Damage Per Second, or How I Learned to Love The Buffs

[Ensign]

I've done a bit of work with actual skill effectiveness the last couple days - I just posted a table with the results HERE (http://www.guildwarsguru.com/content/skill-effectiveness-id1103.php).

What does this data tell us?

The most important thing is that if you want to kill things quickly, you have to learn to embrace buff stacking. Enchantments, Stances, Preparations, and just normal skills all stack with each other, dealing damage simultaneously. With a good Sword, an Conjure, Warrior's Cunning, and Frenzy, you can easily get up to 40+ Damage per Second with just normal attacks - add in something like Power Attack to make things even worse.

Elementalist damage is severely underpowered. Rangers and Warriors can do better burst damage than you outside of two noteworthy cases (Immolate and Lightning Orb) - and they spend just a fraction of the resources to do so. The reasons for this are twofold - first, casting times and energy costs on Elementalist nukes are so high that the benefits of their high damage ratings are nullified, and second, Elementalists can't stack buffs the way that Warriors and Rangers can. Straight up Elementalist attacks are fine compared to Warrior and Ranger shots, but once those other classes start buffing up the Elementalist gets left in the dust.

Which means that if you want to win, don't use Elementalists as your nukers - use Warriors and Rangers who have learned to love the buff.

Peace,
-CxE

[SpineLok]

Very nice work. The amount of insight this brings to those of us who like to look at the raw numbers is invaluable.

Would it be possible to add a damage/min to Energy Duty ratio for each skill for each profession. If you did your chart in excel this should be a quick modification. That would be really nice when trying to decide on skills when you're trying to be conscience of the energy duty to dmg/min cost or vica versa.

Also is your Energy duty calculated as energy/min? I'm assuming it was due to the large numbers. Might be helpful to label that column though.

Very Nice!

[Ensign]

Yeah, Energy Duty is energy per minute. Time Duty is seconds casting per minute. I'll edit that into the note at the bottom. Yes, most Elementalist skills have energy duties higher than your natural regen, so you either need a ton of energy management, or you need to get used to doing burst damage. =)

Damage Per Minute to Energy Duty is actually just skill damage divided by skill cost. All the other terms cancel. So it ends up being a pretty obvious number so I didn't include it - but if people think that the division will be useful, I'll put it in. =)

Peace,
-CxE

[Sausaletus Rex]

The numbers are a bit startling, really. They point to something that's become apparent recently and that's Rangers and Warriors being far superior in terms of damage output to Elementalists.

Here are some quick scratch calculations I made from the table. Everything's against a 60AL target (Chuck, did you use a Str of 12 when calculating the Warrior damages? I suppose you didn't as you'd have to consider the armor of the target a bit more. It doesn't matter as if you didn't it would just make it worse...). But look at how much damage per second Warriors and Rangers can crank out by stacking up just two damage buffs with a skill. Or even with their normal attacks.

Sword War - 18.92
w/Conjure - 28.67
W/Cunning - 32.56
Power Attack - 37.67
W/Conjure - 47.42
w/Cunning - 51.31

Axe War - 19.37
w/Conjure - 29.12
W/Cunning - 33.11
Power Attack - 38.12
W/Conjure - 47.87
w/Cunning - 51.86

Hammer War - 21.6
w/Conjure - 29.03
W/Cunning - 33.48
Power Attack - 35.89
W/Conjure - 43.32
w/Cunning - 47.77

Ranger - 15.07
w/Conjure - 21.67
w/RTW - 29.07
Power Shot - 32.07
w/Conjure - 38.67
w/RTW - 46.07

Elementalist - 12.63
w/Conjure - 20.06

Fire Ele
Flare - 22.29
Fireball - 25.07
Immolate - 48

Air Ele
Lightning Orb - 38.67
Lightning Strike - 30.38

Water Ele
Ice Spear - 16
Shard Storm - 40
Water Trident - 33.14

Earth Ele
Obsidian Flame - 34.18
Stone Daggers - 22.86
Stoning - 43.43


First off, Ice Spear is a joke. That's worse than a wand/conjure for the same amount of skill slots and a lot more energy/time.

Secondly, I think I have to give up my hammer. They get the highest numbers, yeah. But not high enough to really matter. So they can deal an extra 100 damage in a minute, big whoop. For that I have to give up that shield slot where I can get an extra two modifiers that are gong to help me out? To say nothing of the faster swing rate of swords and axes working better with further damage adds such as Order of Pain or Barbs and the like.

But most importantly, the Elementalist numbers are shockingly low. Considering just how much energy and time those Elementalists are putting into pumping out damge they just don't compete with Rangers and Warriors who can stack up their damage buffs. Not even in burst damage where you'd expect Elementalists would be far ahead of the game. What Warriors and Rangers have is sustainability. They can keep going long after an ele has to stop because they've run out of energy. Their damage should be *lower*, then, because they'll make up for it over time.

An Elementalist also gives up a lot in terms of protection, too. They've got a big weakness compared to the more well armorored Rangers and Warriors. That should be because they're more front-loaded. They can go off on a target with a lot of quick, fast, unsustainable damage, taking the chance that they can drop it before it can drop them while the others are more willing to trade blows and win the battle of endurance. That's just not the case with the publically known data (I can't say whether or not the picture is different in the test now, but you'll all get a chance to see in a few weeks or so.) and, honestly, I can't see much of a point to playing as an Elementalist at the moment.

[Vermilion Okeanos]

Hmm.. those are interesting points... however, I don't think magical damage should be calculated in the same format as physical damages... they are just not the same way of usage nor play style.

[Pharalon]

Just a small point Charles, but shouldn't all the normal attacks have a time duty equal to their attack rate, as you can't perform any actions during an attack?

[Sausaletus Rex]

Yes, the time duty for a weapon attack, in Charles's table, would be 60. Or 100%, however you want to calculate it. You only get the DPM by attacking constantly, and you can't do anything else while you're attacking the same way you can't while casting.

This means that the time you take out to buff Warrior's Cunning or Conjure Whatever reduces your DPM a bit and then you start getting into hideously complicated scenarios about just how things play out. Which, since you can probably assume all skill information to be at least slightly innaccurate these days, I'm not going to really get into.

Although it doesn't touch the metric of Chuckles's DPS, that's just Effect/casting.

Hmm.. those are interesting points... however, I don't think magical damage should be calculated in the same format as physical damages... they are just not the same way of usage nor play style.

How would you calculate it, then? Damage isn't all that difficult, regardless of play style or usage. You find a target, you get in range of that target, and you deal damage until they drop or you find a better target. Now, the ways you go about doing that are certainly different. But what Charles points out is the effect, not the methodology. All that matters in the end is the raw numbers. That's what tells you how good something is at dealing damage. And that's the concern here. Elementalists are, of necessity, going to have a different style of doing things because of the way their damage works. So will Rangers compared to Warriors, for that matter. But it doesn't matter how you have to use them to get the most damage out of them, it matters how quickly they can drop that target.

[Ensign]

Just a small point Charles, but shouldn't all the normal attacks have a time duty equal to their attack rate, as you can't perform any actions during an attack?

Yes, they should. I'd been kicking around the best way to list those and just giving them a maximum time duty does the trick.

And yes, having to stop to re-do your preparations cuts into your attacking time, which nullifies some of your damage per minute. Read the Wind is actually surprisingly mediocre because of that - you could just fire off another Power Shot or whatever during that time and dealt another 60+ damage, instead of using a buff that gives another 90 damage. It's still a solid damage boost, and the arrow speed is definitely worth it, but you do have to take those casting times into account.

Peace,
-CxE

[Vermilion Okeanos]

Unlike the warrior and ranger... the elementalist would spam their spells one after another which are all different but working together (most of these skills stay on battle field as the next one is being casted). While warrior ranger attack skills pretty much work alone with itself in just that instant that are buffed up by other skills.

Warrior and Rangers stack their buffs to increase overall DPS
Elementalist stack their spell damages to increase overall DPS

Now also in terms of AoE, which everyone know...

[SpineLok]

Damage Per Minute to Energy Duty is actually just skill damage divided by skill cost. All the other terms cancel. So it ends up being a pretty obvious number so I didn't include it - but if people think that the division will be useful, I'll put it in. =)

Peace,
-CxE
Just want to confirm that you are taking each skill damage individually and dividing it by its recycle time (includes cast time right?) to get damage/s then multiply by 60s to get the duty cycle @ 1 min?

yeah, I realize its just the energy / dmg. I was just saying you could take Energy duty and Damage duty columns and divide them as long as the duty cycle was the same and the program you used to create the table was similar to Excel.

Anyhow I think it would be useful to have this column in your table because its a quick way to overview a large number of skill's energy:dmg/s or mult by 60 get energy:dmg/min. In particular its just a means of defining a skills efficiency at converting energy to damage or energy to health in the case of monks etc etc. Might be useful when trying to select between two skills of similar energy or damage duty cycles. This way you don't make us do the division in our heads... arrrrrrrg :p heheh A bit easier/faster for skill analysis, no guessing and no arithmetic mistakes :D

Just a thought, definitely open for discussion.

Yes, most Elementalist skills have energy duties higher than your natural regen, so you either need a ton of energy management, or you need to get used to doing burst damage. =)
Heheh unless you have a good supporting Necromancer on your team whose purpose is to generate energy. Energy regen spikes and no more problems :D

Again nice spreedsheet.

[Ensign]

Just want to confirm that you are taking each skill damage individually and dividing it by its recycle time (includes cast time right?) to get damage/s then multiply by 60s to get the duty cycle @ 1 min?

For skills that are easy to calculate:

Damage per Second = Skill Damage / (Casting Time + .75)
Damage per Minute = Skill Damage * 60 / (Casting Time + Cooldown)
Energy Duty = Energy Cost * 60 / (Casting Time + Cooldown)
Time Duty = (Casting Time + .75) * 60 / (Casting Time + Cooldown)


Anyhow I think it would be useful to have this column in your table because its a quick way to overview a large number of skill's energy:dmg/s or mult by 60 get energy:dmg/min.

Er, the /s or /min falls out when you do the division...

Anyway, yeah, it's a useful number and one that I'd like to include at some point. Actually I want to make an article out of all of this and explain things in a bit more detail, and divide things up in a way so that they make more sense (try making all of the Time Duties work in a way that's consistent. It's maddening.) Adding statistics like that are useful, and I'd like to get something like that done at some point for every skill in the game (again with all sorts of unification problems that have to be worked out), but for now that table's up there as a teaser, to get people thinking about how these things work.


Unlike the warrior and ranger... the elementalist would spam their spells one after another which are all different but working together (most of these skills stay on battle field as the next one is being casted).

Well, you have two classes of skills really - those that buff damage, like Warrior's Cunning/Conjure Element, and those that deal damage, like Power Attack/Immolate. Basically, skills that you stack, and those that you chain. You have to treat each one differently and the distinction is one that I'd like to draw a bit better in the future.

The fact that damage over time and area of effect skills aren't listed (besides fireball) was done on purpose, because they introduce other problems. For example, people rarely stand in a Firestorm, so figuring out how much damage one actually deals is problematic. What's the DPS from Firestorm, including the cast time? As I said, problematic, which is why they aren't included.

Of course, I am saved by the fact that the various storm skills, with the notable exception of Maelstrom, are awful beyond words, so analyzing their DPS and DPE isn't all that useful. Maelstrom you don't use for the damage anyway, but the interrupting - there are probably more Maelstroms cast at level 0 than any other level.

If there's a lesson in this, it's that skills are varied enough that single metrics cannot be applied to all of them.

Peace,
-CxE

[Sausaletus Rex]

Just want to confirm that you are taking each skill damage individually and dividing it by its recycle time (includes cast time right?) to get damage/s then multiply by 60s to get the duty cycle @ 1 min?

Damage per second or DPS is equal to Damage divided by casting time. Not recycle time. It's a measure of burst damage, just how much damage you'll do of that skill per cast. When you're not casting it you're free to do other things and that's where DPM comes in because that's includes the recycle time somewhat by dint of including the overall time. For a skill like Immolate there's a 17.5 Duty Time. That means that over the 60 seconds to get that 840 damage you'll spend 17.5 seconds of it casting and will have 42.5 seconds of down time where you're recharging. That's time you can fill up with other things, including more skills to cast to damage or refreshing your buffs or whatever else.

To get DPM then, you multiply DPS times the Duty Time. Immolate has a DPS of 48. 17.5 times that equals 840.

I know, it's weird. But that's how Chuck's done it so that's what we're discussing.

Unlike the warrior and ranger... the elementalist would spam their spells one after another which are all different but working together (most of these skills stay on battle field as the next one is being casted). While warrior ranger attack skills pretty much work alone with itself in just that instant that are buffed up by other skills.

Well, if you want to look at it that way, Warriors and Rangers also "stack" skills to cause more damage. A Warrior doesn't just stand there and swipe with his sword. He uses Power Attack and Sever Artery and Gash and Galrath and Pure Strike and Final Thrust in rapid succession the same way an Elementalist would use Fire Storm, Fireball, Immolate, Rodgort's Invocation, and Searing Heat in rapid order. But those Warrior skills are attack skills, meaning they add on to weapon damage and anything modifying weapon damage. All the Conjures and Barbs and whatever else just pile on top of that base damage that a Warrior/Ranger doesn't have to spend any energy to get.

It's more accurate to say that damage dealers stack their skills to increase their DPS. And Warriors and Rangers can stack buffs to increase their skills.

But, if you want to look at it in terms of what a character can produce, we'll have to turn from DPS to DPM because we'll need to include strings of skills. So, to calculate how an Elementalist or a Warrior will deal damage over time with a series of skills you need first to determine the amount of time and then what skills you're dealing with. We'll use Chuck's 60 seconds as a starting point. Then, for each source of damage you need to figure out how much of the time it's going to be used during that time, that's the Time Duty. Multiply that times the DPS, add up all the sums and you'll have the amount of damage a character will do during that 60 seconds. Here's some more quick scratchings on how that works out, I've included the energy duty and the amount of damage per energy so we can compare just how efficient a particular character is being :

Sword War - 1135
Time Spent Attacking - 60
Energy Spent - 0
Damage per en - N/A

Sword War w/ Conjure - 1692
Time Spent attacking - 59
Energy Spent - 10
Damage per en - 169

Sword War w/Conjure and Cunning - 1921
Time Spent Attacking - 59
Energy Spent - 25
Damage per en - 77

Sword War w/Conjure and Cunning spamming Power Attack - 2202
Time Spent Attacking - 59 (44 with regular, 15 with Power Attack)
Energy Spent - 81
Damage per en - 27

Sword War w/ Conjure and Cunning spamming Power Attack, Galrath, and Final - 2653
Time Spent Attacking - 59 (15 w/ Power Attack, 8 w/ Galrath, 6 w/ Final, 30 w/ regular)
Energy Spent - 81
Damage per en - 33

Sword War w/ Conjure and Cunning spamming Galrath and Final - 2372
Time Spent Attacking - 59 (8 w/ Galrath, 6 w/ Final, 45 w/ regular)
Energy Spent - 25
Damage per en - 95


Ranger - 904
Time Spent - 60
Energy Spent Attacking - 59
Damage per en - N/A

Ranger w/ Conjure - 1284
Time Spent Attacking - 59
Energy Spent - 10
Damage per en - 128

Ranger w/ Conjure and Read The Wind - 1454
Time Spent Attacking - 50
Energy Spent - 31.4 (variable depending on Exp)
Damage per en - 46.3

Ranger w/ Conjur and Read the Wind spamming Power Shot - 1482
Time Spent Attacking - 50 (15 with Power Shot, 35 with regular)
Energy Spent - 106.4 (variable depedning on Exp)
Damage per en - 14

Elementalist w/ Wand - 758
Time Spent Attacking - 60
Energy Spent - 0
Damage per en - N/A

Elementalist w/ Wand w/ Conjure - 1189
Time Spent Attacking - 59
Energy Spent - 10
Damage per en - 118

Air Elementalist - 1443
Time Spent Attacking - 41 (24 w/ Lighting Orb, 18 w/ Lightning Strike)
Energy Spent - 179
Damage per en - 8

Air Elementalist w/ Wand w/Conjure - 1804
Time Spent Attacking - 59 (41 w/ skills, 18 w/ wand)
Energy Spent - 189
Damage per en - 10

Fire Elementalist - 1869
Time Spent Attacking - 60 (17 w/ Fireball, 17 w/ Immolate, 26 w/ Flare)
Energy Spent - 169
Damage per en - 11

Earth Elementalist - 1566
Time Spent Attacking - 42 (24 w/ Obsidian Flame, 18 w/ Stoning)
Energy Spent - 236
Damage per en - 7

Earth Elementalist w/ wand w/ Conjure - 1907
Time Spent Attacking - 59 (42 w/skills, 17 w/ wand)
Energy Spent - 246
Damage per en - 8

Water Elementalist - 1672
Time Spent Attacking - 62 (26 w/ Ice Spear, 26 w/ Water Trident, 10 w/ Ice Spear)
Energy Spent - 205
Damage per en - 8

I just did sword wars because they're roughly equal to axes and not that far off from hammers and they're the "worst" case as far as initial damage goes.

So, once again, Elementalists do less damage and they take a lot more energy to do it with. The most efficient are Warriors, but that's not surprising given how they can skirt around energy with adrenaline and also the fact that they need to work harder to get and stay in ranger than Rangers and Elementalists. They need to be a bit more efficient because they won't get as much of a chance to do damage (That's if you start off with everyone being equal. Consider Rangers the baseline for damage. Warriors need to do more damage at higher efficiencies than Rangers because of the lack of range. Elementalists need to do more damage at lower efficiencies because of their vulnerabilities. In a perfect world, of course). And here's something else to keep in mind :

Energy Gain over 60 seconds
Warrior - 40
Ranger - 60
Elementalist - 80

Any energy usage at or below those numbers means that character can keep going indefinitly. Anything above and that character is going to have to stop before the 60 seconds are up at some point. Then you're just waiting for a BiP or regen because your damage output is 0.

Add up all the Elementalist skills you want and try and pump that DPS higher, what Ensign has on his list are the most efficient ele spells. Things like Fire Storm and Earthquake are less efficient because of the AoE effects. They do more damage to more targets so they do a bit less damage per each individual target. But no matter what you add you're still going to burn through energy at a prodigious rate. Far more than you can sustain. Compare that to that Sword War using just a Conjure, Cunning, and the adrenal sword skills (Chuck doesn't have Pure Strike on there, that'd be even worse. There's still time to toss in the occasional Sever and Gash, as well) who'll spend just 25 energy - far less than they'll have to worry about - to get more damage than most Elementalists. That's energy that can be spent elsewhere, perhaps on the Sprint to catch up to a target, or the Hamstring to keep them close, or the rez your team suddenly needs.

[Freyas]

It is surprising how much of a difference there is between warrior and elementalist damage, though it does fit with my experience well. The reason that elementalists show up so much worse here, is that they are generally based around burst damage. In a short period of time, no character can match the sheer damage output that an elementalist can do. However, in order to get that sheer damage output, they lose efficiency. Regardless of damage buffs, no warrior will be able to deal the same damage as a PBAoE elementalist spamming all their skills as fast as they can- 100 damage from inferno, 90-100 damage from flame burst x2, 85 damage(ignoring armor) from crystal wave, throw in an immolate and maybe an obsidian flame or two, and an aftershock, all within 10 seconds, and you output an enormous amount of damage(to anyone or anything foolish enough to not run away). However, after blowing your energy on this string of spells, you're pretty much out of the battle for a minute trying to regenerate your energy, while a warrior or ranger can continue to deal their damage consistantly for the entire period of time.

AoE spells such as fireball or phoenix serve different purposes- they can be some of the most efficient ways of dealing damage, but it's circumstancial. If you can hit 6 people with your fireballs, it does excellent both in dps and dpe(damage per energy). If you only hit one person, it's not too great. Storm spells such as firestorm have the potential to output great dpe, and decent dps, if you can get your enemies to stay in them. Firestorm has the potential to deal 240 damage per target, for 4 seconds casting time and 15 energy. That's 60 damage per second you spend casting it, and 16 damage per energy. If you've got multiple enemies inside the effect, it's efficiency skyrockets. However, unless you're playing PvE, it becomes far less appealing, as people will just move out of the area of effect, and you will often only get a few hits for 24 damage in for your 15 energy and 4 seconds. The exception is when you're dealing with crucial territory in PvP, such as KoTH maps in tombs. If you're trying to take the central altar in the Hall of Heroes, you can cast a firestorm, and be guaranteed to at least be hitting the enemy ghostly hero, and likely get quite a bit of damage in on enemy characters that are trying to protect the hero.

Overall, Elementalists have a lower general dps and higher dpe than warriors or rangers, as they tend to either depend on certain situations to deal large amounts of damage, or do a lot of damage in a small period of time, but then next to no damage due to having used up all their energy. Warriors come out on top, as they can consistantly deal their damage over an extended period of time- attacks are free. The biggest lopsidedness that I can see, however, mostly comes from the fact that warriors can benefit from a large number of effects to increase their damage, whereas there is very little that an elementalist can do to increase their damage output.

[Narcism]

Ensign, I'm not curious as to whether or not you considered hitting more than one character with some of the elementalist skills you specified. I do agree (especially after seeing these tables) that warriors and rangers are best at single target damage, but, elementalists can SURELY surpass them if 2-3 targets are being considered.

[cpukilla]

Another issue with ranger damage is that their arrows miss, especially at long range. This is not a problem for warriors, and less of one for ele's especially with splash damage attacks. For warriors the problem is energy, its difficult to have energy to spam power attack and keep up warriors cunning for example.

[Narcism]

Another issue with ranger damage is that their arrows miss, especially at long range. This is not a problem for warriors, and less of one for ele's especially with splash damage attacks. For warriors the problem is energy, its difficult to have energy to spam power attack and keep up warriors cunning for example.

Warrior's problem is when they can spend too much time chasing their target, or whatever. Ranger's have a little area to work with.. but.. still area...

As well, there are tons of skills that can block/evade/negate Warrior and Ranger damage, you can't really say the same for a Firestorm, or Fireball, w/e (though some spells can be evaded, like Water Trident, which has 10% accuracy :D )

[GhostRaptor]

Warrior's Cunning (60 AL) 3.89 140 0 15
Warrior's Cunning (100 AL) 6.36 229 0 15
The above confuses me. I am reading that as being a case of using Warrior's Cunning against two targets, one in 100AL armour and one in 60AL armour. How in heck does the DPS result in being higher against the 100AL target?

Looks to me like there's something really wrong with your calculations. The problem is duplicated throughout your table for all skills that have armour penetration.

[Narcism]

100 AL -> 82 AL (after Warrior's Cunning) (-18 AL off target)
60AL -> 49 (after Warrior's Cunning) (-11 AL off target)

11 / 18 = 0.61
to verify Ensign's numbers:
3.89 / 6.36 = 0.61

18% off 100 AL benefits more than 18% off 60 AL... So the damage benefit from armor penetration (Warrior's Cunning) on someone with higher armor is bigger (more damage added).

He's not talking about the base damage. The base damage on a 100 AL target is still significantly worse than on a target with 60 AL...

[Sausaletus Rex]

I believe what you're looking at with the 60AL vs. 100AL descriptions is the increase to DPS. That table can be seen as sums to be totaled. And sums that largely ignore the effects of armor (otherwise there's be something like "All damage is against a 60AL target, against a 50AL target it's raised by x%"). You figure out what sort of attacks a character is doing, you add that all up and then figure out how armor is to see what you're really really doing.

Warrior's Cunning, then, and all other skills that penetrate armor have a jump when used against higher armor because they're more effective. 100AL is effectively 82AL. While 60AL is only effectively 49AL. You can check out Ensign's essay on the damage equation and more here (http://www.guildwarsguru.com/content/game-mechanics-id674.php) for exactly what that means (bring your slide rule, folk!) but the percentage of damage you're dealing will be raised more with a difference of 18AL versus a difference of 11AL. The damage will increase more against that more armored target and that's what the table shows. The number is higher against the 100AL target because that's how much of a jump you'll get from the normal DPS or DPM. However, the actual damage will be lower because that's still more armor when you actually loook to how your damage will be influenced by the armor of your target.

[GhostRaptor]

OK, I was reading it wrong then. In which case ... it needs a far better set of explanations on what is being displayed ;).

[Sausaletus Rex]

I'm sure Charles sees it as some sort of intelligence test. If you can't understand the table without detailed explanation then you're not smart enough to be able to talk about it. If you can understand the table then there's no need for explanation. Therefore, explanations are sub-optimal and supperfluous.

That or, you know, he showed me the table because he's working on an article about it, my head spun around and I asked him to slap it up *now* rather than waiting, so there's not much in the way of explanation yet.

[GhostRaptor]

Problem: It needs the explanation to become a relevant talking point. Before anyone can properly discuss those numbers, they need to know under what conditions he conducted his 'experiment' so they can understand what is being discussed. Were all skills checked? Was AoE taken into account? Were specialist effects taken into account (eg if an Elem slows a target with an AoE snare then quite often the DPS from focus fire increases markedly ... but this, obviously, cannot be quantified very easily - especially if most of the added DPS comes from other characters who otherwise weren't dealing damage)?

//edit: and yes, I had a major "doh!" moment a few minutes ago ... Cunning doesn't do any damage by itself, so of course the numbers must've been the extra damage due to the penetration. Can't believe I had to ask for clarification ;).

[Notin]

Couple comments on elementalists:
First, why are Mind Shock/Freeze/Burn missing? Lightning Surge? Those all hit pretty hard, but they aren't listed.

Second, the reason burst damage is good is because it makes healing difficult. Therefore, if you are measuring burst damage, it's best to figure it out in terms of how long the enemy has to heal. Imagine chaining Fireball and Immolate. The healer only really has 1 second to recover from the majority of the damage (not counting the burning here). The cast time of the fireball doesn't really come into play if you cast it first. Note this isn't so much an issue with the table, just the data derived from it.

[Falconer]

Notin, exactly what I was going to bring up...

The other problem is those peak DPS figures don't mean much without knowing total damage (EG: how long the spell takes to cast). Also... the last bit matters as well... the initial casting time as well as the final recharge don't matter. If one fireball up front... and another one 10s later is adequate to the task... great.

Also I feel that the full minute is misleading... as it's pretty clear elementalists are designed to slam something hard with 5-10seconds. Also flare is in my mind, merely a more effective wand attack to use if you got the energy. And even then the energy costs are very misleading... (how many elementalists do you know who leave home without their favorite elemental attunement for example).

And this also avoids the issue of packet-size... Obviously there are many many buffs which simply reduce the amount of damage. The reason that healing hands is so effective in many cases (and healing seed, and mark of protection, etc. etc. etc.) Is because the rapid fire attacks which benefit the most from buffs get nailed the hardest by these. (EG: if you're delivering damage in 50 point packets... you're going to notice that defensive enchant a lot less than if you deliver it in 25 point packets twice as often).

As I pointed out in channel... the problem currently I have with the system is that the synch is too powerfull.. During december BWE 5 rangers it typically took 2 volleys to drop a target. What does this mean... healer watching the health bars... sees healthbar reduced from full and healthy to 25-33% percent almost instantly... He has 2 seconds to react before the next shots slam home... (with 3 elementalists you can get similar results... a nice big skill up front... followed by nice 1s fast casts). With synch the point is landing a singular large amount of damage from multiple sources with no warning ideally and not giving the monks any chance to react.

[Ensign]

The above confuses me. I am reading that as being a case of using Warrior's Cunning against two targets, one in 100AL armour and one in 60AL armour. How in heck does the DPS result in being higher against the 100AL target?

It's an effective 'before armor' DPS. You can pull the armor rating of the target out of the equation and just calculate the damage part, and you can fudge that for armor penetration by making it a rough increase in base damage.

The reason you do it this way is so that you can compare damage across a range of armor levels. When you're looking at a base damage normalized for 60 AL, comparing that to a damage boost normalized for 100 AL is just confusing. So you renormalize the 100 AL damage boost for 60 AL and just list it as being appropriately higher. In fact, you're right - the 100 AL numbers would deal half of the listed damage to a 100 AL target. But so would every other attack on that list.

Basically if you want the numbers for a 100 AL target, just divide everything by two.

Peace,
-CxE

[FrogDevourer]

First of all, thanks a lot for sharing this insightful table. I wish I could read more GW threads like this. That kind of number crunching shows why GW shines when compared to many other online games.

Then to the point. Although I agree with most of your article, there are a few points that could be taken into consideration :



A - AoE skills. As mentioned by Freyas and others, they can dramatically increase the dmg output of elementalists. If I recall correctly my table, using Inferno then multiple Flame Burst can unleash quite a lot of dmg. Of course the elementalist has to be in melee but I'm not taking that into account since we're talking about pure dmg numbers. Note that the AoE effect is also valid for Warriors (cyclone axe, etc...).

Freyas mentioned the circumstancial aspect of this dmg. Although it is mainly true, I think that having at least 1.5/2 targets in average for Fireball or Flame Burst is something that can be achieved frequently. In PvE, it's obvious due to the basic AI. In PvP it can be done, either when the opponent casters stay grouped (your beginner/average opponent) or when someone in your team is fire focused (by say 2 warriors + 1 pet).


B - Range. Having melee dmg somewhat higher than range dmg (and notably spells that can't be dodged) is not that surprising. Most of the time your average warrior won't be standing still beating his opponent to a pulp. He must spend a lot of time running, and this does reduce his output dmg. That's also why point blank spells from elementalists can dish a lot more dmg than spells you show in your table, and notably burst dmg.

Note that melee dmg is also more conditional than dmg from spells. It is quite easy to snare a warrior, many professions can do that (without much effort). Then the output dmg drops dramatically. Disrupting a spellcaster to reduce his output dmg can be done too obviously, but it requires more time and more effort. Basically only a mesmer and stop a caster efficiently, while a lot of builds can reduce the efficiency of a melee character.



C - Does a buffer mean more overall dmg ? Your numbers shows only the dmg dealer part of the combination, but having someone buff your warrior with Judge's Insight means that you have a smitting monk in your group. Is this better than an extra dmg dealer instead ? While buffing is good, how does your table translate into group dmg ?

That's also true for attribute points. Being able to buff oneself (e.g: Conjure) is good, but it requires a high buff attribute. You mention Warrior's Cunning and Power Attack. That means tactics + strength + weapon. That's 3 attributes. A classical elementalist is using only 2 at best (energy storage + one element). He still has a third attribute he can use to increase the group's efficiency (for instance by snaring opponent melee characters).

My point is : buffing is unarguably a good thing but how do you quantify its real cost in term of overall damage output loss ?



D - How does your 60s table translate into a 5 minute fight ? Most buffs have to be recast, that means output dmg are not as good as expected. Moreover using skills quickly drains a lot of energy, both for the elementalist and for the warrior (enchants+power attack). Obviously the warrior seems to have a big advantage for long fights since he can attack endlessly. The same question stands for a shorter fight such as 20s (e.g: the time needed to take down a pesky monk or mesmer), the elementalist can dish his dmg without any preparation : no buff needed, no running needed, it's just instant dmg.



E - Skill rotation. Few elementalists use only one same skill repeatedly, particularly in a 60s fight. You'll most often use muliple skills in rotation. As mentioned by Freyas, it does increase dramatically the output dmg of the elementalist as well as the the energy burnt. Once his energy is depleted, the caster's out of the loop for a while. Basically a good fire elementalist is truly a nuker : he quickly unloads a ton of dmg, enough to kill 50 to 90% of one or two opponents then has to refuel.

Sausaletus Rex told that a warrior could do the same. This is partly true, but a warrior cannot use his adrenaline skills from the start, particularly if he must buff himself first.

Sword War w/ Conjure and Cunning spamming Power Attack, Galrath, and Final - 2653
Time Spent Attacking - 59 (15 w/ Power Attack, 8 w/ Galrath, 6 w/ Final, 30 w/ regular)

If I recall correctly using an adrenaline skills drains one adrenaline point from all other skills so you just can't spam 8 x Galrath, and 6 x Final in 59 seconds (60s minus conjure).

Also note that a warrior cannot afford to spam 15 power attacks in 60s. That's 75 energy. A primary warrior (high strength needed) can use only 60 energy in one minute. And if you take into account the cost of enchant buffs, you have 8 P.Attacks at best.




Don't take me wrong, I'm not arguing about Charles' point, which I tend to agree with. I'm not arguing about Rex's strings of skills examples for I'm convinced he's mostly right. I also think that the dmg you can get from spells is not that exciting compared to the long term capability of a warrior (and notably non-elementalist DoTs).

I just wanted to point out that this table doesn't show all the combat numbers. I'm not talking about flexibility, or metagame or conditional effects or other non dmg abilities. Just about pure dmg numbers.

Sure, a pure Flare elementalist is worth less than a basic warrior, but crunching numbers down to a 60s 1v1 combat doesn't necessarily reflect the true dmg capability of a profession. IMHO, having 4 warriors with buff skills in a 6 people group won't be much better than having 2 warriors and 2 elementalists.

Anyway if you're still reading this I guess I should thank you for bearing my poor and awkward grammar.

I'm looking forward to reading Charles' next theory article.

[Ensign]

I'm sure Charles sees it as some sort of intelligence test. If you can't understand the table without detailed explanation then you're not smart enough to be able to talk about it.

Vicious.

Really it's more of a problem of trying to take a problem with a ton of variables, and fit them into just a couple so that you can compare. In order to do that, you have just have to pick your targets and run with them. Everything besides a few boring damage skills requires a kludge to make fit. Weapons require assumptions about their damage, and their critical hit rate to make accurate comparisons. Anything with armor penetration needs assumptions about the type of target. Buff stacking gets tricky because they cut into each other by preventing you from attacking. Anything with damage over time makes a real mess of things. Then you have conditional damage, you have area of effect damage...

Basically explaining the numbers in that table are an article unto themselves - and one that I'm writing. And I don't just want to talk about these simple numbers, I want to talk about even more useful things, like packing damage, into short timeframes and minimizing the skill slots needed for maximum effect.

So if something isn't explained too well, that's the reason why. It isn't because I don't think these things need explination - they do - but because getting a satisfactory explination for everything is going to take me a week or more, plus refinements, and getting the information out there *now* is more important.


The other problem is those peak DPS figures don't mean much without knowing total damage (EG: how long the spell takes to cast).

What you're looking for there is the right numbers for skills to deliver a 1-2 punch. Basically over short timeframes, the first spell in the chain just uses the aftercast (.75 seconds), and the last just uses the casting time. Everything in between would be limited by the cast/aftercast animation.

However, you'll notice that in any volley of N attacks, there will be N-1 casting times, and N-1 aftercasts. So the net result, for calculating a single burst of damage, is that the very first attack in the volley effectively happens with a 0 casting time - they can't react until it hits so the time you spent casting beforehand really doesn't matter too much.

That happens with everything. Your 1-2 punch DPS fom a Bow is twice what is listed, because the second volley lands two seconds after the first volley. If you take 3 volleys, you get 3/2 the listed peak DPS, again, because the first shot was free.

It's an important note to make, though, because you can pack even more damage than what's listed into extremely short timeframes. The right kludge, though, isn't to affect the peak DPS, but to make the note that the first attack in a volley happens instantly. As soon as your opponent gets a chance to react, you're into known territory.


Also I feel that the full minute is misleading... as it's pretty clear elementalists are designed to slam something hard with 5-10seconds.

What they're designed to do, and what they actually do can be very different things.

But, yes, the Damage Per Minute number is extremely misleading on its own. That's the biggest problem I have with it. If you just look at that number, you come to absurd conclusions like Flare being the best Elementalist skill, or Word of Healing really not being that good compared to Orison.

The Damage Per Minute is *only* useful in the context of Energy and Time Duties. You could add up all the skills you wanted if it weren't for those two, but skills have costs that have to be paid in the form of action time and energy costs. Once you're skills demand more time or energy than you can provide, you have to start cutting back on some of them and that eats up their effectiveness.

In the case of Elementalists, it's those Energy Duties that kill you. Even something relatively cheap like Immolate has an energy duty of 100 - over the course of a minute you only regenerate 80 energy. That's even before trying to synch up with other skills. The Elementalist is a class that's fundamentally Bound By Energy - you're going to run out of energy long before you run out of casting time.

So the DPM numbers for the Elementalist are as misleading as a long chart of attack skills would be if you don't take them in context of duties. In the case of attack skills, you're trying to maximize your damage by fitting things in based on their time duties - in the case of the Elementalist, you're trying to fit things in based upon their Energy Duties.

What's perhaps surprising about that result is that to maximize your damage over a long timeframe, a Warrior or Ranger, bound by time, is more concerned about maximizing the burst DPS of their skills - skills with high burst DPS give the best returns on damage per time duty - while an Elementalist is more concerned with burst EPS, how quickly they're burning off their bar. Which is pretty much the opposite of what these classes are looking for in the short term.


(how many elementalists do you know who leave home without their favorite elemental attunement for example).

I actually don't know a single Elementalist that runs an Elemental Attunement, since you asked. Elemental Attunements aren't anything special. Using an Elemental Attunement is a commitment to spending every single drop of your energy on skills of the attribute in question - use even one skill from your secondary class or a different element, and you're likely to be better off just using Glyph of Lesser Energy. Basically they force you into an extremely narrow tactical scope, and if you want to stray from that you've wasted a skill slot. There are builds that can make use of them, but I generally don't bother - there are just better energy management options out there.

But, again, this does bring up an important point - energy duties and time duties are not set in stone. Even if Elemental Attunements aren't all that good, how many Rangers decline to use Expertise? That knocks a big chunk off of their energy duties. How about Fast Casting? That affects the Time Duty of every spell you cast (and not in a way that's easy to calculate). Since Duties are what are ultimately crimping your long term effectiveness, skills that help you service them, be it by lowering cast times or giving you more energy to work with, are potentially extremely valuable.


the problem currently I have with the system is that the synch is too powerfull.

That's the big problem you face when you get DPS up into these ranges - killing someone before they have a chance to respond. Elementalists are actually the least dangerous in this regard, because their casting times are naturally so high - they can kick out a solid first hit, since those effectively ignore casting times, but they can't kick out a good followup quickly. Your best is something like Shard Storm or Lightning Strike, and each of those is worse than just shooting someone with a bow. Weapons are more dangerous in this regard because of their fast refire rates. If you can get a significant damage per hit out of a weapon (and you sure can by stacking buffs) you get knock people out with a volley of 1-2 punches, and that's where you have problems.

How you address this is different for the various classes. You need to make sure that Warriors and Rangers can't deal appreciable burst damage, so they require at least three volleys to drop someone. As for Elementalists, they need to have excellent damage per skill, but you need to make the casting times long enough that healers have a chance to respond to the first volley. Elementalist damage needs to be high to justify the energy and time you spend casting the spells - but as long as you can't knock someone out with a single hit, it's fundamentally more fair because you can't follow it up with a knockout punch.

So basically, if you want to make changes, tone down the damage buffs that weapon users can stack, and make sure that the most effective skills for an Elementalist are the big nukes. 1 second casting time skills are potentially dangerous, 2 second cast skills are borderline, but anything with a casting time of 3 or greater is going to be hard to abuse, just because people have so much time to react to it. Just be sure that they can't drop a target in a single volley. =)


A - AoE skills. As mentioned by Freyas and others, they can dramatically increase the dmg output of elementalists. If I recall correctly my table, using Inferno then multiple Flame Burst can unleash quite a lot of dmg.

Yeah, PBAoEs have the best burst damage in the game, in the mid/low 50s at level 12. That isn't too impressive on a single target (well, as it stands - 50 burst DPS *should* be good), but it's clearly within the range we care about.

Area of Effect skills are definitely part of the equation - dealing damage to multiple targets is the easiest way to crank up your own damage output. Things start to get a bit fuzzy as far as actual effectiveness goes, because divided damage clearly isn't as valuable as focused damage. 50 damage on a single target is a whole lot more valuable than 25 damage on two targets - and not just considering focused fire concerns. A character's health is just a buffer between life and death - it doesn't matter if a target has 1 health or 1000, they still cast spells and deal damage at the same rate. Similarly, damage is only as scary as its threat of taking away that last health point. Health, fundamentally, just buys a target time, and by dividing up your damage you're giving an opponent a ton of it.

Basically, AoE damage can give you some nice damage / time figures, but they just aren't as effective as knocking out a single target. Area of Effect skills need to scare people - they need to knock a noticible chunk of health off of multiple targets, not just a bit of damage that they can shrug off. That's the only real problem with AoE's in Guild Wars at the moment. Generally they need a damage buff so that they're actually an *attack* on everyone within the blast radius, not just an annoyance.


Note that melee dmg is also more conditional than dmg from spells.

Inherently so - that's why Warriors have a naturally higher DPS when attacking than Rangers - they need it to make up for all of the time they spend running around. Just because of that melee damage is fundamentally more fair.

There's another concern that melee damage has - packing density. Unlike Rangers or Elementalists with ranged attacks, Warriors and PBAoE Elementalists have to be next to a target to deal damage to them, and you aren't going to get more than 3 guys around a single target with any consistency. Even if you found an absurd Warrior DPS combo, you can't put more than 3 or so of them in a group without them stepping on each other. In that way Warrior damage is a bit more 'fair', in that a broken Warrior combo won't spawn 6 Warrior teams - but it is an additional constraint.


C - Does a buffer mean more overall dmg ?

Generally I'm assuming that characters will be buffing themselves since that's what they end up doing in game. Warrior/Elementalists with Conjure, or Warrior/Monks with Judge's Insight (and Scourge Healing) are pretty common sights. Self buffing characters are perfectly reasonable with 3 attribute builds (Strength + Weapon + Other).

I don't know where Tactics fits into any of this. People use Tactics? I guess on a Monk or something. Meh.


D - How does your 60s table translate into a 5 minute fight ?

Generally you're going to see that in the Energy Duties. Generally you're consuming your energy bar as the fight goes on so you end up being bound by energy in the long run.

A five minute fight isn't a particularly good timeframe for reference, though. While a particular match may run for several minutes, a given battle never lasts more than a couple of minutes, and the winner is usually decided in well under a minute. I'd say that in a straight up fight, one team gets enough of an advantage to push to victory in under thirty seconds most of the time - someone's healers break, a teammate goes down, and one side has to concede the battle in hopes of winning the war.

Also, remember that on most PvP maps teams auto-resurrect every two minutes, with full health and energy. That more than anything puts a limit on the actual, meaningful length of any battle.


E - Skill rotation. Few elementalists use only one same skill repeatedly, particularly in a 60s fight.

Right, and the way you model this, at least in the short term, is to string together burst DPS, with the first hit being effectively free, until the energy boundary hits you.

The same is done with Warriors and Rangers, who rotate attack skills to kick up their damage output.


Also note that a warrior cannot afford to spam 15 power attacks in 60s. That's 75 energy.

That's only true if you're relying entirely on your base two pips for your energy supply. Any competitive Warrior has a +1 energy per hit weapon, which provides 45 energy per minute, and that's on top of any other energy boosts you might want to use. A Warrior has at least 100 energy to use in that first minute, clearly enough to fit in constant Power Attacks on top of your buffs.

Remember that part where I said that good equipment fundamentally changes how you build characters? There's your example.


I just wanted to point out that this table doesn't show all the combat numbers. I'm not talking about flexibility, or metagame or conditional effects or other non dmg abilities. Just about pure dmg numbers.

Of course it doesn't. This table is just a tool, just as casting times, energy costs, and all the other numbers listed in a skill description are tools for figuring out what a skill does.

Every skill behaves differently. You have different damage types, stacking damage, different boundaries, different interactions with each other, the target, multiple targets, types of targets. There are too many parameters to make a generalized equation of skill effectiveness or whatever.

What we're doing is creating a toolset for figuring out how effective a skill is going to be in a build, or how effective different builds are going to be compared to each other. Ultimately what you have to do is take two skills, put them next to each other, and analyze them in depth with their tradeoffs. But ultimately, you compare the two skills in similar ways to what we've been discussing in this thread, and you pick out the best skills for comparison by looking at the general numbers and finding skills that look promising.

We aren't filling in details at this point, we're barely laying the groundwork.

Thanks to everyone for the replies, the more angles we can discuss techniques for analysis from, the better our toolset will be in the end.

Peace,
-CxE

[Sausaletus Rex]

C - Does a buffer mean more overall dmg ? Your numbers shows only the dmg dealer part of the combination, but having someone buff your warrior with Judge's Insight means that you have a smitting monk in your group. Is this better than an extra dmg dealer instead ? While buffing is good, how does your table translate into group dmg ?

Right, buffs keep stacking up. And they don't necessarily have to be cast by the attacking. The skill listed largely do, you can't conjure or prep someone else, but you can cast Judge's Insight or Strength of Honor or Barbs or Order of Pain and lots of other things to help people out. But the point there would be that you can actually have a character with Conjure, prep, Judge's Insight, Strength of Honor, Barbs, OoP, OoV and any other buff you might like - to say nothing of the AoE damage hexes like Mark of Pain or Thunderclap (which doesn't deal damage but that's another thing not being discussed here - side effects. The Burning/Blinding/Knockdown/Energy loss and more on many skills also matters. That's why people use Maelstrom, after all. But it's not really important in a discussion of how to maximize sheer damage best.) - if only for a few brief seconds at a time. You can have one stance, one prep, and any number of enchantments, any number of hexes on your target, and things like the Orders and Rituals. It all adds up, pushing both burst damage and damage over time through the roof.

That's the real lesson here. Stack buffs and other damage adds and maximize your DPS. That's the real way to kill things. Not lobbing Fireballs and Fire Storms at them.

That's also true for attribute points. Being able to buff oneself (e.g: Conjure) is good, but it requires a high buff attribute. You mention Warrior's Cunning and Power Attack. That means tactics + strength + weapon. That's 3 attributes. A classical elementalist is using only 2 at best (energy storage + one element). He still has a third attribute he can use to increase the group's efficiency (for instance by snaring opponent melee characters).

Not really, although your point is largely true. The gold standard for a Warrior or Ranger right now would not be merely Strength+Weapon as high as possible plus whatever's left over. Runes and other things make possible 12(11+1)/12(10+2)/10 builds. 3 full attributes is makable, whether it's Tactics (That'd be 12/12/11, actually) or Blood or Healing or Smiting or Death or Domination or whatever you want for that thired attribute. Within those three attributes there's room to include things that also increase group efficiency. Swordsmanship has Hamstring, Tactics has defensive stances to lower the burden on your healers, Strength has speed boost to keep you in range, and so on.

The question isn't so much how many attributes are you putting into your role but how much you need to do outside of that role. If you're a Warrior all you need to do is keep swinging away and keeping in range. That's your job, that's what you do, anything else is a bonus which you might or might not need based on the circumstances. You're there in the party to kill people. Same for the Ranger and for the Elementalist (At least in their purest incarnations). If you're dong that job efficiently what else do you need to do? That's what other characters are there for. You don't need a snare on your main damage dealer, you need a snare on someone else so the damage dealer doesn't spend the time to snare away from doing damage.

It's all well and good to have a well rounded and versatile character capable of filling many slots but what we're talking aobut here is the best way to build a killing machine.

How does your 60s table translate into a 5 minute fight ? Most buffs have to be recast, that means output dmg are not as good as expected.

That's why the time duty and energy duty tables are there. The recharge and recycling times are included in the DPM measure, it's a measure of how much damage you'll do if you cast a skill as many times as possible including all relevant casting and recharging times over a full minute. To use a different time frame DPM becomes DPT and you make T=5 minutes, that's just 5xDPM. And 5 times the time and energy duty as well. A skill with a time duty of 20 over a minute becomes a time duty of 100 over five minutes. A skill that takes 50 energy over five minutes takes 250 energy over five minutes (And Warriors regen 200 energy, Rangers 300, and Elementalists 400). You can do the same thing for 10 minutes, for 30 seconds or 3 minutes 46 seconds. 60 seconds is just an arbitrary number. Most battles aren't goign to last that long, at least not the important parts...


Vicious.

I learn from the best.

Basically explaining the numbers in that table are an article unto themselves - and one that I'm writing......It isn't because I don't think these things need explination - they do - but because getting a satisfactory explination for everything is going to take me a week or more, plus refinements, and getting the information out there *now* is more important.

Which means roughly a month or more in Chuck time, I'm sure. Which is why I asked for the table now. Yes it's dense and arcane but, hey, we've got lots of things that are simple and easy to explain laying around for those who can't hack it. We also want to have the sort of deep and complicated information that not everyone is going to understand or even care about at this site.

But, again, this does bring up an important point - energy duties and time duties are not set in stone. Even if Elemental Attunements aren't all that good, how many Rangers decline to use Expertise? That knocks a big chunk off of their energy duties. How about Fast Casting? That affects the Time Duty of every spell you cast (and not in a way that's easy to calculate). Since Duties are what are ultimately crimping your long term effectiveness, skills that help you service them, be it by lowering cast times or giving you more energy to work with, are potentially extremely valuable.

Right. And what about things like Flurry with Warriors? That doesn't just make their normal swings better or their Conjure better but any other per hit buff you add to it.

Anything that plays around with energy or casting or recharging is going to affect the DPS of things.

[Pharalon]

Anything that plays around with energy or casting or recharging is going to affect the DPS of things.

As will anything that effects swing rate, %chance to hit etc etc etc. You can't just look at the table and draw out absolute truths. Sheild of Deflection will destroy a warrior or ranger's DPS, in the same fashion that daze will cut a Ele's burst DPS by two thrids. Warriors have to spend time closing to melee, arrows can be dodged, spells can be interrupted. If you worry about too many variables at once, all you end up with is a headache. If you want to extract meaniingful conclusions from the sea of data, you need to keep things as simple as possible, and eliminate (or conviently ignore) as many variables as is reasonable.

Once you have some basic principles to work off, which is what the work Charles is doing here is moving towards, you can bring in added levels of complexity, that will bring the numbers floating around here closer to what yoiu would expect to occur in-game.

[Ensign]

That's why the time duty and energy duty tables are there.

Right, and parameters change once you've spent your energy reserve. The first minute of the battle a Monk has 40 + 80 regen energy to play with, but just 80 in each minute after that. Or if you contract the battle more, 80 energy in the first 30 seconds, and 40 every 30 seconds after that. If you have energy to burn, time duties end up being most important, but once you're on regen it's the energy duty that's going to keep you from running wild.


I learn from the best.

You hear that Blackace? Saus just conceded.


Which means roughly a month or more in Chuck time, I'm sure...We also want to have the sort of deep and complicated information that not everyone is going to understand or even care about at this site.

I'd guess finishing it up around the week after the BWE at this point. There's a ton to talk about, and a lot of issues I hadn't thought about have been raised in this thread.

Which brings up the other good reason for putting up a table like this early - it gets people to talk about other issues that perhaps haven't been addressed yet, or points out issues that need more clarification. Basically it's a way of getting feedback on an article that hasn't even been written yet, and that's invaluable. Thanks guys.

Peace,
-CxE

[Keramon]

Fire Elementalist - 1869
Time Spent Attacking - 60 (17 w/ Fireball, 17 w/ Immolate, 26 w/ Flare)
Energy Spent - 169
Damage per en - 11

Earth Elementalist - 1566
Time Spent Attacking - 42 (24 w/ Obsidian Flame, 18 w/ Stoning)
Energy Spent - 236
Damage per en - 7



Shouldn't conjure flame be added to the damage of the spell casting. Ok ... I havn't tried this, but I thought that stacked with all attack damage. I am guessing since it has not been mentioned, that that GW does not consider spell damage as attack damage? I guess this may be a fine distinction that I had not made to this point.

Conjure Flame - (10,1,60) Lose all enchantments. For 60 seconds, your attacks strike for an additional 18 fire damage.

Of course all of this information is really quite elementary (excuse the pun) ... and needs to be taken into context of a battle and all of the variables that can happen with PVE being quite different from PVP. Smart players/Good team vs Bad opponents.

All of this information is very useful and will help many in making decisions relating to build and maximising damage potential.

Much of it needs to be taken into perspective and seen how they can fit in your build ...
Many of the people considering Wa/El will consider the PBAOE and and higher levels of the element of their choice to get a better conjure vs higher strength for a longer/more efficient warrior's cunning vs higher tactics for better Healing Signet.

Its just this balance and the multitude of different options/ways of dealing damage, combinations and playstyles that make it interesting. I am sure that there will be well published "cookie cutter" options in no time, but there will always be variables that people consider that will change the impact/balance of power in any fight.

[Scrivener]

Shouldn't conjure flame be added to the damage of the spell casting. Ok ... I havn't tried this, but I thought that stacked with all attack damage. I am guessing since it has not been mentioned, that that GW does not consider spell damage as attack damage? I guess this may be a fine distinction that I had not made to this point.
Wherever you see "attack", it means using a weapon. Then it breaks down into melee (warrior weapons) and ranged (caster and ranger weapons).

[Blackace]

You hear that Blackace? Saus just conceded.


So all my early days of being bad on TGH forums has finally paid off? Sweet :p

[Notin]

I must say, that's some 1337 DPM on flare.

[FrogDevourer]

Right, and the way you model this, at least in the short term, is to string together burst DPS, with the first hit being effectively free, until the energy boundary hits you.

The same is done with Warriors and Rangers, who rotate attack skills to kick up their damage output.
Yes. That's why I tend to do the same kind of dmg computation as Saus: using skill rotation. There are a lot of combinations but most of the time you'll end up with similar numbers.

Remember that part where I said that good equipment fundamentally changes how you build characters? There's your example.
As mentioned multiple times in this thread, what you're trying to find here, is a good way to compare dmg dealing capabilities in a simple way. Introducing side effects such as equipment will lead to a pandemonium of arguments. Sure, there are broken mods for warriors but why not taking about account other mods as well.

Basically I'm in favor of pure dmg dealing comparison (i.e: pure killing machine potential) without any outside interference. No external buff, no equipment. When this case of study has been beaten to death, we can talk about equipment or group dmg potential.

The question isn't so much how many attributes are you putting into your role but how much you need to do outside of that role.
I'm not sure I understand your point Saus. Having one spare attribute doesn't mean you won't put it into that role. Many non dmg dealing abilities DO have an effect on your output. A classical fire elementalist has one spare attribute. He also has more attribute points left because his dmg output is not based on an exponential of 12. These spare resources can be used to increase the dmg output of the group (e.g: BiP) OR even his own dmg output (e.g: energy management from inspiration). I'm talking about pure dmg output, not about narrow hexes or metagame effects.

I'm aware that it cannot be quantified easily. It's probably far beyond the scope of the current article. My point was that in Charles' article, we should take into account the fact that the warrior killing machine is using 100% of its potential to do dmg, while the elementalist has some resources left.

Let's go back to the snare example. The best killing machine warrior in the world will have his dmg output reduced if he must run after his target. Note that if your target is running, you won't build 1 adrenaline per second either.

To reflect this, an average ratio should be taken : how much output dmg do you lose by running after a moving target ? 50% ? More ? How much do you get if your target is stopped (knock down) or slowed ? +10% ? more ? If you don't take a precise ratio into account, melee numbers may be totally irrelevant, because your aren't killing a sitting duck.

Area of Effect skills need to scare people - they need to knock a noticible chunk of health off of multiple targets, not just a bit of damage that they can shrug off.
You have a point. But a strong side effect of AoE spells is that they can create chaos and confusion. Most healers can lose a few seconds if they see 3 health bars going down at the same time: they can waste time watching bars more precisely, or watching the battleground to see who's fire focused, or they could be busy healing the wrong target.

Basically, a couple big AoE dmg (Inferno+Flame Burst) could mean one less Word of Healing on your main target. That could be the time window you need to kill this target.

A five minute fight isn't a particularly good timeframe for reference, though.
Note that I mentioned a 20s time frame too. My point was that you currently are considering burst dmg (1s) which would be the first or killing blow, and 1 minute dmg. You also show numbers for a full melee dmg output, not taking into account the time need to be in melee range.



Then another very important point is the "threat potential".

A skill which does 10 times more dmg on 20 targets is pointless. So is using slow DoT. Their threat potential is close to zero since they are easy to counter and since they won't dent the life buffer of your target.

If you can do more dmg (in the long run) than natural regeneration + energy need for healing, you have a good dmg ability, but a very low threat potential. Your damage is very predictable. The healer(s) won't be on their toes. They'll just have to wait until the health bar goes under 50% to throw a WoH.

If a couple of elementalists can pack a lot of dmg to kill their target before the monk can save him, the skill numbers can be worse in Charles' table, but it may have a much better threat potential. Healers won't be able to predict the next blow so they are in unknown territory.



I would classify the damage into 3/4 types :
- Killing blow (3 to 6s)
This configuration is pretty obvious, you're dealing as much dmg as possible, burning everything you have (full skill rotation). You're just looking for good threat potential. No energy pip nor skill recharge time nor running time. To that extent, big dmg skills with exhaustion should be taken into account as well.

- Short term damage (20 to 30s)
With this time frame, casting various buffs and running after your target is not as exciting as dealing a lot of spell damage. The elementalist will probably hit the energy boundary. On the other hand, the warrior won't be able to swing his sword 20 to 30 times. Time and energy duties become relevant, but they may not be as important as a final blow from a high threat potential exhaution skill.

- Long term damage (60 to 120s)
In this configuration you have to recast buffs, and you'll have to rely mostly on your energy/adre regeneration. That's when duties become predominant.

and may be :
- Continuous damage (more than 2min)
Both teams are perfectly balanced, and the fight lasts for more than 2min without any casualty. In this configuration, the energy management becomes critical, you don't take into account the initial energy pool. Threat potential is mostly irrelevant. Time/energy duties all the way.


EDIT : Is anything we're talking about in this thread (or more generally in this forum) forwarded to A.Net through the alpha forum ?

[Notin]

Is anything we're talking about in this thread (or more generally in this forum) forwarded to A.Net through the alpha forum ?

Anything that's on the alpha forums falls under the NDA, so even if alpha testers knew, then they couldn't say anything.

[Terphin]

I know i dont post much but just wanna say you guys are doing some excellent research about all this stuff. Its great info.

[FrogDevourer]

Anything that's on the alpha forums falls under the NDA, so even if alpha testers knew, then they couldn't say anything.
I'm not asking for details, I'm just curious to know if he're just discussing for the sake of discussing, or if it could actually be useful. I guess it doesn't matter anyway.

[Sausaletus Rex]

I learn from the best.

Now, Charles, how arrogant of you. I neither said I learned from you nor did I say whether the "best" were those best at being good at what they do or the best at being awful at what they do. You learn more from failure than from success, after all, and I think in studying you people could learn a wealth of information indeed.

Shouldn't conjure flame be added to the damage of the spell casting. Ok ... I havn't tried this, but I thought that stacked with all attack damage. I am guessing since it has not been mentioned, that that GW does not consider spell damage as attack damage? I guess this may be a fine distinction that I had not made to this point.

Scrivener's right. Conjures don't add anything to spell damage. They're not considered "attacks" by the game, only weapon swings are. Attack skills like Power Attack which take place within the swing of a weapon are counted as attacks so the Conjures stack with those.

I'll also point out that - I hope - I didn't include Conjures with Fire or Earth Elementalists. The former because it's more effective, in terms of damage at least, to use Flare to blast away in place of a wand/conjure attack. And the later because Earth doesn't have a Conjure. Sure, you can get one by mixing and matching your attribute lines but a pure Earth damage dealer with ES/Earth/Something else isn't going to have a Conjure available. Probably the only thing keeping Earth from being hands down the best attribute line for an Elementalist. I still think it is but it's not running away with it.


I'm not asking for details, I'm just curious to know if he're just discussing for the sake of discussing, or if it could actually be useful. I guess it doesn't matter anyway.

As Notin said the alpha board are strictly covered by the NDA. We can't tell you anything about them from what any and all of the boards are named to how many posts there are, it's all off limits.

However, that doesn't mean it doesn't matter. We can't tell you what's being said but at the very least there are alphas on these board (quite a few of us, actually) and it's not a big secret that the developers keep an eye on fansite forums, too. At the very least what's being said here is influencing those alphas and devs that are *here* and may influence our views on things in futher discussion elsewhere even if we don't discuss things specifically. The thoughts and opinions from non-alphas are taken into account and valued, have no doubt of that.

[cpukilla]

Could you at least tell us if you ARE talking about our ideas on the alpha boards? It seems a little wierd that you could discuss things with us here, go back to the alpha boards and change something based on it, and we would never know. In fact, if we found out someone would be getting in trouble! Darn NDA, grumble grumble. :(

[Narcism]

Could you at least tell us if you ARE talking about our ideas on the alpha boards?

No, no we can't.

[Weezer_Blue]

haha. narcism you're evil.


anyway... i guess from all this that I should sub in Read the Wind instead of Favourable Winds. Or is the speed increase on Favourable Winds much more dramatic that RTW so that I should hold on to it? It doesn't matter how many buffs you've got if your arrows won't land where you want them (in a warrior's forehead)

[Narcism]

Ditch Favourable Winds?

[Sausaletus Rex]

I shall not concede!

haha. narcism you're evil.

Narc's bush league. Not evil. Of course, he's in Agony so that tells you all you need to know right there.


anyway... i guess from all this that I should sub in Read the Wind instead of Favourable Winds. Or is the speed increase on Favourable Winds much more dramatic that RTW so that I should hold on to it? It doesn't matter how many buffs you've got if your arrows won't land where you want them (in a warrior's forehead)

RTW isn't all that good. Over time you end up spending so much time and energy to keep it up that it's just as easy to use another attack and up your DPS - or DPM, DOT, XFL, KMFDM, or whatever one of Chuck's acronyms you prefer - that way. You'll get just as much from an extra Power Shot every 15 or whatever it is these days seconds than you will from having RTW on everything over that time. Unless you really want to pack as much damage into those 15 seconds as possible, it's forgettable and so are most preps for damage.

Favorable Wind has its own problems, of course. The real switch you want here is to go Ran/Ele, go Exp 12 (11+1), Marks 12 (10+2), Elemental of your choice 10 and pick up a Conjure. That's 60 seconds of +damage on all your bow attacks for just, what, 5 energy? Sold. Combine that with FW or better yet with FW and RTW and there you go.

[Weezer_Blue]

Thanks. I guess I'll stick with Favourable Winds. I can get more attacks off over a longer period of time as well as get in a bunch of power shots + conjure lightning. My current build looks like this:

Hunters Shot
Power Shot
Pin Down
Distracting Shot
Favourable Winds
Marksman's Wager {E} (going to see if I need this - probably not)
Enervating Charge
Conjure Lightning

Hunters Shot adds bleeding to running foes, and from having my artery severed countless times I can tell you that's quite annoying. Power Shot is obviously just for the pure damage. I was considering Precision Shot too but it has a higher recharge time in place of high energy and it doesn't miss as easily. Distracting Shot is there rather than concussion shot because Concussion only works on spells. And I like taking down warriors the most (they're fun to kill). Favourable Winds keeps everything as accurate as possible for long periods of time (while the match lasts a long time, the major battles usually don't take very long and one team ends up retreating by the time it runs out). Pin Down is obvious. Enervating Charge causes a little damage but more importantly it causes weakness. And Conjure Lightning is obvious as well. If I don't need Wager, I could put Glimmering Mark in its place. Some people have pointed out blindness but I'm trusting to a good monk for this build so i don't need Antidote Signet.

...But this isn't a strategy discussion board anyway so er... yeah...

[Ensign]

In fact, if we found out someone would be getting in trouble!

Not really - you can figure out which forum threads are being talked about by following the external links on the alpha forums. If a given thread here is getting a lot of hits from forums.guildwars.com, you can guess that alphas are talking about it.

Just saying.

Peace,
-CxE

[cpukilla]

*sniff alphas hoarding knowledge! :p

Charles, is there an easy way to do that? I'm not quite tech savy enough :confused:

[Ensign]

Not unless you have access to server logs.

Does it really matter? If an issue is hot in the community, you can be certain that testers and the developers are aware of it. The test is a community tool, filled with people from the community. Alpha testers who post on said hot button threads. Connect the dots.

Peace,
-CxE

[cpukilla]

No, it doesn't, and I know already that anet listens to fan feedback simply by the choices they make. But it would be cool to know if they changed something based on a comment you or someone else made, even if its a selfish kind of curiosity ;) .

[Ensign]

I've been thinking about area of effect damage, and just how valuable that is compared to focused fire damage. On the surface, it seems simple - if a skill hits two targets instead of one, you've dealt twice as much damage, which is twice as good.

Not so fast.

This question really gets to the heart of what health and damage is. Your goal when you attack someone isn't to deal damage to them - it is to kill them. The difference is subtle. If I have 480 health, I really don't care if I lose the first 479 health, it's that last point that I care about. The rest of the points are just buffer, health that defends against your attacks until you land a killing blow. If I can kill, shut down, or otherwise defeat you before you take that last point, it really doesn't matter how much damage you dealt me, because I won. In the same way, dealing damage to your opponent isn't a goal unto itself - it's merely a mechanism for setting your opponent up for a killing blow.

What does this have to do with area of effect damage? Area of effect damage is bad for the same reasons that focus fire is good - it kills people faster, and even if it doesn't kill it puts the fear of death in people that makes them change their behavior and worry about survival instead of doing whatever it is they normally do. If you spread your damage around, your enemies can keep doing whatever it is they do longer, putting off their fear of death and instead executing their gameplan, their plan to beat you. Clearly, damage to two targets isn't as valuable as double damage to one target. But how much worse?

The answer to this lies in how much effective time you grant your enemy by dividing damage instead of concentrating.

Let's take a simple example - you're a Warrior attacking two identical targets, and the entirity of your plan is to hit them with your sword until they die. Pretty good plan, for a Warrior. Ahem. Anyway, let's say they don't have any sort of self healing, and you can drop both of them in ten seconds - five seconds if you focus on them one at a time. While you're beating on them they're doing something back to you, be it damage or energy denial or whatever else enemies do. What do you know, you just hit things with swords, maybe an axe if you're feeling particularly adventurous.

Ahem.

Ok, so every five seconds, the enemy does X stuff to you. Now you have a choice. First, you can choose to focus fire them - killing one, then the other. If you do that, the two enemies will do 2X worth of stuff to you, then one drops. Then the remaining enemy will do another X while you finish him off. So overall, focusing fire on one of two enemies allows them to do 3X worth of stuff to you.

Or, you could choose to divide your damage, one hit to one, then one hit to the other. If you do things this way, both enemies will deal X every five seconds until they both die simultaneously at the ten second mark. If you do things this way, they'll get to do 4X back to you.

So dividing your damage up between targets is only 3/4 as effective as focus fire, since the time to kill is identical and the amount of stuff they can do back to you is all that changes, really.

Now, what if you were hitting both of them simultaneously? If that were the case, you'd be able to drop both in five seconds, and they'd only be able to deal 2X damage back to you. Area effect damage is 3/2 times as effective as focused fire on a single target.

You can repeat this process for more targets, and in doing so you get a simple rule - every additional target beyond the first is worth 1/2 as much as hitting a single target.


So as far as simplistic models go, when calculating damage to multiple targets from an AoE attack, add 50% to your DPS figure for each additional target. If an AoE attack would deal 20 DPS against a single target, it would deal 30 DPS to two targets, 40 DPS to 3, 50 DPS to 4, and so forth.

This is just a model for effectiveness. Clearly, the actual damage you're dealing has gone up as you'd expect, +100% damage per target. But from the perspective of killing opponents and denying them the chance to counterattack, your effectiveness only goes up 50% per target. There's more to area effect damage than that - it's worse when you take healing into account since it gives more targets the ability to take care of themselves, but it also has side benefits in causing the fear in enemy monks, from dropping multiple targets at once and forcing them into triage. Plus many more effects - clearly this isn't comprehensive.

Still, for the purposes of working area of effect skills into the equation and comparing how effective they are, it's a good approximation.

Peace,
-CxE

[Blackace]

I can see this is going to end up getting into what Damage,Health and Time really do for you in Guild Wars.

[Ensign]

Yeah, that's the idea. All the different resources in the game interact, they trade with each other, and ultimately some trades are good and some are bad. It isn't so simple that you can just assign a raw effectiveness number to every skill or anything, and ultimately you still have to execute. But what we're doing here is figuring out ways to look at the game, to lay a foundation of concepts and valuations from which more advanced strategies come.

Figuring out relationships between, say, energy and health are pretty straightforward. The hard resource is time. Just how valuable a second is depends so much on advanced strategy, on just what you can do in a second and how much more valuable time is now than later. It's an incredibly difficult subject that will never have a straight equivilency like some of the other resources, like 1 energy = 1.5 armor = 12 health.

So, naturally, time is the resource that interests me the most, and the one that's going to devour me as the game evolves.

Peace,
-CxE

[cpukilla]

I disagree somewhat with what your saying... I think it depends on what you concentrate on, the enemy's damage output, or healing ability. If I focus fire one enemy, then the healer can either get outpaced trying to heal him, or use protecting enchantments and/or stances to keep him alive. Healing generally won't cut it, but enchantments can if they are recast after being removed. This is energy efficient, and can (sometimes) save the ff'd player. But if I aoe 4 enemies at once, or try to keep hitting them with aoe, its no longer viable to use protection, you need to use healing to bring them back up to health. So focus fire vs aoe change the kind of monk you need, protection or healing. Healing is generally done with single target spells, and to do aoe you either need to be close (heal area, divine healing) or pay 25 en for heal party (inefficient). Aoe attacks are a very good way to drain the healer's energy supply, while single target attacks are better to deplete the protector's energy and/or force long recharges on some skills.

Heres what I am going for : while aoe can take off much more health than focus fire, it lacks the killing power. So why not open up with some good aoe's, then focus in on a weakened player you want to focus fire into oblivion. With any luck the enemy team will have too many hurt players to heal them all at once, and you can pick the target you want dead, and focus him before they can react. This causes confusion in the other team, it weakens the entire group, and it makes focus firing even more effective.

[Nefser]

If you were lucky (skilled?) enough to aoe 4 enemies more than once...very nice. But at that point, a) you're probably rather low on energy and b) have 4 rather pissed-off and partially damaged players looking to put some big hurt on someone. You've also certainly got the attention of a healer who may be getting ready to heal one or more of those players, especially if any one of them starts to get hit a bit more...which is what you're just about to do.

If they're all willing and able to turn and attack you...you're toast. You will not be able to surive multiple attacks on you from 4 targets while you busy yourself trying to finish off target #1 (partially damaged).

If they don't all turn and attack...well, lucky for you. But in that case you're still only going to be able to focus on 1 of those targets and I wouldn't be surprised if at least two of them turned on you.

But lets look at this another way: Let's say you had a different build with good, single-target attacks. You unload the same amount of time and energy (or less?) with those attacks on one of those '4 targets sitting temptingly close in AoE range'. Unload and keep unloading until one is dead. While this is occurring, it may be that only one target (the one being damaged most likely) will have turned to attack you. Perhaps two. This now gets you in the same situation as Charles mentioned previously, but you've probably got the first target rather well 'in hand' by now and will be looking to finish that one off and start on #2.

All in all AoE's are more of a tactical or defensive action, not a high-damage set-up for the killing blow.

AoE's are a tough call - you want to hit at least 2 and really 3+ targets to get the most BfyB. But does any player (esp. and El/*) really want to get the undivided attention of 3+ enemies? Probably not. :)

[FrogDevourer]

This is just a model for effectiveness. Clearly, the actual damage you're dealing has gone up as you'd expect, +100% damage per target. But from the perspective of killing opponents and denying them the chance to counterattack, your effectiveness only goes up 50% per target.
I think this AoE model is a good start.

Then how do you adress the loss of efficiency which is due to range effects. More precisely how much melee damage do you lose by running after your target ? Do you make one swing every 2/3s instead of 1.75 ? Is this ratio weapon dependant ?

Then optionally, I'd ask how much "running penetration" damage you get by using slow/stop effects. Do they translate into +xx% damage versus running targets ?


Figuring out relationships between, say, energy and health are pretty straightforward. The hard resource is time. Just how valuable a second is depends so much on advanced strategy, on just what you can do in a second and how much more valuable time is now than later. It's an incredibly difficult subject that will never have a straight equivilency like some of the other resources, like 1 energy = 1.5 armor = 12 health.

So, naturally, time is the resource that interests me the most, and the one that's going to devour me as the game evolves.

I wonder if the value of one second is as universal as one armor/health. Two seconds could be worth 10 burst energy and 3 long term energy. Continuous effects (e.g: DoT) can have a very good time/dmg ratio but they are not as threatening as a big 2s punch.

[Freyas]

Well, when you start looking at the effect on players and teams, AoE can be very helpful when hitting multiple targets, better than the same amount of damage on one target in many situations. From my experience playing a healer, I can keep any one target alive through an entire team of warriors, so long as that target is not myself, and the team does not use rend enchantments more than once or twice(though with another healer/protector, even rend can be handled). However, once you make the healers spread their attention to other players, it makes it more difficult for them to keep your priority targets alive.

The best tactic for damage unless you are able to kill a character before the healer can react/cast their spells, is to deal massive damage to two or more targets. Enchantments can protect one target so that they are basically invincible, but if you're forcing the other team to spread their enchantments out among several targets, those enchantments lose much of their efficiency, and you're much more likely to be able to break through the opponents healing.

One strategy that I've seen to be very effective is to focus one target, and after they get a couple enchantments switch to another target. This is much easier and more effective using Rangers and Elementalists, as target changing with warriors takes time to move into melee range and get around any other players in between you and the new target. On the other hand, Rangers and Elementalists can immediately change to a second target while losing little or no DPS.

As Charles said, the point of damage is not to remove the first 479 health, but rather the last 1. However, you have to factor in the strengths of the healers versus different types of damage. With something like Healing Seed or Healing Hands on a target, the more people you have attacking the target, and the faster they are attacking, the more effective those enchantments become. However, this can be completely bypassed by changing targets as soon as they recieve that enchantment- the healer now has less energy and 25 seconds before they can recast the enchantment. The same goes for most other enchantments that are strong against focus-fire: Shield of Deflection is excellent to protect someone getting hit by lots of warriors and/or rangers, but if the other team immediately switches targets, the SoD does nothing, and even though it's a spammable enchantment, after 2-3 rapid casts to the different targets the opponents are attacking, the enchanter will be out of energy, letting you to deal your damage without having to deal with the enchantments.

Overall, more than just the amount of damage you can output, you also have to consider the utility of that damage- how fast you can change targets, how much you can lessen the ability of the enemy healers to keep their team alive, as well as being able to deal enough damage to break through the healing in the first place. AoE can be helpful in this respect as utility damage, more than just damage over time- if you nail 3 targets with a high-damage AoE spell, and follow up by focusing those targets, switching targets when big enchantments are cast on them, you've done a good job of increasing the burden placed on the enemy healers. If you're using AoE to deal 100 damage to enemy warriors, and your team is focusing all their attacks on the enemy healers or spellcasters, the AoE isn't too worthwhile, as the healer can ignore it in favor of healing the people taking sustained damage. Similarly, Elementalists have the advantage of taking off the last amount of health through enchantments. Healing Seed will do little if you hit the target with a Lightning Orb, and Shield of Deflection is useless against high-damage spells. In this respect, an elementalist can be more effective taking off that last little bit of health than a warrior or ranger, as they can bypass most healing enchantments which can stop the warriors and rangers from killing their target.

[Keramon]

One of the most effective benefits of AOE is the effect of confusion. Casting it into the center of a group often results in the said group being split up and running in different directions and concentrating on getting away from the AOE rather than attacking for a few seconds. That is not something that is measurable however, it should still be mentioned.

[Xapti]

Ensign's post, #50, points out some good information regarding AOE.
A very pertinent addition explaning more the advantages of AOE is in Freyas' post, #56.
I tend to agree more that AoE is more beneficial than it might seem DPS wise due to those arguments (in post #56), even though they still are considerably less powerful than focus fire in some circumstances.

perhaps summary of those two posts?:
Focus fire can be very useful if it can kill the target before it gets many buffs on it, but once certain enchants are enabled, ethier begin focus fire on another target, and/or start use AoE damage.

edit: and about using an energy gain weapon...
For Standard primary warrior, regen would be 2 pips, 2/3 of a point every second, 40 points per minute. Assuming the same warrior is using an energy gain sword, and attacking 100% of the time, he would loose 1 pip, and gain .75 per second. This results in a total gain of +.25 EP per second maximum, resulting in maximum sustainable energy consumption being 65 EP/minute

[Ellestar]

http://www.guildwarsguru.com/forum/showpost.php?p=1450&postcount=50
Ok, so every five seconds, the enemy does X stuff to you. Now you have a choice. First, you can choose to focus fire them - killing one, then the other. If you do that, the two enemies will do 2X worth of stuff to you, then one drops. Then the remaining enemy will do another X while you finish him off. So overall, focusing fire on one of two enemies allows them to do 3X worth of stuff to you.

Or, you could choose to divide your damage, one hit to one, then one hit to the other. If you do things this way, both enemies will deal X every five seconds until they both die simultaneously at the ten second mark. If you do things this way, they'll get to do 4X back to you.

So dividing your damage up between targets is only 3/4 as effective as focus fire, since the time to kill is identical and the amount of stuff they can do back to you is all that changes, really.

Now, what if you were hitting both of them simultaneously? If that were the case, you'd be able to drop both in five seconds, and they'd only be able to deal 2X damage back to you. Area effect damage is 3/2 times as effective as focused fire on a single target.

You can repeat this process for more targets, and in doing so you get a simple rule - every additional target beyond the first is worth 1/2 as much as hitting a single target.
Can't agree with it.

There is a so-called "Basic Effectiveness Equation" and derivations
Power = HP * DPS
He who has a bigger power wins.

Say, a power of two units is equal. Then Power_1 = HP_1 * DPS_1 = HP_2 * DPS_2 = Power_2 or HP_1 / DPS_2 = HP_2 / DPS_1
Defender HP / Attacker DPS = time to kill defender, so if power is equal then units kill each other simultaneously. It's not exactly true if you'll calculate Power of several units, but it's close enough.
Also from that formula, if 1st unit is Power_Ratio = Power_1 / Power_2 more powerful than a 2nd unit, then it will lose (1 / Power_Ratio) percent of his life when 2nd unit will die.

If you have N enemies and you kill them one after another (100% focus fire), then they have N * HP_1 hit points and have DPS = DPS_1 * (N,(N-1),(N-2),..,2,1) = DPS_1 * N * (N + 1) / 2 or average DPS = DPS_1 * N * (N + 1) / 2 / N = DPS_1 * (N + 1) / 2
Power = N * HP_1 * DPS_1 * (N + 1) / 2 = HP_1 * DPS_1 * N * (N + 1) /2

If you have N enemies and you don't focus fire at all (enemies die simultaneously) then HP = HP_1 * N and DPS = DPS_1 * N
Power = HP_1 * DPS_1 * N^2
(N^2 = N * N, just in case)

So, dividing your damage between your targets is (N * (N + 1) / 2) / (N^2) = (N+1) / (2 * N) less efficient. It's equal to 3/4 only when N = 2

If you're hitting all targets simultaneously, then your enemies effectively have HP equal to HP_1 and DPS = N * DPS_1 or N times as powerful as one enemy. So, it's (N * (N + 1) / 2) / N = ((N + 1) / 2) times more efficient than killing enemies one at a time by a focus fire.
If you divide your damage between all targets, then obviously you just kill all targets simultaneously. So i don't understand how you come to a conclusion that every additional target beyond the first is worth 1/2 as much as hitting a single target.

Now, if we'll suppose that we have 8 targets total and we're attacking M targets simultaneously, then we can think about it like enemies have (8 / M)units with HP = HP_1 * M and average DPS = DPS_1 * M. So, Power = (M^2) *((8 / M) * ((8 / M) + 1)) / 2
Power of 8 units if we focus fire is 8*9/2=36
Of course, it makes sense only when 8 / N is a whole number.
So, Power_Ratio is
1: 1 * 36 / 36 = 1
2: 36 / (4 * (4 * 5 / 2)) = 0.9
4: 36 / (16 * (2 * 3 / 2)) = 0.75
So, that's what we lose in effectiveness when we hit several targets with AoE spell instead of focusing same damage on one target out of 8 avaliable. If we have some non-AoE damage, then numbers will be close only when our AoE damages secondary targets that will be attacked by focus fire next.

****************

If I recall correctly using an adrenaline skills drains one adrenaline point from all other skills so you just can't spam 8 x Galrath, and 6 x Final in 59 seconds (60s minus conjure).
Maybe someone will confirm or deny this?


http://www.guildwarsguru.com/forum/showpost.php?p=1221&postcount=27
There's another concern that melee damage has - packing density. Unlike Rangers or Elementalists with ranged attacks, Warriors and PBAoE Elementalists have to be next to a target to deal damage to them, and you aren't going to get more than 3 guys around a single target with any consistency. Even if you found an absurd Warrior DPS combo, you can't put more than 3 or so of them in a group without them stepping on each other. In that way Warrior damage is a bit more 'fair', in that a broken Warrior combo won't spawn 6 Warrior teams - but it is an additional constraint.
What if they surround target first? It works fine in a lot of other games so probably it should be good in GW too. Of course, it means that at least some warriors should move on a manual control instead of a simple T-Space and with a speed buff.

http://www.guildwarsguru.com/forum/showpost.php?p=1226&postcount=28
That's the real lesson here. Stack buffs and other damage adds and maximize your DPS. That's the real way to kill things. Not lobbing Fireballs and Fire Storms at them.
Probably warrior train has a good DPS, but it's easy to debuff it too. Rangers are better in this regard, but they recieve less benefit to their peak DPS from stacking buffs.


http://www.guildwarsguru.com/forum/showpost.php?p=1450&postcount=50
This question really gets to the heart of what health and damage is. Your goal when you attack someone isn't to deal damage to them - it is to kill them. The difference is subtle. If I have 480 health, I really don't care if I lose the first 479 health, it's that last point that I care about. The rest of the points are just buffer, health that defends against your attacks until you land a killing blow. If I can kill, shut down, or otherwise defeat you before you take that last point, it really doesn't matter how much damage you dealt me, because I won. In the same way, dealing damage to your opponent isn't a goal unto itself - it's merely a mechanism for setting your opponent up for a killing blow.
I like my term for this "HP shield" - IMHO it has very strong meaning that your HP is effectively your shield and it's different from a common terms such as "damage shield" etc. that can be already used in a game mechanics.
Also, sometimes it's useful to think that a target you plan to kill don't have that HP Shield (equal to (HP - 1)) while all other targets have this shield. Probably 2nd character in a queue also don't have that HP shield. Same idea, but more widely used.

****************

There are 3 stages of killing a target with a focus fire

1) Preparation damage
2) Burst damage
3) Breakthrough

(1) Preparation damage is the starting damage what you already have on a target when you start to synchronize damage.
(2) Burst damage is obvious - any damage you deal on a target before enemy defence starts to defend this target (that may include reaction to focus fire damage or to damage at all).
(3) Breakthrough is when you're trying to deal more damage than enemy defence can soak in and remove what's left off target health.

(1) Preparation damage is tricky. If enemy decides to heal, then you don't have an advantage of Preparation damage and/or an advantage of Burst damage (Burst damage actually becomes Breakthrough in that case). Generally noone heals damage that is somewhat less than lowest possible instant heal. Also, it's hard to choose a right target to heal when a lot of targets are damaged, but on the other side generally priority targets are a priority for later damage and for that reason they are a priority for healing too. AoEs are probably the most efficient here. Health loss because of the Death penalty can also be counted as a Preparation damage.

(2) Burst damage is more efficient when you can do it unexpected (so generally warriors or PBAoE spells aren't that useful). In other case enemy at the very least will be able to heal preparation damage. The most important thing is to deal a lots of raw damage in this phase. High-damage long casting time attacks may work better because it's impossible to heal a target before it actually recieves damage.

(3) In a Breakthrough phase, the most important things are DPS and disables. Even if enemy has 1 health, attack will be unsuccessful if you can't outpace enemy healing and defence.
Generally it's better to save limited power you have for this final assault. If you use a limited power in a burst damage phase and then lose the arms race in a Breakthrough phase, then you lose everything you have - energy, skills with long recharge times etc. Also, the more damage target has after a Burst damage phase, the bigger immediate responce you'll get from enemy defence that further reduces your chances for success. If you didn't used your most precious skills or a significant amount of energy, then you can still win a resource race - either you cancel a Breakthrough early enough after enemy spends his resources to defend. Or you drain some resources first that enemy will use after a Burst damage phase and some more resources it will use to defend against an increased pressure during a Breakthrough. So, effectively, if you can consecutively and significantly increase your pressure on enemy, and enemy tries to save his most efficient defence resources, then it works just like several additional Burst damage phases.
Obviously, Interrupting/removing/disabling enemy defence in this phase is almost as good as dealing respective damge while it don't have the same effect in other phases.

[Blackace]

Interesting part about the damage phases here. I've thought about it(mostly with necro hexes and when to apply them) but thats something that we havent really covered at Gurus yet. I never actually thought of names for the phases, but one day I'll have to make a topic about it.

[torry]

Charles, it would be valuable to see the DPS of base adrenaline attacks in your table.

Why? Because they stack independantly and with everything else. So 3 adrenaline attacks could add 3+3+3 = 10 DPS.

While Energy attacks do not stack, as they draw from the same energy pool. Skill A will mean less of Skill B, or no Skill C.

[Ensign]

Great post Ellestar - our models and mathematics are virtually identical, though yours is more rigorous. Thanks for posting it.


So i don't understand how you come to a conclusion that every additional target beyond the first is worth 1/2 as much as hitting a single target.

The answer to this is sitting a paragraph earlier, actually:


So, it's (N * (N + 1) / 2) / N = ((N + 1) / 2) times more efficient than killing enemies one at a time by a focus fire.

If you're hitting N enemies simultaneously with an attack, that's (N + 1) / 2 times as efficient as just hitting one enemy - which should look familiar, as that's the relationship I referred to. It equals 1 for a single target (normal damage), 1.5 for two targets, 2 for three targets, 2.5 for four targets, etc. Hence, +50% DPS per additional target.


If you divide your damage between all targets, then obviously you just kill all targets simultaneously.

The difference is that you're looking at it from the perspective of dividing your damage - and, indeed, dividing your damage cuts into effectiveness. But the original basis wasn't one of dividing your damage, but of 'how much extra effectiveness you gain by catching people in an AoE.'

Basically, if an AoE skill deals 20 DPS to a single target, it'll deal (effectively) 30 DPS if you can catch two people in the AoE, 40 DPS if you can catch three people, etc. This matches the relationship and basis as given.

Now, if instead we compared it to your basis of dividing damage, catching two people in an AoE isn't 1.5x as efficient as hitting a single target, it's .75x as efficient as though that damage were all concentrated onto a single target. But if the damage were concentrated you'd have 40 DPS, and the penalty for division gives you 30 DPS. Similarly, for three targets you'd have 60 DPS raw, but the penalty for division is (N + 1) / 2N = 2/3, so you effectively only get 40 DPS - again, the same result.

It's just a difference of basis. Our mathematics are identical - thank you again for spelling out the rigorous details.


Maybe someone will confirm or deny this?

Confirmed - stacking a whole bunch of adrenal skills makes the chains a bit messier than simply adding up all of your DPS numbers.


What if they surround target first?

It might be possible to get eight Warriors attacking a single target. It might not be possible - I've never tried. I do know that after around three Warriors collision detection starts to be a big deal, and you can't just pile Warriors onto a target without very careful packing. In practice this means that after three Warriors you're starting to cut into your effectiveness because you're spending a lot more time trying to make room than actually dealing damage - hence it becomes impractical and not recommended.


Generally it's better to save limited power you have for this final assault.

I couldn't disagree with this more. Get your best damage in early when the target is still soft. Once the opponent has had a chance to react you're going to see armor buffs, deflection bonuses, and a bunch of other defenses that will cut into your assault. If you couldn't deal enough damage to kill someone before the defenses come up, you aren't going to deal anywhere near enough once they drop a Sheild of Deflection on the target.

So get as much in early as you can, then concentrate on punching holes in their defense so you can finish them off. In practice this means damage that's slower but ignores defenses, because punching through defenses (aka, enchantment removal) is impractical.

Peace,
-CxE

[Rellok]

Confirmed - stacking a whole bunch of adrenal skills makes the chains a bit messier than simply adding up all of your DPS numbers.



I believe that any adrenal skill drains 1 adren from each adren pool, however, if the skill successfully hits, you gain one everywhere anyway. So, basically if you are trying to put together adren trains, you need to get the highest adren you will need without adren attacks (unless they hit multiple targets).

Great info guys,

Matt

[Ellestar]

If you're hitting N enemies simultaneously with an attack, that's (N + 1) / 2 times as efficient as just hitting one enemy - which should look familiar, as that's the relationship I referred to. It equals 1 for a single target (normal damage), 1.5 for two targets, 2 for three targets, 2.5 for four targets, etc. Hence, +50% DPS per additional target.
First, it's true only when there are N enemies total. In GW, you have 8 enemies total. So it's wrong. Second, there is a difference between dealing all damage only with that kind of AoE attack and killing N enemies or just hitting N enemies with that attack for some percent of their life.

I couldn't disagree with this more. Get your best damage in early when the target is still soft. Once the opponent has had a chance to react you're going to see armor buffs, deflection bonuses, and a bunch of other defenses that will cut into your assault. If you couldn't deal enough damage to kill someone before the defenses come up, you aren't going to deal anywhere near enough once they drop a Sheild of Deflection on the target.

So get as much in early as you can, then concentrate on punching holes in their defense so you can finish them off. In practice this means damage that's slower but ignores defenses, because punching through defenses (aka, enchantment removal) is impractical.
So, if your primary damage-dealers are warriors, then you save your limited resources to finish your enemy, but you don't use it at the beginning - say, in your example it's the damage that ignores defence and energy you need to use it.


http://www.guildwarsguru.com/forum/showpost.php?p=1302&postcount=35
As mentioned multiple times in this thread, what you're trying to find here, is a good way to compare dmg dealing capabilities in a simple way. Introducing side effects such as equipment will lead to a pandemonium of arguments. Sure, there are broken mods for warriors but why not taking about account other mods as well.

Basically I'm in favor of pure dmg dealing comparison (i.e: pure killing machine potential) without any outside interference. No external buff, no equipment. When this case of study has been beaten to death, we can talk about equipment or group dmg potential.
That way you'll get a meaningless result because in practice things will be different. Metagame is always important in Min/Maxing.

One of the most effective benefits of AOE is the effect of confusion. Casting it into the center of a group often results in the said group being split up and running in different directions and concentrating on getting away from the AOE rather than attacking for a few seconds. That is not something that is measurable however, it should still be mentioned.
Experience teams will not be confused under normal conditons, but AoE DoTs still may be very useful if used right. For example, you cast a fire storm on a monk. It's not a good idea to stay under a fire storm for a full duration so monk will try to move out of this area. Of course, it's easy to do it under normal conditions. But when someone is under a heavy focus fire, monk may not have a free time to move out of a fire storm area immediately. If he's snared, then it's even more efficient because he either spends more time moving or recieves full damage. Also, it reduces strain on a friendly monks if you catch some enemy ranged damage-dealers in an AoE because they will spend time moving too.

[Blackace]

Too bad that the fire line sucks horribly. Chaos storm>Fire magic.

[Lamaros]

The problem with focus fire is it's easy to counter. The problem with AOE is that you wont kill anyone that way. So do a bit of both:

AOE to 'soften' multiple targets.

Then you hit burst damage on a likely target.

As the advantage of AOE, in damaging multiple targets, is that it allows you to set up many options of what to kill without the enemy knowing where they should concentrate defence. If you pick one they dont heal straight away (and they are unable to heal them all) with the burst damage (we know Galrath+Final can do near 300 damage to the proper undefended targets), before they can get the heavy defences up then you'll kill someone.

Simple philosophy:

Focused Healing > Focused Attacking
Group Healing > AOE attacking.

The aim is to make them group heal when you're focusing, and focus heal when you're group attacking.

But even the above is probably too specific an example...

"Metagame is always important in Min/Maxing"
"Attack is the best defence"
"<insert random Sun Tzu quote here>"

Winning isn't about Burst or AOE damage, it's about doing something that your enemy has no counter for. Knowing what your enemy is going to do, or forcing them to act in a specific way, provides you with this opportunity.

So much better to look at Burst DPS and AOE DPS as different tools at your disposal, rather then in direct comparison.

What is useful is looking at what does the best Burst DPS, Best AOE DPS, and how. This allows you to know how to deploy it most sucessfuly, how it can be countered.. etc etc.