Are +5 Infusions really worth 4k DPS? — Guild Wars 2 Forums
Home Fractals/Dungeons/Strike Missions/Raids

Are +5 Infusions really worth 4k DPS?

Dahkeus.8243Dahkeus.8243 Member ✭✭✭✭

Edit - Title adjusted from 6k to 4k after incorporating feedback in this thread

Ok, looking for some experienced insight here. I'm trying to get a character back raid-ready and am focusing on condi banner warrior. I've followed the SnowCrows setup and have all the gear/traits the same, except that I don't have any of the +5 infusions. Beyond that, I have full ascended gear with the exact same rune/sigils. I also am testing on the exact same golem settings.

Here's the SnowCrows benchmark video:

And here's a video of me matching the same first 30 seconds of the rotation:

I noticed that I'm not hitting the same 25 stacks of Furious, so maybe there's something else at play? However, I'm landing all the same skills shown in the benchmark (including the burst skills) within the same window of Berserk.

Maybe there's something else I'm missing, but for all my testing, I can't seem to get higher than 27k DPS at any point. I know that Expertise is a big stat for condi builds and that some of the +5 infusions consist of this stat, but this still feels like a huge gap between what I am producing and what the benchmark is producing. Is it really just the infusions that make the difference?

(Also, here's a link to a log from a similar test against the golem: https://dps.report/WN4r-20200527-172833_golem)

Comments

  • RaidsAreEasyAF.8652RaidsAreEasyAF.8652 Member ✭✭✭
    edited May 27, 2020

    They are definitely not worth 6k dps.
    your rotation just isn't super clean, you start the fight with an auto attack and on your first shattering blow you are already 2 seconds behind the benchmark. Speed is important for a rotation and you can already lack behind 2 seconds in the first 3 seconds of a log. That's a huge time loss.

    Was only looking at the log. In the video you also don't reset your burst skill with your healing skill.

  • Opopanax.1803Opopanax.1803 Member ✭✭✭✭

    Is it really 6k extra damage per second, or damage per rotation?

  • Dahkeus.8243Dahkeus.8243 Member ✭✭✭✭
    edited May 28, 2020

    @RaidsAreEasyAF.8652 said:
    They are definitely not worth 6k dps.
    your rotation just isn't super clean, you start the fight with an auto attack and on your first shattering blow you are already 2 seconds behind the benchmark. Speed is important for a rotation and you can already lack behind 2 seconds in the first 3 seconds of a log. That's a huge time loss.

    Was only looking at the log. In the video you also don't reset your burst skill with your healing skill.

    Hmm, I appreciate you taking the time to speak up, but your insights don't really add up. The fight doesn't start with an auto attack in either the log or the video. The first attack in the log is Pin Down since the auto-attack shown is interrupted during the weapon swap. Scorched earth is also cast twice with the heal in both video and log. It's hard to see in the video, but it's shown in the log.

    Your timing in comparing the videos may be off, but you can count the cast burst skills to see that they match. In my video, there's 6 Scorched Earths and 3 Flaming Fury casts before berserk runs out. That is one shy of the SC video, but only because I missed one at the end. I've tested the rotation again without missing that one and it's still below 27k DPS.

    Here's an updated log with that last scorched earth included:
    https://dps.report/5LNK-20200527-190156_golem

    Updated image from recent test:

  • Dahkeus.8243Dahkeus.8243 Member ✭✭✭✭
    edited May 28, 2020

    @Opopanax.1803 said:
    Is it really 6k extra damage per second, or damage per rotation?

    Damage per second. You can watch my DPS during the vid or look at the full graph in the log. At no point do I hit or exceed 28k DPS.

    I'd really like to hear "No, you just forgot " something, but I've been sitting here trying to figure this out for quite a while now and no matter how much I mirror exactly what is shown from the benchmark/raid guide without having the +5 infusions for armor, I cannot come close to the 33-34k DPS being shown.

  • Lan Deathrider.5910Lan Deathrider.5910 Member ✭✭✭✭

    I do not have the stat infusions either, and I can hit 35k on the golem with Power Decapitate Berserker, which is pretty much the benchmark for that build. I think you may be missing some of the skill activations be a few frames, which adds up over the rotation. Queue them up as the previous skill is activating.

  • Alex.9106Alex.9106 Member

    You started the fight with torch 5 while standing in the golem which instantly puts you infight, if you see the benchmark video he starts casting torch 5 while standing away from the golem then switching weapons casting lb 5 and running into the golem, if you check the sc log you notice it doesnt even show his torch 5 and lb 5 because he wasnt infight yet. Your lb 5 showing up in your log is a sign that you were slow at the start, it shouldnt be visible, as arc only starts logging once you are in combat. (This is just concerning the opening first 2 seconds)

  • Dahkeus.8243Dahkeus.8243 Member ✭✭✭✭

    @Lan Deathrider.5910 said:
    I do not have the stat infusions either, and I can hit 35k on the golem with Power Decapitate Berserker, which is pretty much the benchmark for that build. I think you may be missing some of the skill activations be a few frames, which adds up over the rotation. Queue them up as the previous skill is activating.

    Which skill activations?

    I get that it seems that I should be missing something like this, but the logs and vids all show me hitting them.

  • Shikaru.7618Shikaru.7618 Member ✭✭✭
    edited May 28, 2020
    1. Your opening torch 5 needs to be done out of combat so that you dont have a weapon swap cooldown when going into bow. Notice in the SC video when they're in bow the first time the weapon swap is not on cooldown.
    2. On the third row aka the second time you're in sword, you have a full cast of sever artery in between shattering blow and flaming flurry.
    3. All of the red squares in your rotation log are moments where you cancelled auto attacks and thus are losing time. If you mouse over the red squares you'll see that you're spending anywhere from 50-250ms on cancelled autos each time. That is time that can be shaved off which will end up adding up to whole seconds where you're behind simply because you're not hitting skills fast enough.

    Doing some quick math, if you shaved off all of the cancelled skills, you would have dealt the same damage in 29.888 seconds instead of 33.019 which would have raised your dps from 25,819 to 28,523. It makes a pretty big difference.

    To get ahead of potential questions you may have. Yes the timing is actually that tight if you want to be optimal. Do you need to achieve these tight timings in raid scenarios to be successful? absolutely not.

  • Lan Deathrider.5910Lan Deathrider.5910 Member ✭✭✭✭

    @Dahkeus.8243 said:

    @Lan Deathrider.5910 said:
    I do not have the stat infusions either, and I can hit 35k on the golem with Power Decapitate Berserker, which is pretty much the benchmark for that build. I think you may be missing some of the skill activations be a few frames, which adds up over the rotation. Queue them up as the previous skill is activating.

    Which skill activations?

    I get that it seems that I should be missing something like this, but the logs and vids all show me hitting them.

    I mean that your timing may be off just slightly. It helps if you are pushing DPS to queue the next skill while the current one is activating.

  • Dahkeus.8243Dahkeus.8243 Member ✭✭✭✭

    @Shikaru.7618 said:
    1. Your opening torch 5 needs to be done out of combat so that you dont have a weapon swap cooldown when going into bow. Notice in the SC video when they're in bow the first time the weapon swap is not on cooldown.
    2. On the third row aka the second time you're in sword, you have a full cast of sever artery in between shattering blow and flaming flurry.
    3. All of the red squares in your rotation log are moments where you cancelled auto attacks and thus are losing time. If you mouse over the red squares you'll see that you're spending anywhere from 50-250ms on cancelled autos each time. That is time that can be shaved off which will end up adding up to whole seconds where you're behind simply because you're not hitting skills fast enough.

    Doing some quick math, if you shaved off all of the cancelled skills, you would have dealt the same damage in 29.888 seconds instead of 33.019 which would have raised your dps from 25,819 to 28,523. It makes a pretty big difference.

    To get ahead of potential questions you may have. Yes the timing is actually that tight if you want to be optimal. Do you need to achieve these tight timings in raid scenarios to be successful? absolutely not.

    Helpful feedback. Being able to tighten my rotation to 28.5k DPS sounds reasonable. By following that math, it still puts it at between 4.5k and 5.5k DPS loss from not having the infusions, meaning that I'm losing around 15% DPS without having them.

    I'll practice more and see what my actual DPS comes out to with my rotation tightened to match what you're describing.

    @Lan Deathrider.5910 said:
    I mean that your timing may be off just slightly. It helps if you are pushing DPS to queue the next skill while the current one is activating.

    You should be able to see me doing that in the posted videos.

  • Alex.9106Alex.9106 Member
    edited May 28, 2020

    Infusions increase your dps by rougly ~1k, everything else is a rotation issue

  • Shikaru.7618Shikaru.7618 Member ✭✭✭
    edited May 28, 2020

    @Dahkeus.8243 said:

    @Shikaru.7618 said:
    1. Your opening torch 5 needs to be done out of combat so that you dont have a weapon swap cooldown when going into bow. Notice in the SC video when they're in bow the first time the weapon swap is not on cooldown.
    2. On the third row aka the second time you're in sword, you have a full cast of sever artery in between shattering blow and flaming flurry.
    3. All of the red squares in your rotation log are moments where you cancelled auto attacks and thus are losing time. If you mouse over the red squares you'll see that you're spending anywhere from 50-250ms on cancelled autos each time. That is time that can be shaved off which will end up adding up to whole seconds where you're behind simply because you're not hitting skills fast enough.

    Doing some quick math, if you shaved off all of the cancelled skills, you would have dealt the same damage in 29.888 seconds instead of 33.019 which would have raised your dps from 25,819 to 28,523. It makes a pretty big difference.

    To get ahead of potential questions you may have. Yes the timing is actually that tight if you want to be optimal. Do you need to achieve these tight timings in raid scenarios to be successful? absolutely not.

    Helpful feedback. Being able to tighten my rotation to 28.5k DPS sounds reasonable. By following that math, it still puts it at between 4.5k and 5.5k DPS loss from not having the infusions, meaning that I'm losing around 15% DPS without having them.

    I'll practice more and see what my actual DPS comes out to with my rotation tightened to match what you're describing.

    @Lan Deathrider.5910 said:
    I mean that your timing may be off just slightly. It helps if you are pushing DPS to queue the next skill while the current one is activating.

    You should be able to see me doing that in the posted videos.

    I only included cancelled autos. There are instances where you have completed autos in between skill usage that I didn't take out since it'd make the quick math too difficult but those will also contribute a lot to your dps as well. Typically auto chains will do more damage the further along in the chain you are so a hypothetical auto chain could look something like this:

    Auto 1 - 1000 damage
    Auto 2 - 2000 damage
    Auto 3 - 5000 damage

    Now lets say in a given weapon rotation you can fit 3 big skills and 3 autos before you must weapon swap. Here's what optimal looks like:
    Big skill - Big skill - Big skill - Auto 1 - Auto 2 - Auto 3 - weapon swap
    which means your auto chain did 8000 damage total.

    Here's what bad looks like:
    Big skill - Big skill - Auto 1 - Big skill - Auto 1 - Auto 2 - weapon swap
    Notice in this example your autos only did 4000 damage total in the same time span.

    You can actually see an example of this in your log where you fit auto 1 in between 2 big skills. If you instead tighten up the rotation, you can move the time spent on auto 1 to the end and get an auto 3 in there.

  • Dahkeus.8243Dahkeus.8243 Member ✭✭✭✭
    edited May 28, 2020

    @Shikaru.7618 said:

    @Dahkeus.8243 said:

    @Shikaru.7618 said:
    1. Your opening torch 5 needs to be done out of combat so that you dont have a weapon swap cooldown when going into bow. Notice in the SC video when they're in bow the first time the weapon swap is not on cooldown.
    2. On the third row aka the second time you're in sword, you have a full cast of sever artery in between shattering blow and flaming flurry.
    3. All of the red squares in your rotation log are moments where you cancelled auto attacks and thus are losing time. If you mouse over the red squares you'll see that you're spending anywhere from 50-250ms on cancelled autos each time. That is time that can be shaved off which will end up adding up to whole seconds where you're behind simply because you're not hitting skills fast enough.

    Doing some quick math, if you shaved off all of the cancelled skills, you would have dealt the same damage in 29.888 seconds instead of 33.019 which would have raised your dps from 25,819 to 28,523. It makes a pretty big difference.

    To get ahead of potential questions you may have. Yes the timing is actually that tight if you want to be optimal. Do you need to achieve these tight timings in raid scenarios to be successful? absolutely not.

    Helpful feedback. Being able to tighten my rotation to 28.5k DPS sounds reasonable. By following that math, it still puts it at between 4.5k and 5.5k DPS loss from not having the infusions, meaning that I'm losing around 15% DPS without having them.

    I'll practice more and see what my actual DPS comes out to with my rotation tightened to match what you're describing.

    @Lan Deathrider.5910 said:
    I mean that your timing may be off just slightly. It helps if you are pushing DPS to queue the next skill while the current one is activating.

    You should be able to see me doing that in the posted videos.

    I only included cancelled autos. There are instances where you have completed autos in between skill usage that I didn't take out since it'd make the quick math too difficult but those will also contribute a lot to your dps as well. Typically auto chains will do more damage the further along in the chain you are so a hypothetical auto chain could look something like this:

    Auto 1 - 1000 damage
    Auto 2 - 2000 damage
    Auto 3 - 5000 damage

    Now lets say in a given weapon rotation you can fit 3 big skills and 3 autos before you must weapon swap. Here's what optimal looks like:
    Big skill - Big skill - Big skill - Auto 1 - Auto 2 - Auto 3 - weapon swap
    which means your auto chain did 8000 damage total.

    Here's what bad looks like:
    Big skill - Big skill - Auto 1 - Big skill - Auto 1 - Auto 2 - weapon swap
    Notice in this example your autos only did 4000 damage total in the same time span.

    You can actually see an example of this in your log where you fit auto 1 in between 2 big skills. If you instead tighten up the rotation, you can move the time spent on auto 1 to the end and get an auto 3 in there.

    I'm making some progress on some of the stuff mentioned to get closer to matching the exact rotation in terms of included full auto attacks, but the DPS numbers aren't really changing much as I get closer to having this absolutely perfect. There's still a very big gap between the ~27k mark and the 33k mark, but I'll reply with an updated log/video when I get it perfected to see where exactly it lands (will probably be tomorrow since I won't have another chance tonight).

    @Alex.9106 said:
    Infusions increase your dps by rougly ~1k, everything else is a rotation issue

    Take a page from Shikaru's book and put some actual substance behind your claims.

  • Shikaru.7618Shikaru.7618 Member ✭✭✭
    edited May 28, 2020

    @Dahkeus.8243 said:

    @Shikaru.7618 said:

    @Dahkeus.8243 said:

    @Shikaru.7618 said:
    1. Your opening torch 5 needs to be done out of combat so that you dont have a weapon swap cooldown when going into bow. Notice in the SC video when they're in bow the first time the weapon swap is not on cooldown.
    2. On the third row aka the second time you're in sword, you have a full cast of sever artery in between shattering blow and flaming flurry.
    3. All of the red squares in your rotation log are moments where you cancelled auto attacks and thus are losing time. If you mouse over the red squares you'll see that you're spending anywhere from 50-250ms on cancelled autos each time. That is time that can be shaved off which will end up adding up to whole seconds where you're behind simply because you're not hitting skills fast enough.

    Doing some quick math, if you shaved off all of the cancelled skills, you would have dealt the same damage in 29.888 seconds instead of 33.019 which would have raised your dps from 25,819 to 28,523. It makes a pretty big difference.

    To get ahead of potential questions you may have. Yes the timing is actually that tight if you want to be optimal. Do you need to achieve these tight timings in raid scenarios to be successful? absolutely not.

    Helpful feedback. Being able to tighten my rotation to 28.5k DPS sounds reasonable. By following that math, it still puts it at between 4.5k and 5.5k DPS loss from not having the infusions, meaning that I'm losing around 15% DPS without having them.

    I'll practice more and see what my actual DPS comes out to with my rotation tightened to match what you're describing.

    @Lan Deathrider.5910 said:
    I mean that your timing may be off just slightly. It helps if you are pushing DPS to queue the next skill while the current one is activating.

    You should be able to see me doing that in the posted videos.

    I only included cancelled autos. There are instances where you have completed autos in between skill usage that I didn't take out since it'd make the quick math too difficult but those will also contribute a lot to your dps as well. Typically auto chains will do more damage the further along in the chain you are so a hypothetical auto chain could look something like this:

    Auto 1 - 1000 damage
    Auto 2 - 2000 damage
    Auto 3 - 5000 damage

    Now lets say in a given weapon rotation you can fit 3 big skills and 3 autos before you must weapon swap. Here's what optimal looks like:
    Big skill - Big skill - Big skill - Auto 1 - Auto 2 - Auto 3 - weapon swap
    which means your auto chain did 8000 damage total.

    Here's what bad looks like:
    Big skill - Big skill - Auto 1 - Big skill - Auto 1 - Auto 2 - weapon swap
    Notice in this example your autos only did 4000 damage total in the same time span.

    You can actually see an example of this in your log where you fit auto 1 in between 2 big skills. If you instead tighten up the rotation, you can move the time spent on auto 1 to the end and get an auto 3 in there.

    I'm making some progress on some of the stuff mentioned to get closer to matching the exact rotation in terms of included full auto attacks, but the DPS numbers aren't really changing much as I get closer to having this absolutely perfect. There's still a very big gap between the ~27k mark and the 33k mark, but I'll reply with an updated log/video when I get it perfected to see where exactly it lands (will probably be tomorrow since I won't have another chance tonight).

    @Alex.9106 said:
    Infusions increase your dps by rougly ~1k, everything else is a rotation issue

    Take a page from Shikaru's book and put some actual substance behind your claims.

    Combing over your rotation a bit more, I also noticed that you're missing one cast of long bow 2 each time you're in berserk longbow. You should be doing swap from sword, lb 2, f1, filler stuff, f1, lb2 before the swap comes of cd with some variation when you land on lb after opening burst. Try to fix that in your rotation.

  • Henry.5713Henry.5713 Member ✭✭✭✭
    edited May 28, 2020

    The build in question runs 11 Malign and 7 Spiteful infusions (to allow you to max out Condition Duration). We are talking about 55 Condition Damage and 35 Expterise (~2.3% Condition Duration). You should be hovering at around 3000 Condition Damage in total. That 55 Condition Damage coming from your infusions is an increase of roughly ~2%.

    There is no way that increasing your main damage stat by 2% and prolonging the duration of your conditions by another 2% would ever come close increasing your damage by 6-7k. This would be a wooping 25% increase of your overall damage.

    Progress isn't made by early risers. It's made by lazy men trying to find easier ways to do something. ~ Robert Heinlein

  • Cyninja.2954Cyninja.2954 Member ✭✭✭✭
    edited May 28, 2020

    https://dps.report/Nrn9-20191003-150603_golem
    Dps log of the benchmark video.

    0.17 Headbutt
    1.9 Berserk
    2.17 Dual Shot
    2.2 Scorched Earth
    2.5 Blood Reckonig
    2.85 Dual Shot
    2.9 Scorched Eartch
    3.2 Fan of Fire
    3.5 Weapon Swap

    https://dps.report/5LNK-20200527-190156_golem
    Your latest log.

    0.17 Sever Artery (not sure why you would not start on Headbutt)
    0.1 Weapon Swap (wasted weapon swap)
    0.45 Pin Down (probably fat fingered)
    1.65 Dual Shot
    1.69 Headbutt (you are already 1.5 second behind ideal rotation)
    3.45 Berserk
    3.74 Scorched Earth
    4.1 Blood Reckoning
    4.3 Scorched Earth
    4.73 Dual Shot
    5.0 Fan of Fire

    Look at the exact amount of skills in the ideal rotation. The simple rotation including auto attacks is almost a block, indicating that at nearly every available time frame a skill was cast or active(the fact that most skills have similar baseline cast times is beneficial here).

    Now take a look at your simple rotation. It's a string of different long lines of skills between weapon swaps. This indicates that your are either slower in executing/chaining skills, or miss clicking skills. Especially your sword rotation is sloppy compared to the benchmark.

    As far as infusions and the damage increase they bring, this was mentioned and Henry.5713 did some quick math for you. It's around 1-1.5k depending on build (power or condi) and if one is crit capped before (or condi duration capped) or not. Those 2 factors make up the biggest difference. Based on personal testing of myself, I can say that calculating 1k for infusions is correct. You can easily try this out by simply taking lower level food or reducing your setup by approximately 90 stat points (you could take off spotter for example for ease of testing, or add some effect worth around 90 beneficial stat points).

  • @Dahkeus.8243 said:
    Hmm, I appreciate you taking the time to speak up, but your insights don't really add up. The fight doesn't start with an auto attack in either the log or the video. The first attack in the log is Pin Down since the auto-attack shown is interrupted during the weapon swap. Scorched earth is also cast twice with the heal in both video and log. It's hard to see in the video, but it's shown in the log.

    Your timing in comparing the videos may be off, but you can count the cast burst skills to see that they match. In my video, there's 6 Scorched Earths and 3 Flaming Fury casts before berserk runs out. That is one shy of the SC video, but only because I missed one at the end. I've tested the rotation again without missing that one and it's still below 27k DPS.

    Here's an updated log with that last scorched earth included:
    https://dps.report/5LNK-20200527-190156_golem

    Updated image from recent test:

    No, you dont cast that LB Burst Skill in the Video. Its at 0:04. You cast your heal skill first and then the LB skill. Your adrenaline even stays full until you cast its after the healskill.
    You also interupting your Auto Attack doesnt mean that you dont start the fight with it. It means that you wasted time. On your first Weaponswap (or rather your second) you are already 2 seconds behind.

  • DutchRiders.2871DutchRiders.2871 Member ✭✭✭

    @Alex.9106 said:
    Infusions increase your dps by rougly ~1k, everything else is a rotation issue

    On the golem it can get a bit wonky cause of the 80% number. If that 1k allows you to reach that 80% just after your burst it inflates the number. On arc it should be around 1k dps.

    But yeah to the starter just practice your rotation and keep in mind some rotations are grinded by the best gw2 players (not all are grinded).

  • Alex.9106Alex.9106 Member
    edited May 28, 2020

    @Dahkeus.8243 said:

    @Alex.9106 said:
    Infusions increase your dps by rougly ~1k, everything else is a rotation issue

    Take a page from Shikaru's book and put some actual substance behind your claims.

    This is not a scientific forum, but whatever here you go:

    If you look at the benchmark the damage dealt by conditions from Luna Nightstalker consist of:
    35.86 average stacks of bleeding
    20.18 average stacks of burning
    3.26 average stacks of poison

    Burning adds 17641.3 dps
    bleeding adds 10703.6 dps
    poison adds 1024.4 dps

    Now infusion give you 11*5= 55 condi dmg and 7*5=35 expertise as stats.
    35/15 = 2.33 So the 35 expertise increase your condition duration by 2.33 percent.

    So lets see how much dps you lose if the conditions lose 2.33 percent of their duration.

    Burning: (17641.3*100)/102.33 = 17239.62
    Bleeding: (10703.6*100)/102.33 = 10459.88
    Poison: (1024.4*100)/102.33 = 1001.07

    So you lose (17641.3+10703.6+1024.4)-(17239.62+10459.88+1001.07)=668.73 dps

    Now lets see how much dps you lose if you miss 55 condi dmg stats.
    Formula for condi dmg is:
    Burning: (0.155 * Condition Damage) + 131 per stack per second
    Bleeding: (0.06 * Condition Damage) + 22 per stack per second
    Poison: (0.06 * Condition Damage) + 33.5 per stack per second

    This gives us:
    0.155*55=8.525 dps loss per stack of burning
    0.06*55=3.3 dps loss per stack of bleeding/poison

    So now we need to compute the avg number of stacks of conditions after our 2.33 duration increase is missing:
    Burning: (35.86*100)/102.33=35.04
    Bleeding: (20.18*100)/102.33=19.72
    Poison: (3.26*100)/102.33=3.19

    Now lets see:
    35.04*8.525=298.72 DPS loss for burning
    19.72*3.3=65.08 DPS loss for bleeding
    3.19*3.3=10.53 DPS loss for poison

    With this we have 668.73+298.72+65.08+10.53=1043.06 total dps loss if you dont have infusions.

    So the fact that the golem has a fixed number of health will create some problems, because depending on which part of the rotation you end at the golem your dps may experience some differences.
    But it should not be to far off from our calculated number of 1043.06 dps loss.

    What this also tells you is that should you ever get infusions you probably want to get the expertise infusions first.

  • Dahkeus.8243Dahkeus.8243 Member ✭✭✭✭

    @Shikaru.7618 said:

    @Dahkeus.8243 said:

    @Shikaru.7618 said:

    @Dahkeus.8243 said:

    @Shikaru.7618 said:
    1. Your opening torch 5 needs to be done out of combat so that you dont have a weapon swap cooldown when going into bow. Notice in the SC video when they're in bow the first time the weapon swap is not on cooldown.
    2. On the third row aka the second time you're in sword, you have a full cast of sever artery in between shattering blow and flaming flurry.
    3. All of the red squares in your rotation log are moments where you cancelled auto attacks and thus are losing time. If you mouse over the red squares you'll see that you're spending anywhere from 50-250ms on cancelled autos each time. That is time that can be shaved off which will end up adding up to whole seconds where you're behind simply because you're not hitting skills fast enough.

    Doing some quick math, if you shaved off all of the cancelled skills, you would have dealt the same damage in 29.888 seconds instead of 33.019 which would have raised your dps from 25,819 to 28,523. It makes a pretty big difference.

    To get ahead of potential questions you may have. Yes the timing is actually that tight if you want to be optimal. Do you need to achieve these tight timings in raid scenarios to be successful? absolutely not.

    Helpful feedback. Being able to tighten my rotation to 28.5k DPS sounds reasonable. By following that math, it still puts it at between 4.5k and 5.5k DPS loss from not having the infusions, meaning that I'm losing around 15% DPS without having them.

    I'll practice more and see what my actual DPS comes out to with my rotation tightened to match what you're describing.

    @Lan Deathrider.5910 said:
    I mean that your timing may be off just slightly. It helps if you are pushing DPS to queue the next skill while the current one is activating.

    You should be able to see me doing that in the posted videos.

    I only included cancelled autos. There are instances where you have completed autos in between skill usage that I didn't take out since it'd make the quick math too difficult but those will also contribute a lot to your dps as well. Typically auto chains will do more damage the further along in the chain you are so a hypothetical auto chain could look something like this:

    Auto 1 - 1000 damage
    Auto 2 - 2000 damage
    Auto 3 - 5000 damage

    Now lets say in a given weapon rotation you can fit 3 big skills and 3 autos before you must weapon swap. Here's what optimal looks like:
    Big skill - Big skill - Big skill - Auto 1 - Auto 2 - Auto 3 - weapon swap
    which means your auto chain did 8000 damage total.

    Here's what bad looks like:
    Big skill - Big skill - Auto 1 - Big skill - Auto 1 - Auto 2 - weapon swap
    Notice in this example your autos only did 4000 damage total in the same time span.

    You can actually see an example of this in your log where you fit auto 1 in between 2 big skills. If you instead tighten up the rotation, you can move the time spent on auto 1 to the end and get an auto 3 in there.

    I'm making some progress on some of the stuff mentioned to get closer to matching the exact rotation in terms of included full auto attacks, but the DPS numbers aren't really changing much as I get closer to having this absolutely perfect. There's still a very big gap between the ~27k mark and the 33k mark, but I'll reply with an updated log/video when I get it perfected to see where exactly it lands (will probably be tomorrow since I won't have another chance tonight).

    @Alex.9106 said:
    Infusions increase your dps by rougly ~1k, everything else is a rotation issue

    Take a page from Shikaru's book and put some actual substance behind your claims.

    Combing over your rotation a bit more, I also noticed that you're missing one cast of long bow 2 each time you're in berserk longbow. You should be doing swap from sword, lb 2, f1, filler stuff, f1, lb2 before the swap comes of cd with some variation when you land on lb after opening burst. Try to fix that in your rotation.

    @Cyninja.2954 said:
    https://dps.report/Nrn9-20191003-150603_golem
    Dps log of the benchmark video.

    0.17 Headbutt
    1.9 Berserk
    2.17 Dual Shot
    2.2 Scorched Earth
    2.5 Blood Reckonig
    2.85 Dual Shot
    2.9 Scorched Eartch
    3.2 Fan of Fire
    3.5 Weapon Swap

    https://dps.report/5LNK-20200527-190156_golem
    Your latest log.

    0.17 Sever Artery (not sure why you would not start on Headbutt)
    0.1 Weapon Swap (wasted weapon swap)
    0.45 Pin Down (probably fat fingered)
    1.65 Dual Shot
    1.69 Headbutt (you are already 1.5 second behind ideal rotation)
    3.45 Berserk
    3.74 Scorched Earth
    4.1 Blood Reckoning
    4.3 Scorched Earth
    4.73 Dual Shot
    5.0 Fan of Fire

    Look at the exact amount of skills in the ideal rotation. The simple rotation including auto attacks is almost a block, indicating that at nearly every available time frame a skill was cast or active(the fact that most skills have similar baseline cast times is beneficial here).

    Now take a look at your simple rotation. It's a string of different long lines of skills between weapon swaps. This indicates that your are either slower in executing/chaining skills, or miss clicking skills. Especially your sword rotation is sloppy compared to the benchmark.

    As far as infusions and the damage increase they bring, this was mentioned and Henry.5713 did some quick math for you. It's around 1-1.5k depending on build (power or condi) and if one is crit capped before (or condi duration capped) or not. Those 2 factors make up the biggest difference. Based on personal testing of myself, I can say that calculating 1k for infusions is correct. You can easily try this out by simply taking lower level food or reducing your setup by approximately 90 stat points (you could take off spotter for example for ease of testing, or add some effect worth around 90 beneficial stat points).

    I appreciate the feedback. I tightened up the rotation to match what is shown, but it's not accounting for the full amount of DPS expected from the napkin math people are talking over here.

    Also, @Cyninja.2954 , your log doesn't have your rotation as far as I can tell since the Player Summary section is empty.

    However, here's what I have after incorporating the feedback:

    Full log: https://dps.report/RI1Y-20200528-114041_golem

    Vid:

    DPS still peaks at 29k, which is still 4k below 33k.

  • Cyninja.2954Cyninja.2954 Member ✭✭✭✭
    edited May 28, 2020

    You should work an creating a full dps log, not stopping midway.

    It would both allow for you to directly see where you lose time, as well as what the total time is for you to kill the golem versus the benchmark. It also helps in creating muscle memory for extended periods of time.

    Also the logs I posted are not mine. They are 1 from the video, and the other from you. Worked fine just now.

  • Shikaru.7618Shikaru.7618 Member ✭✭✭

    This doesn't show up in the log, but you canceled your torch 4 with flaming flurry the first time you're in sword. You can see it happen at 0:06 in your video.

  • frqnbbsov.9806frqnbbsov.9806 Member ✭✭
    edited June 4, 2020

    I think you have problem in golem settings, in standart you need to add all conditions on golem, first time I had this mistake and I said why my dps low, golem settings very important. Average golem with 4000000 and all conditions.Some skills damage associate with conditions on golem so you not added all.

  • Besetment.9187Besetment.9187 Member ✭✭
    edited June 7, 2020

    [SC] are a bunch of robots with insanely fast and accurate fingers. Luna reaches the first weapon swap in 3.5 seconds. You do it in 5.7 seconds. At the next weapon swap, Luna squeezes in an extra hamstring, sever artery, gash and hamstring before weapon swapping and Luna hits the second weapon swap faster than you did. You appear to have interrupted your first sword auto chain and are completely missing the second auto chain and you are still behind on time.

    If you want to hit [SC] numbers good luck but it helps if you are young, do not suffer from any type of chronic wrist injury and were a guitar hero world champion in a previous life.

  • Dahkeus.8243Dahkeus.8243 Member ✭✭✭✭

    Yea, the more I've practiced, the more I've realized that there are just some things going on with rotation, particularly with auto attacks that made the difference. I was able to fit in all the big skills in the rotation in about the same amount of time, but the benchmark vid was set with someone who was able to get in all of the big skills so efficiently that they were able to get in more auto attacks without having to delay weapon swaps or things like that. It still frustrates me that this makes so much of a difference, but I do have big respect for the people that pull this off.

  • Besetment.9187Besetment.9187 Member ✭✭
    edited June 7, 2020

    Right. The order of skills matter and the time between them also matters. You can queue actions by pressing a skill button slightly before the previous one finishes. When starting out, you need to watch the fill bar, get the rhythm down, then drill it until it becomes muscle memory and you don't need to look at the fill bar anymore.

    If you queue actions perfectly, there should be no auto 1s between your skill buttons. If you open up your simple rotation and see an auto 1 in between big skills, this means your rhythm is slightly off. Because nobody has perfect metronomic rhythm, this is why you see buttons getting mashed in benchmark videos to ensure they get queued with no gaps in between.

    I'm learning Condi Firebrand right now and of course I'm using the [SC] benchmark and arcDPS log to identify areas I can improve. After memorizing the rotation, my first logged run ended up being 25k/s average. The [SC] benchmark was 31k/s. I let 64 scepter autos through. The [SC] benchmark let zero through. The skill queueing and zealot's fire timings were so good, they did not let a single scepter auto attack go off. Honestly, I'm not convinced that some of those guys are even human.

    The good news for you is that if you can chill at 27-29k/s on a golem, thats good enough for pretty much everything in the game except high end speed clears.

  • Hannelore.8153Hannelore.8153 Member ✭✭✭✭
    edited June 7, 2020

    If you're trying to hit these benchmarks in regular gameplay I ask that you please wear wrist braces or you'll regret it later. Its exteremely hard on your fingers, hands and wrists, and is mostly done for experimental proof of concept reasons and can cause permanent damage over just a few years. And then you'll struggle to do any demanding tasks in games at all.

    Consider that a short term boon becomes a long-term liability.

    There's no content in the game that needs those DPS numbers to clear, not even raids. A guild used to clear raids in Masterwork (green) gear by doing hardcore rotations so imagine what level of effort you actually need in Exotic/Ascended.

    Hannah | Daisuki[SUKI] Founder, Ehmry Bay, NA | 22 charas, 15k hours, ~27k AP | ♀♥♀
    Mains Mariyuuna/Tempest & Alisha Kei/Druid(PvE), Terakura/Spellbreaker & Kitty Koume/Reaper(WvW)
    The fun of being a link is tagging up and hearing "who are you?"

  • Yep. I have tendinitis and have had it since I was 32. Mashing buttons in games like Diablo 3 and decades of playing guitar with ropey technique really took its toll. (It isn't worth it).

  • Dahkeus.8243Dahkeus.8243 Member ✭✭✭✭

    Well, it's not really the act of hitting skills early to queue them or to hit them quickly enough that has been the challenge , but the toughest thing is really knowing exactly when to press to cast certain skills since it's so easy to cancel-cast certain skills if you hit the next one too early as well as to waste time if you don't do so early enough.

  • Kulvar.1239Kulvar.1239 Member ✭✭✭

    @RaidsAreEasyAF.8652 said:
    They are definitely not worth 6k dps.
    your rotation just isn't super clean, you start the fight with an auto attack and on your first shattering blow you are already 2 seconds behind the benchmark. Speed is important for a rotation and you can already lack behind 2 seconds in the first 3 seconds of a log. That's a huge time loss.

    Was only looking at the log. In the video you also don't reset your burst skill with your healing skill.

    I wonder which class build is less impacted by network lag.

  • @Kulvar.1239 said:
    I wonder which class build is less impacted by network lag.

    I guess that depends on on what level you want to play a certain class. If you wanna squeeze out even a tiny bit more DPS, every class is impacted by it.

  • Astralporing.1957Astralporing.1957 Member ✭✭✭✭
    edited June 8, 2020

    @Dahkeus.8243 said:
    It still frustrates me that this makes so much of a difference

    It's actually one of the problems devs have to live with now. There's a ton of theoretically small, hard to notice (or hard to reproduce) things that have a surprisingly big impact on dps effectiveness. It's one of the factors for why the dps disparity between top and average players is so massive in this game.

    The whole point of a social game is to play with the people you want to play with, not be forced to play with the people you don't.

  • Kulvar.1239Kulvar.1239 Member ✭✭✭

    @RaidsAreEasyAF.8652 said:

    @Kulvar.1239 said:
    I wonder which class build is less impacted by network lag.

    I guess that depends on on what level you want to play a certain class. If you wanna squeeze out even a tiny bit more DPS, every class is impacted by it.

    I was wondering which is less impacted, not which is not impacted.

  • LadyKitty.6120LadyKitty.6120 Member ✭✭✭

    @Astralporing.1957 said:

    @Dahkeus.8243 said:
    It still frustrates me that this makes so much of a difference

    It's actually one of the problems devs have to live with now. There's a ton of theoretically small, hard to notice (or hard to reproduce) things that have a surprisingly big impact on dps effectiveness. It's one of the factors for why the dps disparity between top and average players is so massive in this game.

    And it's the combination of bad build choice on top of bad rotation further multiplying the difference. One of the worst examples might be power sword berserker. By itself it already has half of axe+axe's damage with optimized zerker gears. If that player takes warhorn as off-hand, that's another 30% off easily. If they run survival traits in Strength, another 30% off. Taking Tactics over Discipline with survival traits is also about 20% off. And then, if they use lvl 80 boost Soldier's gears, that's more than 50% off. And bad rotation is another 50%. Result? Top-tier metaplayer does 20x the damage of that random sword+warhorn/shield wannabe-Boromir (and Kitty's not trying to insult anyone here, Boromir just came to mind when Kitty had already written those percentiles and weapon choices. That example is theorethical but Kitty's seen similar stuff happening a lot in-game build/numbers-wise.)

    It's Kitty. The young lady who streams and records videos playing various (non-)metabuilds. Raid/fractal videos at youtube.com/LadyKitty, Kittymarks test results at youtube.com/Kittymarks and tinyurl.com/Kittymarks.

  • Noodle Ant.1605Noodle Ant.1605 Member ✭✭✭
    edited June 8, 2020

    @Kulvar.1239 said:
    I wonder which class build is less impacted by network lag.

    There’s no definite answer afaik, but I assume it’s classes that rely the least on:

    • combination(s) of different low cast time skills pressed in quick succession. In between each cast, I assume there is a minuscule delay from lag, which adds up to massive amounts of time wasted and hence why benchers with optimal latency can fit x more auto chains in a rotation/finishes rotation in x amount of time
    • skills with abnormally long aftercasts (or just ‘aftercast management’). I have a feeling that high latency goes into more of the aftercast vs optimal latency even when queuing skills. Or the skill gets cancelled and completely wasted if you queue the next skill too quickly. This is probably why it’s difficult to determine a lag-friendly class at first glance even if its rotation is ‘simple’/‘simpler’

    Sometimes the rotation needs to be altered to get higher numbers.

    Power staff weaver (when it used to be meta) and pre-rework condi (banner?) berserker at some point of time used to be lag-friendly specs (or at least ppl w/ ~300 ping managed to get within 1-2k difference). Mainly just power staff weaver tho.

    pChrono (main), cWeaver and pReaper. An asura who likes snowflakes.