrible at being glass cannons in general.

kaffaljidhma.1496:

Actually Kangyi’s calcuations aren’t that useful for different reasons – in situations where you need to apply sustained DPS over several utility skill cycles, trait spec matters. I wouldn’t use greatsword without honor in real application because it is the least defensive weapon.

Putting aside the fact that most non-boss battles last for 1 cycle or less, what would be a lot more useful is DPS calcs for the most popular trait specs. I actually don’t know which situation calls for max DPS on the guardian since they’re horrible at being glass cannons in general.

The point of the calculations is to denote an acceptable average to conclude what is what. Just because your not going to fully utilize the weapon does not invalidate the data on potential.

I think the problem is math is to hard to understand, the concept even harder for some. There’s probably two out of ten people that can comprehend. What’s even there. Not because its harf but because most people are at grade 5-8 in math knowledge. I think he did a really good job with what he had. Its more than acceptable.

Its an unfortunate fact if he used words like good better ect it would probably be received better.

AerictheBoss.9172:

kaffaljidhma.1496:

Actually Kangyi’s calcuations aren’t that useful for different reasons – in situations where you need to apply sustained DPS over several utility skill cycles, trait spec matters. I wouldn’t use greatsword without honor in real application because it is the least defensive weapon.

Putting aside the fact that most non-boss battles last for 1 cycle or less, what would be a lot more useful is DPS calcs for the most popular trait specs. I actually don’t know which situation calls for max DPS on the guardian since they’re horrible at being glass cannons in general.

The point of the calculations is to denote an acceptable average to conclude what is what. Just because your not going to fully utilize the weapon does not invalidate the data on potential.

I think the problem is math is to hard to understand, the concept even harder for some. There’s probably two out of ten people that can comprehend. What’s even there. Not because its harf but because most people are at grade 5-8 in math knowledge. I think he did a really good job with what he had. Its more than acceptable.

Its an unfortunate fact if he used words like good better ect it would probably be received better.

That’s not the issue. The issue is that DPS comparisons that switch up traits instead of weapons are not as practical as ones that assume the same trait setup with different weapon and stat combinations, which are actually changeable for situations that require them. I wouldn’t be caught dead with a 10/30/30/0/0 build, for instance, except in very specific WvW situations where my party already has enough boons, and even then I’d like to know the comparative weapon DPS with that trait distribution more than the arguable knowledge that it’s the best for greatsword damage.

Players carry several weapon sets but stick with one trait set. Comparisons should reflect that.

You realize my numbers also include the base numbers before damage bonuses are applied, right? I also gave you the formula you need to calculate the damage bonuses for any given setup. If you just want to compare relative damage between weapons for one build just take the base damages and factor in any differences from traits that only affect one over the other.

Guanglai Kangyi.4318:

You realize my numbers also include the base numbers before damage bonuses are applied, right? I also gave you the formula you need to calculate the damage bonuses for any given setup. If you just want to compare relative damage between weapons for one build just take the base damages and factor in any differences from traits that only affect one over the other.

I’m lazy and would much prefer those calculations done for me rather than theorycrafting the absolute maximum for each weapon. Ideally, a set of trait distributions with the DPS of each weapon under it, with the highest DPS weapon (or pair of weapon sets) of each category highlighted.

On topic for this thread I would also like DPS comparisons for the 45 second battle, or however long a typical skirmish lasts.

No offense, but not everybody enjoys math and/or number crunching (min/maxing), and perfectly valid logic and theorycrafting may not translate well to the diverse playstyles out there. I.E. 4+3=7, but it doesn’t mean that 7 will work for everyone just because it’s the “right” answer arithmetically.

Not saying that you who enjoy those things should stop doing so, though, as long as you respect others’ “wrong/ mathematically incorrect choices” as well. Peace to all.