Guild Wars 2

I posted this in the game bugs forum, but since it’s technically more of a design oversight than a bug, and it’s one that most people seem to not notice, I wanted to post it here in hopes to generate conversation and awareness:

It’s very clear that skills were balanced in part around their activation speeds. However, in cases where a skill has no cooldown and is repeated (such as the #1 skills), the activation speed is superseded by the length of the animation. This doesn’t really affect potential DPS where the activation speed + cooldown is longer than the length of the animation, but otherwise it does. It is especially a problem with the #1 skills on some weapons, where the animation length is significantly longer than the activation speed. I think some people notice this and assume it is intentional, but doing a little comparative analysis it actually appears more likely to be a design oversight.

As an example, the Thief’s Vital Shot has a 1/2 second activation speed and a 4 second bleed. The Warrior’s Bleeding Shot has a 3/4 second activation speed and a 6 second bleed. It’s easy to infer that Bleeding Shot was meant to be 50% stronger while Vital Shot was 50% faster. However, because the length of the firing animation is similar for both, the reality is that Bleeding Shot is 50% stronger, while Vital Shot is only about 10% faster. This is because the animation for the Rifle is only slightly longer than its activation speed, while the animation for the Pistol is significantly longer than its activation speed.

This would be the major reason (not Body Shot, though it’s kind of weak too) why most people complain about Pistols. This also severely affects Longbows (both Long Range Shot and Dual Shot are exceptionally weak autoattacks), and the Elementalist’s staff. On the latter, Fire is stronger than the other 3 elements, but the other 3 elements are supposed to be faster. They in fact all have the same recast because they use the same animation, so therefore Fire’s autoattack is strictly better than the other 3 elements, even though it’s still weaker than it should be due to the length of the animation.

It’s a major problem causing issues with most professions and making certain weapon sets much less usable than they should be.

i was expecting mesmer’s scepter to be mentioned here
the auto attack is a chain and goes faster the closer you are to your target

lupercaleb.6041:

i was expecting mesmer’s scepter to be mentioned here
the auto attack is a chain and goes faster the closer you are to your target

I didn’t mention Mesmer’s Scepter only because I haven’t specifically analyzed it in any depth, whereas I have the ones I mentioned above. And it is possible that the range and projectile speed has more to do with it than whether or not it is a chain.

At any rate, there is no question that something is off with a lot of these #1 skills and it really needs to be looked at.

i dont use auto attack so i wouldnt know. from what i understand the game wasnt going to have an auto attck but there early test showed that some people are not to bright do they added the abilty to set skills to auto attack.

i feel its to clunky in this game and when useing any skill that dose malie or cone dammage its not a good idea to use it because as soon as the mob your targiting moves out of range you stop attacking.

but even though the mob your targiting moves you can still hit any mobs in your cone

UnderdogSMO.9428:

i dont use auto attack so i wouldnt know. from what i understand the game wasnt going to have an auto attck but there early test showed that some people are not to bright do they added the abilty to set skills to auto attack.

i feel its to clunky in this game and when useing any skill that dose malie or cone dammage its not a good idea to use it because as soon as the mob your targiting moves out of range you stop attacking.

but even though the mob your targiting moves you can still hit any mobs in your cone

I’m pretty sure that isn’t the case, and there’s nothing wrong with having an autoattack. It has nothing to do with being ‘bright’.

Einlanzer.1627:

UnderdogSMO.9428:

i dont use auto attack so i wouldnt know. from what i understand the game wasnt going to have an auto attck but there early test showed that some people are not to bright do they added the abilty to set skills to auto attack.

i feel its to clunky in this game and when useing any skill that dose malie or cone dammage its not a good idea to use it because as soon as the mob your targiting moves out of range you stop attacking.

but even though the mob your targiting moves you can still hit any mobs in your cone

I’m pretty sure that isn’t the case, and there’s nothing wrong with having an autoattack. It has nothing to do with being ‘bright’.

it dose when your not smart enough to figger out there isnt one. this was when the game still had mana and your 1 skill would actualy use some of iit from what i remamber.

Hm… so you’re saying that the devs took it for granted? I’m used to CoH (rest in peace) where the animation time was the activation time, and the biggest obstacle to calculating DPS was the baked in game refresh rate to processing actions.

But if the attack delay on skills is applied prior to animation time, and animation time is just tacked on, then this does lead to a very big problem in balancing. This throws the effectiveness of many classes into the air, and also explains why scepter mesmers are still quite horrible.

I did notice something while playing a thief. In PVE I like to use s/d and d/d dual skills to evade enemy attacks and fight at the same time, so I spend half the fight auto attacking and the other half using the dual skill to evade. This works… only about half the time. The other half of the time, the delay between button press and move execution is so wide that the dodges in those skills become unreliable. Death Blossom has a quarter second dodge, and there are times when I press the button and the enemy attack lands long before the attack even executes. Then, on the next use, I’ll press the button and the attack will execute so quickly that the dodge period occur before the enemy attack even lands.

Ideally, there should be 3 factors that are accounted for when balancing skills:

#1: Delay in processing input due to the game’s refresh rate. These are usually a fraction of a second, but in long term it adds up, and also it factors in more with quickly used skills.

#2: Delay in skill activation baked into the skill.

#3: Animation time.

You miss any one of them, and it leaves a glaring hole in game design that makes balancing impossible.

Blood Red Arachnid.2493:

Hm… so you’re saying that the devs took it for granted? I’m used to CoH (rest in peace) where the animation time was the activation time, and the biggest obstacle to calculating DPS was the baked in game refresh rate to processing actions.

But if the attack delay on skills is applied prior to animation time, and animation time is just tacked on, then this does lead to a very big problem in balancing. This throws the effectiveness of many classes into the air, and also explains why scepter mesmers are still quite horrible.

I did notice something while playing a thief. In PVE I like to use s/d and d/d dual skills to evade enemy attacks and fight at the same time, so I spend half the fight auto attacking and the other half using the dual skill to evade. This works… only about half the time. The other half of the time, the delay between button press and move execution is so wide that the dodges in those skills become unreliable. Death Blossom has a quarter second dodge, and there are times when I press the button and the enemy attack lands long before the attack even executes. Then, on the next use, I’ll press the button and the attack will execute so quickly that the dodge period occur before the enemy attack even lands.

Ideally, there should be 3 factors that are accounted for when balancing skills:

#1: Delay in processing input due to the game’s refresh rate. These are usually a fraction of a second, but in long term it adds up, and also it factors in more with quickly used skills.

#2: Delay in skill activation baked into the skill.

#3: Animation time.

You miss any one of them, and it leaves a glaring hole in game design that makes balancing impossible.

Yep, pretty much. While I’ll admit there’s a slight chance I’m wrong, I’m pretty sure I’m not. It seems pretty obvious when you start looking at skills across professions that something is off and that there was a failure at balancing the skills around total cast time rather than activation speed.

If you play a Warrior or Ranger, have you ever noticed how pitifully weak the #1 skill on the Longbow is? If you look at the activation speed, you’ll see it’s 3/4 seconds. However, if you time it you’ll see the rate of fire is about 1 shot per 1.25 seconds. That’s a .5 second delay between activations that isn’t taken into account by any metric. The warrior’s rifle also has a 3/4 second activation speed, but the animation is much quicker, so the rate of fire is about 1 shot per .85 seconds, meaning there’s only a .1 second delay between activations. Unsurprisingly, the #1 on the Rifle packs a much bigger punch than the #1 on the Longbow. This affects the DPS of the whole set, so you can pretty objectively state that the Rifle is a superior weapon to the Longbow in terms of sustained DPS. The Rifle is also superior to the Pistol, the Staff, and likely the Scepter also.

I really don’t think it’s intentional, I feel quite confident that it was an oversight. One that is causing serious problems.

Maybe the devs are actually balancing on damage per second rather than time per event. We don’t really know how they’re balancing.

It may be difficult for us to figure out the best damage, but I’m pretty sure the devs would use specific tools that would mitigate the need for them to realize this. I mean it makes sense.

I suspect the devs have a bevy of tools to measure stuff that aren’t available to us.

But it’s an interesting thing to consider from a player vantage.

Vayne.8563:

Maybe the devs are actually balancing on damage per second rather than time per event. We don’t really know how they’re balancing.

It may be difficult for us to figure out the best damage, but I’m pretty sure the devs would use specific tools that would mitigate the need for them to realize this. I mean it makes sense.

I suspect the devs have a bevy of tools to measure stuff that aren’t available to us.

But it’s an interesting thing to consider from a player vantage.

It would be easier to believe that if the discrepancies weren’t so obvious. Go attack something only using autoattack with all weapons across professions, and you’ll immediately see inconsistency in damage potential. You could make the argument that the inconsistency was intentional, however, there’s another piece of data you can use to infer otherwise:

The longer the difference between the activation speed of the skill and the animation speed of the skill, the weaker the skill is.

The Rifle for both Engis and Warriors is a good weapon, and the autoattack is strong, because there is only a .1 second difference between the activation speed and the total animation time. It does comparable damage to most melee weapons and works well as a standalone set.

The Pistol for both Thieves and Engis is a bit weaker. The Thief can’t stack as many bleeds nor do as much direct damage while also having less range compared to the Rifle Warrior. The activation speed is 1/2 second and total animation is about .8 seconds, so there is about a .3 second difference. The Pistol should in fact be slightly stronger than the rifle due to the reduced range, but it’s actually weaker. Theorycrafting has determined this conclusively.

The Longbow for both the Ranger and Warrior is very obviously weaker than it should be, as everyone complains loudly about. The #1 skill on both does pitiful damage so the weapon has to be used tactically only in niche situations and is otherwise very poor. The activation speed is a 3/4 second, like the Rifle, but the total animation is about 1.25 seconds, so that means the difference is about .5 seconds.

The Staff for the Elementalist has different speeds for the various elements. Fire has 1 second while Earth, Air, and Water all have 3/4 second. However, the animation is the same for all of them so they all have the same recast speed, about 1.25 seconds. That’s a .25 second delay for Fire and a .5 second delay for the others. This means they are all 4 weaker than they should be, but Water, Air and Earth are dramatically weaker while Fire comes closer to the range of usability. This is one of the reasons you almost always see Staff Eles using Fire outside of situations where they are needed for heavy support.

Vayne.8563:

Maybe the devs are actually balancing on damage per second rather than time per event. We don’t really know how they’re balancing.

It may be difficult for us to figure out the best damage, but I’m pretty sure the devs would use specific tools that would mitigate the need for them to realize this. I mean it makes sense.

I suspect the devs have a bevy of tools to measure stuff that aren’t available to us.

But it’s an interesting thing to consider from a player vantage.

I would not assume the devs have some mysterious device that allows them to compensate for flaws in methodology. I would assume the opposite.

Because “best damage” is a really simple formula. It’s damage/ time, AKA DPS. The damage is fairly easy to adjust, but the time requires the 3 things I mentioned above to be taken into account. And to get those, there really isn’t anything much better than the average stop watch.

Blood Red Arachnid.2493:

Vayne.8563:

Maybe the devs are actually balancing on damage per second rather than time per event. We don’t really know how they’re balancing.

It may be difficult for us to figure out the best damage, but I’m pretty sure the devs would use specific tools that would mitigate the need for them to realize this. I mean it makes sense.

I suspect the devs have a bevy of tools to measure stuff that aren’t available to us.

But it’s an interesting thing to consider from a player vantage.

I would not assume the devs have some mysterious device that allows them to compensate for flaws in methodology. I would assume the opposite.

Because “best damage” is a really simple formula. It’s damage/ time, AKA DPS. The damage is fairly easy to adjust, but the time requires the 3 things I mentioned above to be taken into account. And to get those, there really isn’t anything much better than the average stop watch.

Especially considering that the only ‘time’ based metric, Activation speed, is entirely superseded by the length of the animation, which is not measured anywhere in game. I repeat, it is in fact very likely that this is a design oversight and not intentional. Assuming that is the case, it needs to be addressed immediately.

Meh, chalk it up to design esthetics overruling game mechanics.

Honestly i wonder if these hidden delays are why i have such issues dodging various stuff. It is like the actual evasion kicks in a half second or so after the actual animation starts…

digiowl.9620:

Meh, chalk it up to design esthetics overruling game mechanics.

Honestly i wonder if these hidden delays are why i have such issues dodging various stuff. It is like the actual evasion kicks in a half second or so after the actual animation starts…

Except it’s very likely to be an oversight on Anet’s part, and it’s a very serious one. This is the major culprit behind why the Longbow feels really off and why people complain very loudly about it in the forums. Even the devs have admitted that it needs work.

I did some testing similar to this a couple of days ago for the guardian, the seconds aren’t perfectly accurate but are close enough to be taken into consideration.
Scepter: 157 animations over 117.99 seconds
Sword: 90 animations over 76.09 seconds
Mace: 123 animations over 134.43 seconds
Hammer: 69 animations over 82.89 seconds
Great Sword: 122 animations over 101.77 seconds

Given that the scepter has a “cast” time of 0.25 seconds and the fact that I was in melee range (negating the slow travel time) I should have achieved ~472 scepter animations, in reality I was achieving 1.33 animations every second which is quite obviously different than the claimed 4 animations a second that the “cast” time would indicate.

Einlanzer.1627:

digiowl.9620:

Meh, chalk it up to design esthetics overruling game mechanics.

Honestly i wonder if these hidden delays are why i have such issues dodging various stuff. It is like the actual evasion kicks in a half second or so after the actual animation starts…

Except it’s very likely to be an oversight on Anet’s part, and it’s a very serious one. This is the major culprit behind why the Longbow feels really off and why people complain very loudly about it in the forums. Even the devs have admitted that it needs work.

I wonder where i read it, but i swear ANet claimed their combat system was to be animation based or something.

Hmm. An interesting read to be sure, and an interesting insight into why some of these attacks seem to be lacking. No way to know for sure if its accurate without a dev weighing in, but the evidence does seem to support it.

Well, rifle is two handed, so in a sense it should be stronger. CBF reading the rest, my eyes are dead and ima go bed xD

In a game where the ‘challenge’ seems to revolve around giving bosses unreasonable amounts of health, I’d say a working auto-attack was essential for most builds…

I’d like a dev to see this, because you make a great case. After all, WHY would the different ele staff auto-attacks be given different cast times if it makes absolutely no functional difference? As you have answered, the most obvious explanation is: because attacks were initially balanced around cast time, but animation time unintentionally superseded cast time entirely post-balance.

I’ve always felt ele staff autos were unjustifiably slow, and now I understand why. It’s terrible to think that an animator arbitrarily deciding to make certain animations more elaborate can have such negative and far reaching impacts on game balance. Not surprisingly, it is the easily overemphasized motions like casting a spell with a staff or scepter that have caused the biggest problems.

Would be interesting to measure the difference between races doing the same exact attack.

digiowl.9620:

Would be interesting to measure the difference between races doing the same exact attack.

I’ve thought about this as well, unfortunately I don’t have space in my bar to make duplicate classes for two different races and see if there’s a difference. It would be a great test, though. Maybe I will delete one of my underused toons and give it a whirl.

I also feel like we should really get this rolling on the game bug forums to try to bring more dev attention to it. If you have read this and agree with my assessment, please take a second to respond to the similar post I have in the Game Bugs forum.

Very interesting thread. I look forward to the developments. Is there a disparity? Or can the dev’s justify it somehow?!

Agree. It drives me absolutely nuts. My suggestion would be, if I use a skill that does not require a target (for example, a shout), my character should not start autoattacking the nearest thing…

IndigoSundown.5419:

digiowl.9620:

Honestly i wonder if these hidden delays are why i have such issues dodging various stuff. It is like the actual evasion kicks in a half second or so after the actual animation starts…

Yeah, many’s the time I was caught half in and half out of Subject Alpha’s crystal of imprisonment — obviously right in the middle of the dive portion of the dodge animation. I think the evade procs on hitting the ground.

Something that is a real problem if you try to avoid a virtually instant attack, like those recent sonic periscopes. Dodging the moment i hear the sound usually just see me starting the dodge and then go flying as the shockwave hits.

And much the same seems to be the case with block actions. I wonder how many times i have tried to use the various damaging blocks on a mesmer and seen the mob land the blow and then my toon raising his weapon into position, even tho i hit the button the moment the mob attack was noticeable.

These kinds of actions needs to be instant, or they are pointless and you further fuel the DPS race.

Chuo.4238:

Agree. It drives me absolutely nuts. My suggestion would be, if I use a skill that does not require a target (for example, a shout), my character should not start autoattacking the nearest thing…

Noticed something similar when using stimulants from the engineer med kit to traverse distances. I find myself targeting fauna left and right, and if it was not a kit i would likely attract a whole lot of needless attention.

Guys, this is about a very specific issue that is broadly affecting weapons and professions in very serious ways- many autoattacks being much slower than they were intended to be and egregiously gimping the weapon. I’m on a crusade to bring as much attention to this issue as possible, so I’d appreciate being able to stay on topic.

Don’t you think this is intended? I.e. the cast time gives the delay of the skill effect after pushing the button while the total execution time of the skill (until you can cast it again) is determined by the animation?
I don’t think it would be intended to cast a 0.25s skill 4 times per second. I think it’s merely there so that the skill feels nearly instant opposed to a long cast time skill which you have to time carefully.
Not an expert in these things, but that’s what I always thought.

If the action was instantaneous(block, evade, attack, etc) and the animation lagged a bit, it wouldn’t be so bad. A visual incongruity can be tuned by the devs without affecting gameplay.
Someone has an advantage when you are encountering latency for skill activation waiting on the animation.

Tintifax.2614:

Don’t you think this is intended? I.e. the cast time gives the delay of the skill effect after pushing the button while the total execution time of the skill (until you can cast it again) is determined by the animation?
I don’t think it would be intended to cast a 0.25s skill 4 times per second. I think it’s merely there so that the skill feels nearly instant opposed to a long cast time skill which you have to time carefully.
Not an expert in these things, but that’s what I always thought.

No, I don’t, mostly because the evidence available doesn’t support that it’s intended. What I can’t seem to be able to drive home to some people is this:

The total animation speed only matters for skills set to autoattack, it does not matter otherwise. So balancing a skill on its activation speed is adequate for all skills but the #1 skill. Since the #1 skills have activation speeds, it’s very likely they were all balanced in the same way initially.

However, whenever a skill is set to autoattack (usually just the #1 skills), whatever its activation speed is is absorbed into the total time of the animation, making Activation speed a non-functioning metric on those skills. And the length of the animation relative to the activation speed varies hugely between weapons, making the delay between the two erratic from weapon to weapon. Unsurprisingly, weapons that have a very large difference between their activation speeds and their total animation times, such as the Longbow, have pitifully weak autoattacks while weapons that have a short delay, such as the Rifle, have fairly strong autattacks.

Maybe it’s intentional, but it kitten sure doesn’t look like it is and I think some explanation is warranted.

You don’t have to “drive something home” with me, I’m just offering a differing perspective and don’t care if I’m wrong :-) Of course a dev clarifying things would be nice…

My gist on it:
The main balancing is done with the cooldown, the activation time ist just another dimension of balancing/altering the feel of a skill. Now while there is no cooldown listed for the #1 skills they still have a base cooldown, namely their animation time. All other skills have a cooldown longer than their animation time. The fact that it’s not listed as a cooldown with the #1 skills is where the misunderstanding comes from.
Of course if you set the #1 skill to autoattack the activation time is irrelevant, but the skills are still balanced via the animation time. However it is possible, that this balancing could (should) be improved…
That activation time is irrelevant with autoattacking #1 skills could very well not mean anything. If I remember right, autoattack was added relatively late in development, when the rest of the skill system was already done.

So in essence, you may see a higher DPS by hammering #1 manually vs leaving to in the hands of the autoattack mechanic.

No, digiowl, I’ve just tested it. You can’t cast faster manually than using autoattack. What Einlanzer means is that you don’t notice the skill activation delay when autoattacking. You don’t care that the skill activates half a second after the computer presses the button for you, as long as the skill fires, say, once a second. Although you notice it when casting manually.

Tintifax.2614:

You don’t have to “drive something home” with me, I’m just offering a differing perspective and don’t care if I’m wrong :-) Of course a dev clarifying things would be nice…

My gist on it:
The main balancing is done with the cooldown, the activation time ist just another dimension of balancing/altering the feel of a skill. Now while there is no cooldown listed for the #1 skills they still have a base cooldown, namely their animation time. All other skills have a cooldown longer than their animation time. The fact that it’s not listed as a cooldown with the #1 skills is where the misunderstanding comes from.
Of course if you set the #1 skill to autoattack the activation time is irrelevant, but the skills are still balanced via the animation time. However it is possible, that this balancing could (should) be improved…
That activation time is irrelevant with autoattacking #1 skills could very well not mean anything. If I remember right, autoattack was added relatively late in development, when the rest of the skill system was already done.

If what you say is true, it would be very likely that we would expect to see similar output from all autoattacks, when in fact we don’t. As I also outlined above, we can predict how weak an autoattack is relative to other autoattacks by looking at the difference between its activation speed and its total recast time due to the animation.

We can, for example, predict that the Longbow’s autoattack is very weak, the Pistols autoattack is somewhat weak, and the Rifle’s autoattack is relatively strong compared to both before even looking closely at the damage values. We make this prediction by looking at the difference that exists between the two metrics, or the delay. The delay for the Rifle is about .1 second, the delay for the Pistol is about .3 seconds, and the delay for the Longbow is about .5 seconds.

This lines up exactly with where the actual damage values are at when you look at them in game. This is why I am very doubtful that it is intentional.

Speaking of “Auto attack problem”, my auto attacks don’t auto attack quite often of late. I attack something and then won’t attack again, even when there’s no interference (LoS, obstacle etc.).

Tintifax.2614:

No, digiowl, I’ve just tested it. You can’t cast faster manually than using autoattack. What Einlanzer means is that you don’t notice the skill activation delay when autoattacking. You don’t care that the skill activates half a second after the computer presses the button for you, as long as the skill fires, say, once a second. Although you notice it when casting manually.

In that case i am not sure it should be called the autoattack problem, as the problem is independent of autoattack. It is rather that autoattack masks the problem, that animations influence execution times and so distort the effectiveness of skills.

As such, i fear it to be a “side effect” of their “play the game, not the UI” mantra. I am almost surprised they bothered with tooltips at all, as the combat video they have up seems to describes a shooter rather than a RPG.

With added irony being that shooters are notorious for their latency sensitivity, yet supposedly they dropped GW1 style interrupts because they were latency sensitive.

Yet the game we have now is as much, or even more, latency sensitive.

Sadly it seems they have trashed their blog, as i suspect within it hid the article where they mention the combat being animation based…

Einlanzer.1627:

If what you say is true, it would be very likely that we would expect to see similar output from all autoattacks, when in fact we don’t. As I also outlined above, we can predict how weak an autoattack is relative to other autoattacks by looking at the difference between its activation speed and its total recast time due to the animation.

We can, for example, predict that the Longbow’s autoattack is very weak, the Pistols autoattack is somewhat weak, and the Rifle’s autoattack is relatively strong compared to both before even looking closely at the damage values. We make this prediction by looking at the difference that exists between the two metrics, or the delay. The delay for the Rifle is about .1 second, the delay for the Pistol is about .3 seconds, and the delay for the Longbow is about .5 seconds.

This lines up exactly with where the actual damage values are at when you look at them in game. This is why I am very doubtful that it is intentional.

Well, in my point of view it doesn’t depend on the difference between animation time and activation time, but on animation time alone. Because I think activation time has nothing to do with how often you can execute a skill per time unit, but is only about the delay between pressing the button and the skill taking effect. So yes, the longbow is weaker than the rifle because the animation time (effective cooldown) is higher without the damage being higher in an equal amount. If the rifle had an activation time of 0.25s instead, the difference to the animation time would be greater than with the longbow, but still the longbow would be weaker (even more so because the rifle would feel snappier without autoattack).

You expect all autoattacks (#1 skills) to have more or less equal dps and expect a bug since it isn’t so. Now I simply ask the question: why do the #1 skills of all weapons across all professions have to be equally strong? Isn’t it possible that the devs don’t try to balance skill for skill, but rather take all weapon skills into account? As in, maybe they feel that other longbow skills than #1 offer more than the respective rifle skills and so make the longbow autoattack intentionally weaker than the rifle autoattack by having a longer animation time?

Tintifax.2614:

Einlanzer.1627:

If what you say is true, it would be very likely that we would expect to see similar output from all autoattacks, when in fact we don’t. As I also outlined above, we can predict how weak an autoattack is relative to other autoattacks by looking at the difference between its activation speed and its total recast time due to the animation.

We can, for example, predict that the Longbow’s autoattack is very weak, the Pistols autoattack is somewhat weak, and the Rifle’s autoattack is relatively strong compared to both before even looking closely at the damage values. We make this prediction by looking at the difference that exists between the two metrics, or the delay. The delay for the Rifle is about .1 second, the delay for the Pistol is about .3 seconds, and the delay for the Longbow is about .5 seconds.

This lines up exactly with where the actual damage values are at when you look at them in game. This is why I am very doubtful that it is intentional.

Well, in my point of view it doesn’t depend on the difference between animation time and activation time, but on animation time alone. Because I think activation time has nothing to do with how often you can execute a skill per time unit, but is only about the delay between pressing the button and the skill taking effect. So yes, the longbow is weaker than the rifle because the animation time (effective cooldown) is higher without the damage being higher in an equal amount. If the rifle had an activation time of 0.25s instead, the difference to the animation time would be greater than with the longbow, but still the longbow would be weaker (even more so because the rifle would feel snappier without autoattack).

You expect all autoattacks (#1 skills) to have more or less equal dps and expect a bug since it isn’t so. Now I simply ask the question: why do the #1 skills of all weapons across all professions have to be equally strong? Isn’t it possible that the devs don’t try to balance skill for skill, but rather take all weapon skills into account? As in, maybe they feel that other longbow skills than #1 offer more than the respective rifle skills and so make the longbow autoattack intentionally weaker than the rifle autoattack by having a longer animation time?

You’re consistently misunderstanding me. For skills that aren’t on auto, the Activation speed is the only important metric, animation length is mostly irrelevant. For Autoattacks, it’s the opposite because the activation speed is absorbed into the Animation length. What I was saying is that the greater the difference between the activation speed and the animation length, the weaker the autoattack. You can see this across weapons. This is because the skill was balanced with the activation speed in mind and not with the animation. Therefore, the longer the animation relative to the activation speed, the slower the weapon actually fires vs what it was intended to fire, and therefore the weaker the attack is.

Your second paragraph is a strawman. Most autoattacks do equivalent DPS. The only ones that don’t are the ones that have large ‘hidden’ delays due to their animations being significantly longer than their activation speeds. Moreover, you can easily observe that the greater this delay (or the difference between the animation length and the activation speed) is, the weaker the autoattack is.

It really isn’t rocket science to reason out that this means it’s probably an oversight.

Eh, I’m seeing a lack of evidence. You’re stating that autoattack is causing delays between activation time and animation time yet the only statement you provide is support for the argument “there is a discrepancy between activation time and animation time”. Unless we’re talking about different “activation time”, it exists even when you don’t have skills on autoattack. Where is your evidence comparing spamming the #1 key and setting it on autoattack? Does the Vital Shot vs Bleeding Shot example not apply when people just spam the #1 key then? It’s a reasonable and plausible theory, I just don’t see enough evidence here to convince myself.

I use autoattack but often I spam the #1 key anyway. If others do the same then your theory should be able to be experienced firsthand by noticing differences between when people leave it up to the game to autoattack and when they spam the #1 key to override it.

Off topic: If the argument was the discrepancy between activation time and animation time, I would definitely have to agree. Mesmer scepter #1 definitely feels like the animation is delaying the skill to fire at the proper intervals (unless the proper intervals were intentionally designed to be so slow as to make the DPS worthless) and there are a bunch of other weapon/utility skills that I feel they have “weird, unexplainable” delays that take longer than they should.

ExZee.8109:

I use autoattack but often I spam the #1 key anyway. If others do the same then your theory should be able to be experienced firsthand by noticing differences between when people leave it up to the game to autoattack and when they spam the #1 key to override it.

No, like some other people you are misunderstanding what I’m saying. I never once suggested that you should experience this. In fact, the fact that this is precisely what doesn’t happen helps illustrate my point. Try rereading my original post.

ExZee.8109:

Off topic: If the argument was the discrepancy between activation time and animation time, I would definitely have to agree. Mesmer scepter #1 definitely feels like the animation is delaying the skill to fire at the proper intervals (unless the proper intervals were intentionally designed to be so slow as to make the DPS worthless) and there are a bunch of other weapon/utility skills that I feel they have “weird, unexplainable” delays that take longer than they should.

This is not off-topic, this is precisely what I’m saying. What annoys me is people acting like it isn’t a big deal when it’s a huge deal egregiously affecting the balance of several weapons across professions, and people arguing that it’s probably intended when evidence drawn from analysis and comparison makes it seem more likely that it’s an oversight resulting from a breakdown between skill development/balancing and animation design.

Einlanzer.1627:

No, like some other people you are misunderstanding what I’m saying. I never once suggested that you should experience this. In fact, the fact that this is precisely what doesn’t happen helps illustrate my point. Try rereading my original post.

Actually I was more or less forced to read it multiple times to understand what you were trying to say in the first place. If I’m not the only person misunderstanding, maybe the problem isn’t in the readers.

ExZee.8109:

Einlanzer.1627:

No, like some other people you are misunderstanding what I’m saying. I never once suggested that you should experience this. In fact, the fact that this is precisely what doesn’t happen helps illustrate my point. Try rereading my original post.

Actually I was more or less forced to read it multiple times to understand what you were trying to say in the first place. If I’m not the only person misunderstanding, maybe the problem isn’t in the readers.

Several other people were able to understand it just fine, and furthermore it isn’t a simple thing to articulate in a forum post.

edit: I just reread it and I don’t really see a problem, other than the topic itself being a little muddy.

For any number cruncher like me, the animation time is very important when determining the DPS of any skill or chain of skills, let alone the auto attack. It is also important since being locked into a skill’s animation means that other skills can’t be used as effectively, hence why it is on my thief the weapon/utility dodge skills are erratic. That or it is just random lag causing such an issue… pretty certain it is the auto attack though, since it tends to line up more and doesn’t affect regular dodge.

So I do fear that this issue extends beyond auto attack. It is just a lot more apparent in auto attacks, though, since those are used more than once every 30 seconds.

I still don’t know about this. Take the longer auto attack time of the long bow over the short bow. The short bow does a much better autoattack but the long bow otherwise has much higher damage skills. Taking in the whole picture, it might be possible that the short bow’s strong auto-attack compensates for the significantly lower damage otherwise.

As for Anet having tools to parse stuff, they have referenced having such tool, but they didn’t go into specifics. It might not affect this sort of thing, I don’t know. But I do know they have, according to them, a number of tools at their disposal to parse stuff.

It would be sort of silly if they didn’t.

Blood Red Arachnid.2493:

For any number cruncher like me, the animation time is very important when determining the DPS of any skill or chain of skills, let alone the auto attack. It is also important since being locked into a skill’s animation means that other skills can’t be used as effectively, hence why it is on my thief the weapon/utility dodge skills are erratic. That or it is just random lag causing such an issue… pretty certain it is the auto attack though, since it tends to line up more and doesn’t affect regular dodge.

So I do fear that this issue extends beyond auto attack. It is just a lot more apparent in auto attacks, though, since those are used more than once every 30 seconds.

Scary numbers person, is that you?

a-hem… What you’re saying makes sense. I’d not even thought about it before, I just assumed my timing was off, but… Just reading what you’ve said makes a little lightbulb go off inside my head.

Vayne.8563:

I still don’t know about this. Take the longer auto attack time of the long bow over the short bow. The short bow does a much better autoattack but the long bow otherwise has much higher damage skills. Taking in the whole picture, it might be possible that the short bow’s strong auto-attack compensates for the significantly lower damage otherwise.

The problem with the LB having supposed “higher damage skills” is that the SB outpaces those just with “auto-attack” damage.

With no traits or gear and steady weapons against the heavy golem in SPvP I found;
Long Bow #1 long range: 41 damage 1.25s animation.
Long Bow #1 med range: 30 damage 1.25s animation.
Long Bow #1 short range: 23 damage 1.25s animation.
Long Bow #2 170 damage over 5.5 second animation (5s “cast” time displayed)

Shortbow #1 18 damage, 128 bleed (3s) over an animation that I couldn’t accurately time. (my reflexes kept giving me numbers between 0.76 to 0.89 seconds)

Given that the Shortbow (along with dagger #1.1 and 1.2 for thief and necro) doesn’t display the “cast” time we can assume that the animation and cast time are shorter than 0.25 seconds, otherwise there isn’t a reason to NOT have a time displayed on the tooltip.

Worst case scenario, the animation of the Shortbow is 0.5 seconds, and deals 18 direct damage, we can multiply that damage by 11 to equal the time of Rapid Fire and we have 198 damage for the SB and 170 damage for rapid fire over the same 5.5 seconds, It only gets worse for the LB if the animation of the SB is faster than 0.5 seconds.

[7th edit] Using a 30/30/10/0/0 build the Shortbow can stack a 4s bleed up to 8 times (exact) which means that the Shortbow is pulling off 8 attacks over 4 seconds, as such I can say with confidence that the animation for the SB is 0.5 seconds.

The only Long Bow ability that out paces the Shortbow is Barrage.
For added laughs; over 5 seconds at long range the longbow #1 does 164 damage vs the #2 doing 170 damage over 5.5 seconds, if one takes into account that extra 0.5 seconds then the long bow #1 OUT DAMAGES the #2 over the same 5.5 second time period (assuming nothing crits).

Einlanzer.1627:

Several other people were able to understand it just fine, and furthermore it isn’t a simple thing to articulate in a forum post.

edit: I just reread it and I don’t really see a problem, other than the topic itself being a little muddy.

Ok, let’s try this one last time. You said that the fact people aren’t noticing a difference between the auto-attack and spamming the #1 key helps illustrate your point. Your point is that when the #1 skill isn’t chained, the auto-attack results in a slower rate-of-fire from its original design.

1. Where is your data showing the (amount of) delay from its original design? You provide one analogical example that isn’t based on research or fact. While those that saw and agree with your points have given more ideas to do research on the topic, non have provided actual data that proves your hypothesis in the first place. Coffeebot.3921 is the only person to do any physical research on the topic, yet he didn’t compare his data to another set of data where he manually pressed the #1 key instead of leaving it on auto-attack. Therefore, while it sounds plausible, your argument isn’t going to convince anyone with half a brain without some form of factual evidence.

2. I don’t see how people not noticing a difference between auto-attack and manual key pressing is helping to illustrate your point. If it’s such a major problem as you suggest, it would be obvious to any and every player to the point it seriously affects their gameplay. The forums would be flooded with autoattack users complaining that they’re getting out-DPSed by key spammers, yet the case is exactly the opposite: people don’t see it as a big deal at all. Perhaps the delay is so minimal it doesn’t affect the gameplay to the point it needs a major fix. Again, only factual evidence would prove or disprove any argument you may have.

Blood Red Arachnid.2493:

So I do fear that this issue extends beyond auto attack. It is just a lot more apparent in auto attacks, though, since those are used more than once every 30 seconds.

This is what I’ve also mentioned as “off topic” and according to the thread title, it really is. If it’s a problem where the discrepancy between the activation delay and animation delay applies to a lot of other skills regardless of whether they’re on autoattack or not, then it doesn’t really have anything to do with autoattack at all other than the fact that the autoattack makes it most noticeable. Then this discrepancy should be up in a newly titled thread for argument whether it was intentional or not, and whether it has screwed up balancing to the point it needs a major makeover or not. Can’t do any of this without some numbers to make a convincing argument to the devs.

ExZee.8109:

Coffeebot.3921 is the only person to do any physical research on the topic, yet he didn’t compare his data to another set of data where he manually pressed the #1 key instead of leaving it on auto-attack.

Manually pressing #1 for Long Bow over 10.04 seconds;
1) 1.38s
2) 1.17s
3) 1.15s
4) 1.26s
5) 1.30s
6) 2.5s
7) 1.28s
Average of 1.43s, remove the outlier and average drops to 1.25(6 recurring) seconds.

Furious button mashing version over 20.28s;
1) 0.56s (outlier)
2) 1.25s
3) 1.1s
4) 1.19s
5) 1.27s
6)1.30s
7) 2.37s (outlier)
8) 1.27s
9)1.12s
10) 2.64s (outlier)
11) 2.44s (outlier)
12) 1.35s
13) 1.18s
14) 1.17s
Average total: 1.45s (reasonably consistent with sedate test)
Average after outlier removal: 1.23s (reasonably consistent with sedate test)

ExZee.8109:

Einlanzer.1627:

Several other people were able to understand it just fine, and furthermore it isn’t a simple thing to articulate in a forum post.

edit: I just reread it and I don’t really see a problem, other than the topic itself being a little muddy.

Ok, let’s try this one last time. You said that the fact people aren’t noticing a difference between the auto-attack and spamming the #1 key helps illustrate your point. Your point is that when the #1 skill isn’t chained, the auto-attack results in a slower rate-of-fire from its original design.

1. Where is your data showing the (amount of) delay from its original design? You provide one analogical example that isn’t based on research or fact. While those that saw and agree with your points have given more ideas to do research on the topic, non have provided actual data that proves your hypothesis in the first place. Coffeebot.3921 is the only person to do any physical research on the topic, yet he didn’t compare his data to another set of data where he manually pressed the #1 key instead of leaving it on auto-attack. Therefore, while it sounds plausible, your argument isn’t going to convince anyone with half a brain without some form of factual evidence.

2. I don’t see how people not noticing a difference between auto-attack and manual key pressing is helping to illustrate your point. If it’s such a major problem as you suggest, it would be obvious to any and every player to the point it seriously affects their gameplay. The forums would be flooded with autoattack users complaining that they’re getting out-DPSed by key spammers, yet the case is exactly the opposite: people don’t see it as a big deal at all. Perhaps the delay is so minimal it doesn’t affect the gameplay to the point it needs a major fix. Again, only factual evidence would prove or disprove any argument you may have.

Blood Red Arachnid.2493:

So I do fear that this issue extends beyond auto attack. It is just a lot more apparent in auto attacks, though, since those are used more than once every 30 seconds.

This is what I’ve also mentioned as “off topic” and according to the thread title, it really is. If it’s a problem where the discrepancy between the activation delay and animation delay applies to a lot of other skills regardless of whether they’re on autoattack or not, then it doesn’t really have anything to do with autoattack at all other than the fact that the autoattack makes it most noticeable. Then this discrepancy should be up in a newly titled thread for argument whether it was intentional or not, and whether it has screwed up balancing to the point it needs a major makeover or not. Can’t do any of this without some numbers to make a convincing argument to the devs.

Look, you’re on the wrong page somewhere, but I can’t figure out where. I think you’re just over-complicating what I’m trying to say in your head.

My original hypothesis was this and only this – the skills were weighted and balanced around their activation speeds. This is a given, otherwise that metric wouldn’t exist. And in most cases this is fine, because you aren’t using the same skill more than once every few seconds. When a skill is set to autoattack, though, (typically #1 skills) its activation speed is mostly irrelevant because it is superseded by the animation speed, which is almost always longer (the degree of difference varies from weapon to weapon). This means that many of the #1 skills (or autoattack skills) have refire rates that are slower than they should be and are therefore weaker than they were originally intended to be. This has absolutely nothing to do with whether you have the skill actually set to autoattack or are pressing 1 repeatedly, because the same animation plays out regardless; I don’t even understand where that entered the equation or why it’s being discussed.

The evidence that this is not intended and is instead an oversight is very convincing when you start looking at the data for #1 skills across weapons and professions, as I have illustrated. I don’t have the exact numbers, but it’s very easy to determine even from a cursory analysis that the DPS for both Pistols and Longbows is lacking in comparison to Rifles, etc, and that most of this deficiency comes from the #1 skill. Specifically, it comes from the fact that those skills have animation durations that are far longer than the skill’s activation speed (that would be the ‘delay’ I was referring to) when they were probably balanced around the latter.