You've still completely missed the mark on DoT damage.

LiquidPony

I want to start by quoting something that I think highlights why you've missed the mark, again, on DoTs:

In a future patch during this PTS cycle, we will be increasing the total damage of any Damage over Time abilities that were adjusted in PTS v5.2.0 by approximately 20% to reduce situations where many of the abilities felt difficult to maintain outside of nanosecond perfect rotation. This means there will be an overall reduction in damage from DoT abilities of approximately 33% compared to what is currently on the Live megaservers.

In response to much of your feedback, we have increased the standards of our Damage over Time abilities by 20% compared to the original pass in PTS v5.2.0 to better encompass the fact that there are many “dead-zones” of time where these abilities fall off before you can reapply them.

Ignoring the fact that your percentage math makes no sense at all, I'd like to pick on this implication that the only reason DoTs are underperforming is that people aren't skilled enough to have "nanosecond perfect rotations". That's hogwash. The reason that DoTs are underperforming is because you over-nerfed them.

"Nanosecond perfect rotations" are not the crux of the issue; that's just one small part. You're missing the point. Your DoT values are simply too low.

Here's a simple demonstration. I've created a PTS character and tried to reduce variance as much as possible. 100% crit rate, no glyphs or poisons, no proc sets.

Observation: DoTs suck
Hypothesis: On a stamblade, Rending Slashes is not worth using and I'm better off spamming Surprise Attack again
Experiment: Four setups, all using the same gear (5 x Leviathan, 5 x Hunding's Rage, 2 x NMG; Divines armor w/ Stam enchants; Infused jewelry w/ Weapon Damage enchants; Precise weapons; Thief Mundus; Lavafoot food; Weapon Power pots; CP evenly distributed in Thaum/MaA except in the final test).
1. Spam LA + Surprise Attack 10 times
2. LA + Rending Slashes then LA + Surprise Attack 9 times
3. Replace Rending Slashes with a Fighters Guild ability, repeat test #1
4. Optimize CP by taking it out of Thaum, repeat test #1
5. With the CP from test #4, test with Rending Slashes again

Results
Test 1 setup


Test 1


Test 2 setup


Test 2


Test 3 setup


Test 3


Test 4 setup


Test 4


Test 5 setup


Test 5


Analysis: In the most naive possible test, using Rending Slashes does result in slightly higher DPS (~3.12%). However, using more realistic test scenarios such as replacing Rending Slashes with a skill that grants a passive stat increase and optimizing CP around Direct Damage, higher DPS is achieved by simply spamming one skill. Note: because it's impossible to ensure that parses are all exactly the same duration, it's important to look at the total damage done and mean damage of each cast rather than simply comparing DPS.

Conclusion: Your DoT values are still wrong and it's not hard to see. I don't know what the correct DoT to spammable ratio should be, but what you've got now is wrong. It is not sufficient for DoTs to be marginally stronger than spammables; they need to be significantly stronger. Because, it's not just about "nanosecond perfect rotations". You also have to consider opportunity costs: I can simply slot some skill that grants a passive increase (Fighters Guild, Mages Guild, Animal Companions, Assassination, Grave Lord, etc), I can better optimize CP if I reduce my DoT damage, there are fights with phases where DoTs simply stop working, etc.

Zalathorm

LiquidPony wrote: »

Conclusion: Your DoT values are still wrong and it's not hard to see. I don't know what the correct DoT to spammable ratio should be, but what you've got now is wrong. It is not sufficient for DoTs to be marginally stronger than spammables; they need to be significantly stronger. Because, it's not just about "nanosecond perfect rotations". You also have to consider opportunity costs: I can simply slot some skill that grants a passive increase (Fighters Guild, Mages Guild, Animal Companions, Assassination, Grave Lord, etc), I can better optimize CP if I reduce my DoT damage, there are fights with phases where DoTs simply stop working, etc.

Thank you for your thorough testing.

With this kind of evidence, it seems imperative that PTS is extended to allow for more testing and responsive number tweaking.

The only way to find what the right DoT to spammable ratio should be is through rigorous testing and feedback from players willing to do it. Clearly if ZOS knew how to assess the issue, it would have been fixed properly from patch 5.0.0.

Drako_Ei

If dots are not worth casting... imagine aoe

MartiniDaniels

Zalathorm wrote: »

LiquidPony wrote: »

Conclusion: Your DoT values are still wrong and it's not hard to see. I don't know what the correct DoT to spammable ratio should be, but what you've got now is wrong. It is not sufficient for DoTs to be marginally stronger than spammables; they need to be significantly stronger. Because, it's not just about "nanosecond perfect rotations". You also have to consider opportunity costs: I can simply slot some skill that grants a passive increase (Fighters Guild, Mages Guild, Animal Companions, Assassination, Grave Lord, etc), I can better optimize CP if I reduce my DoT damage, there are fights with phases where DoTs simply stop working, etc.

Thank you for your thorough testing.

With this kind of evidence, it seems imperative that PTS is extended to allow for more testing and responsive number tweaking.

The only way to find what the right DoT to spammable ratio should be is through rigorous testing and feedback from players willing to do it. Clearly if ZOS knew how to assess the issue, it would have been fixed properly from patch 5.0.0.

There is one very easy solution - roll back ALL dots (ST, aoe, class, generic) to U22 and start from that point. From PVE perspective U22 was overall pretty balanced, only magblade and magden and magcro were notably under performing, others were close to each other with stamina slightly forward which is ok given that in many encounters it's harder to be in melee.

SeaUnicorn

I did the test bot forgot to screenshot yesterday. On a Magplar with 56 in direct damage and 56 in thaum I get same dps with 7 dots, 6 dots and spammable and so on down to 3 dots and spammable. However, more dots I add - harder it is to sustain. For the purpose of the test I gradually replaced dots with various utility skills like inner light and rune focus.

Interestingly enough swapping dots back and forth doesn't change the outcome.

I do think damage-wise it's balanced because well you get to play any build you want and pull the numbers, however there are 2 things besides DPS that make DOTs inferior:

- DOTs are harder to sustain, more you add, worse it gets.
- DOTs are all different durations, making DOT builds annoying whack-a-mole mini game. At least make class dots match in duration. For example solar barrage is always an odd kid on the block. Don't understand why.

Finally overall damage went down by 13% for me and LA damage is even more prevalent source of damage now, which increases the vast gap between skilled and new players.

The damage balancing needs to be addressed properly such that new players don't get punished more for the high damage of the select few leaders who took build optimization to a whole new level.

kojou

They will probably fix by lowering passive damage...

relentless_turnip

Zalathorm wrote: »

LiquidPony wrote: »

Conclusion: Your DoT values are still wrong and it's not hard to see. I don't know what the correct DoT to spammable ratio should be, but what you've got now is wrong. It is not sufficient for DoTs to be marginally stronger than spammables; they need to be significantly stronger. Because, it's not just about "nanosecond perfect rotations". You also have to consider opportunity costs: I can simply slot some skill that grants a passive increase (Fighters Guild, Mages Guild, Animal Companions, Assassination, Grave Lord, etc), I can better optimize CP if I reduce my DoT damage, there are fights with phases where DoTs simply stop working, etc.

Thank you for your thorough testing.

With this kind of evidence, it seems imperative that PTS is extended to allow for more testing and responsive number tweaking.

The only way to find what the right DoT to spammable ratio should be is through rigorous testing and feedback from players willing to do it. Clearly if ZOS knew how to assess the issue, it would have been fixed properly from patch 5.0.0.

I would also like to thank you for your near obsessive testing

I think they should just take DOT's back to the original figure's (scalebreaker) and subtract 33% as they said they were intending to in 5.2.2

Every thread i have read is on more than on board with a -33% on live DOT's.

Instead we have +20% on already nerfed value... you don't need to be an accountant to realise that these math's will not equate to your 'intentions'.

I hope they revisit this...

Davadin

LiquidPony wrote: »

Conclusion: ...is not sufficient for DoTs to be marginally stronger than spammables; they need to be significantly stronger.

disagree.

and this is coming from a stamDK with 56-points in Thaum.

i want both "stack-and-forgot" and burst combo both be viables.

so looking at individual skills, i want DoT that ticks marginally weaker than spammable, because they are inherently less risk.

LiquidPony

Davadin wrote: »

LiquidPony wrote: »

Conclusion: ...is not sufficient for DoTs to be marginally stronger than spammables; they need to be significantly stronger.

disagree.

and this is coming from a stamDK with 56-points in Thaum.

i want both "stack-and-forgot" and burst combo both be viables.

so looking at individual skills, i want DoT that ticks marginally weaker than spammable, because they are inherently less risk.

You're free to disagree but you're wrong.

There's simply no point in using DoTs if they are not significantly stronger than spammables. Simply compare test #2 to test #4 above.

As it stands right now, the only DoTs worth using are the occasional class DoT or very long duration DoT, Mystic Orb, and ground-targeted DoTs that are used to keep backbar enchants procced.

Stuff like Rending Slashes, Carve, Soul Trap, and Poison Injection are still trash skills that aren't worth running in PvE.

fred4

Listen to the OP! Great stuff!

SodanTok

I miss 2 important tests here that would actually be telling the story better. One with precise CP distribution (56+56 when you do test where DoT is only 6.7% is very skewed) and one where you go all in on Master-at-Arms but you still use the DoT in rotation.

If not for sake of being correct then for sake of not showcasing data that can be misinterpreted. Like Test3 that shows 700 difference gain over Test2 even tho total damage dealt was 5k less

Taloros

Another way to make DoTs viable is to give them a similar damage than direct damage attacks, but reduce their cost significantly. Then, they'd have a use as a cost-friendly alternative for improving sustain.

Increasing damage would still be better, though. Sustain isn't an issue in many fights.

LiquidPony

SodanTok wrote: »

I miss 2 important tests here that would actually be telling the story better. One with precise CP distribution (56+56 when you do test where DoT is only 6.7% is very skewed) and one where you go all in on Master-at-Arms but you still use the DoT in rotation.

If not for sake of being correct then for sake of not showcasing data that can be misinterpreted. Like Test3 that shows 700 difference gain over Test2 even tho total damage dealt was 5k less

What is "precise CP distribution"? If you do full parses on a DD-focused setup like a stamblade, and use Constellations to optimize CP, it will eventually optimize virtually every point out of Thaumaturge. Probably about 20 in Thaum.

Putting 56 points in Thaumaturge was being extremely generous to Rending Slashes in the first place.

And DPS is ... damage divided by time. There's no way to ensure that each parse takes exactly the same amount of time to complete. You can simply look at the actual parse values and extrapolate. Yes, parse #2 has more total damage than parse #3 (and if you look at the mean values of all the skills, that makes perfect sense; Rending Slashes in parse #2 is very slightly stronger than skipping it altogether in parse #3).

Parse #4 shows that optimizing towards Direct Damage means that you're better off just spamming one skill than incorporating weak DoTs.

I will add a note to make sure that people are paying attention to total damage done and the actual mean damage of each cast in CMX, and also do another test with the loaded MaA CP while also using Rending Slashes. Although this is going to be more skewed in favor of Rending Slashes than I would like since, at this point, something like 33% of its total damage is Direct Damage. It might be more appropriate to do another set of tests using Soul Trap.

Jeezye

Very nice analysis, finally a post with substance. I think technically from your numbers dots and direct damage seem in balance. Also consider dots are used in PvP.

What I dont understand however is the inconsistency between uniquly higher damage of certain dots and their utility. In that regard, cripple is a total waste of a bar space, e.g. compared to sun fire.

Bladerunner1

Test 1 hit 201k total damage
Test 2 - 209.5k total damage
Test 3 - 203.6k total damage
Test 4 - 222.2k total damage

The difference between test 2&3 is 2.8%, and it would have been a whole lot closer on PTS 5.2.0, like maybe 1%.

And test 4 has 100% direct damage. Direct damage is kind of a Nightblade-leaning thing, and physical weapon expert is buffing a huge portion of the parse compared to what you'd get on a normal DPS test.

You could just as easily slot an axe, use flurry as a spammable and go the other way with Thaumaturge, especially with oher classes that lean on their class DOTs. Some of the dots would be a lot harder to pass up on those builds, which could be cast aside on direct damage builds. All I know is my Nightblade is hitting much closer to my DK on the PTS, still not there yet, and each is using opposite damage styles instead of going one-sided on Thaumaturge like it is on live.

SodanTok

LiquidPony wrote: »

SodanTok wrote: »

I miss 2 important tests here that would actually be telling the story better. One with precise CP distribution (56+56 when you do test where DoT is only 6.7% is very skewed) and one where you go all in on Master-at-Arms but you still use the DoT in rotation.

If not for sake of being correct then for sake of not showcasing data that can be misinterpreted. Like Test3 that shows 700 difference gain over Test2 even tho total damage dealt was 5k less

What is "precise CP distribution"? If you do full parses on a DD-focused setup like a stamblade, and use Constellations to optimize CP, it will eventually optimize virtually every point out of Thaumaturge. Maybe 3-5% or something at most.

Putting 56 points in Thaumaturge was being extremely generous to Rending Slashes in the first place.

And DPS is ... damage divided by time. There's no way to ensure that each parse takes exactly the same amount of time to complete. You can simply look at the actual parse values and extrapolate. Yes, parse #2 has more total damage than parse #3 (and if you look at the mean values of all the skills, that makes perfect sense; Rending Slashes in parse #2 is very slightly stronger than skipping it altogether in parse #3).

Parse #4 shows that optimizing towards Direct Damage means that you're better off just spamming one skill than incorporating weak DoTs.

I will add a note to make sure that people are paying attention to total damage done and the actual mean damage of each cast in CMX, and also do another test with the loaded MaA CP while also using Rending Slashes.

I meant that 56/56 distribution is ideal in condition where DoT and Direct portion of DPS are equal. Speccing into Direct is gonna result in more DPS everytime if DoT portion of parse isnt higher.
So Parse 4 shows that optimizing CP around damage dealt results in higher DPS. Nothing more than that. It would result in DPS increase even with Rending being used (by replacing the one fighters guild skill).

As for Constellation, are you using manual input instead of parse import? I have doubts about the 'it will eventually optimize virtually every point out of Thaumaturge' since just poisons with relequen, hail, trap, DW DoT and possible maarselok will put your DoT portion in parse above 30%. Something parse import in constellation is being incorrect about.

LiquidPony

Bladerunner1 wrote: »

Test 1 hit 201k total damage
Test 2 - 209.5k total damage
Test 3 - 203.6k total damage
Test 4 - 222.2k total damage

The difference between test 2&3 is 2.8%, and it would have been a whole lot closer on PTS 5.2.0, like maybe 1%.

And test 4 has 100% direct damage. Direct damage is kind of a Nightblade-leaning thing, and physical weapon expert is buffing a huge portion of the parse compared to what you'd get on a normal DPS test.

You could just as easily slot an axe, use flurry as a spammable and go the other way with Thaumaturge, especially with oher classes that lean on their class DOTs. Some of the dots would be a lot harder to pass up on those builds, which could be cast aside on direct damage builds. All I know is my Nightblade is hitting much closer to my DK on the PTS, still not there yet, and each is using opposite damage styles instead of going one-sided on Thaumaturge like it is on live.

Except that in most cases there is no point in using Flurry; you'll be better served by switching to 2H and using Wrecking Blow/Executioner so that you can continue optimizing CP towards Direct Damage.

vMA DW is still garbage, so there's no benefit to going Flurry to enable that. Except *maybe* on a stamDK. Every other class is going to want to use their class spammable or Wrecking Blow + Executioner. And even on a stamDK ... if you're the only one you'll probably be using Stone Fist for the Stagger proc, so even that's a scratch.

Do a bunch of full parses on the raid dummy, with DoTs, and then let Constellations optimize your CP for you. Eventually you'll drop most DoTs and move most of your points out of Thaumaturge.

Joy_Division

This is what happens when ZOS tries to fix something that was never a problem.

For 4 1/2 years, I dont recall any serious complaints regarding DoT damage, for PvE or PvP. There were only some annoyances that some of them, most notably Engulfing Flames and Spear Shards, the damage was lackluster. But for the most part, class DoTs were fine.

The ZOS had to fix them by standardizing all of the to follow the same formula (hooray for diversity!), removing the first tick on many of them (which makes it even harder to burst players down), which skewed things enough that they felt compelled to put a massive increase on all their damage (even though they were told it was too radical a change), and now with every PTS patch they're just throwing on massive nerfs and buffs to try and recapture the sweet spot the game already had before they tried to "fix" things. With all these changes, some abilities like Poison injection have had their mechanics altered and thus whatever % change ZOS decides upon will still result in a nerf from what it used to be able to do.

This is like a dog chasing it's tail. I know some people feel that with some of the buffs that ZOS is listening, but this sort of development cycle is not something that I think should be applauded.

WrathOfInnos

Great post, and results agree with the testing I’ve been doing on Magicka DPS builds. Basically it comes down to the essentials: Spammables, Wall (for the vMA effect and enchant procs), and Barbed Trap (works better than Accelerate for Minor Force on Mag builds, which is stupid IMO). Mystic Orb also seems worth running on most Magicka builds, although it is annoying to use with Trap because Orb is best from 8m away and Trap requires stepping into melee distance. Beyond these 3 skills, any DoTs I’ve tried to add to the build ends up being a DPS loss. The following are a list of skills that can be used if you want lower damage:
- Entropy
- Soul Trap
- Scalding Rune
- Destructive Reach
- Liquid Lightning/Lightning Flood

Without specifically testing, I’m going to guess that this also extends to many other class DoTs like:
- Twisting Path
- Cripple
- Vamp Bane/Reflective
- Eruption
- Boneyard

And there is definitely some funny math going on here. For example, Degeneration on live (Scalebreaker) deals about 3.3X the damage of one spammable over 12s. If this was reduced by 33% as mentioned in the 3.2.2 patch notes you quoted then it would be doing 2.2X the damage of one spammable, and would still be marginally useful (might even slot it if it kept the Magicka return on Light Attacks). However, it’s power budget was not held to this claim of a 33% reduction, and instead it was nerfed to less than 2X one spammable (a little over 40% reduction), and tests have proven it is worthless now. And this is a generous example, since it is one of the strongest single target DoTs, which also gives Empower on every cast and has relatively good passives just for slotting it.

The new DoT standard is clearly not high enough. I can respect the gradual increase of 20% to attempt to find the correct value, and avoid massively overshooting again, but I believe another increase is needed in 5.2.4. I’m guessing that another 20% would be about right. That would bring DoTs to around 28% weaker than they are on live, or 44% stronger than in PTS 5.2.0. At a minimum they need to get a 12% increase to reach 67% of live values and 34% higher than 5.2.0 (might make DoTs DPS neutral, but they probably wouldn’t be a gain).

LiquidPony

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

I miss 2 important tests here that would actually be telling the story better. One with precise CP distribution (56+56 when you do test where DoT is only 6.7% is very skewed) and one where you go all in on Master-at-Arms but you still use the DoT in rotation.

If not for sake of being correct then for sake of not showcasing data that can be misinterpreted. Like Test3 that shows 700 difference gain over Test2 even tho total damage dealt was 5k less

What is "precise CP distribution"? If you do full parses on a DD-focused setup like a stamblade, and use Constellations to optimize CP, it will eventually optimize virtually every point out of Thaumaturge. Maybe 3-5% or something at most.

Putting 56 points in Thaumaturge was being extremely generous to Rending Slashes in the first place.

And DPS is ... damage divided by time. There's no way to ensure that each parse takes exactly the same amount of time to complete. You can simply look at the actual parse values and extrapolate. Yes, parse #2 has more total damage than parse #3 (and if you look at the mean values of all the skills, that makes perfect sense; Rending Slashes in parse #2 is very slightly stronger than skipping it altogether in parse #3).

Parse #4 shows that optimizing towards Direct Damage means that you're better off just spamming one skill than incorporating weak DoTs.

I will add a note to make sure that people are paying attention to total damage done and the actual mean damage of each cast in CMX, and also do another test with the loaded MaA CP while also using Rending Slashes.

I meant that 56/56 distribution is ideal in condition where DoT and Direct portion of DPS are equal. Speccing into Direct is gonna result in more DPS everytime if DoT portion of parse isnt higher.
So Parse 4 shows that optimizing CP around damage dealt results in higher DPS. Nothing more than that. It would result in DPS increase even with Rending being used (by replacing the one fighters guild skill).

As for Constellation, are you using manual input instead of parse import? I have doubts about the 'it will eventually optimize virtually every point out of Thaumaturge' since just poisons with relequen, hail, trap, DW DoT and possible maarselok will put your DoT portion in parse above 30%. Something parse import in constellation is being incorrect about.

If your DoT damage is 30% of your total damage, and Direct Damage is 60%, and your LA damage is 10%, you're going to end up putting the vast majority of your points in Master at Arms. About a 66/20 split IIRC. And then you'll do another parse and find that your ratio has changed and it will recommend moving a few more points, and so on.

LiquidPony

Joy_Division wrote: »

This is what happens when ZOS tries to fix something that was never a problem.

For 4 1/2 years, I dont recall any serious complaints regarding DoT damage, for PvE or PvP. There were only some annoyances that some of them, most notably Engulfing Flames and Spear Shards, the damage was lackluster. But for the most part, class DoTs were fine.

The ZOS had to fix them by standardizing all of the to follow the same formula (hooray for diversity!), removing the first tick on many of them (which makes it even harder to burst players down), which skewed things enough that they felt compelled to put a massive increase on all their damage (even though they were told it was too radical a change), and now with every PTS patch they're just throwing on massive nerfs and buffs to try and recapture the sweet spot the game already had before they tried to "fix" things. With all these changes, some abilities like Poison injection have had their mechanics altered and thus whatever % change ZOS decides upon will still result in a nerf from what it used to be able to do.

This is like a dog chasing it's tail. I know some people feel that with some of the buffs that ZOS is listening, but this sort of development cycle is not something that I think should be applauded.

Completely agree.

I see people doing the "GG ZOS, nice start" song-and-dance routine in v5.2.3 and I think it's ludicrous.

This patch has been a balance travesty from the start; it shouldn't have taken us 4 weeks to get to the point where DoTs are almost worth using again.

Nothing about this PTS dev cycle should be applauded. This patch has sucked from day 1 and it continues to do so.

SodanTok

LiquidPony wrote: »

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

I miss 2 important tests here that would actually be telling the story better. One with precise CP distribution (56+56 when you do test where DoT is only 6.7% is very skewed) and one where you go all in on Master-at-Arms but you still use the DoT in rotation.

If not for sake of being correct then for sake of not showcasing data that can be misinterpreted. Like Test3 that shows 700 difference gain over Test2 even tho total damage dealt was 5k less

What is "precise CP distribution"? If you do full parses on a DD-focused setup like a stamblade, and use Constellations to optimize CP, it will eventually optimize virtually every point out of Thaumaturge. Maybe 3-5% or something at most.

Putting 56 points in Thaumaturge was being extremely generous to Rending Slashes in the first place.

And DPS is ... damage divided by time. There's no way to ensure that each parse takes exactly the same amount of time to complete. You can simply look at the actual parse values and extrapolate. Yes, parse #2 has more total damage than parse #3 (and if you look at the mean values of all the skills, that makes perfect sense; Rending Slashes in parse #2 is very slightly stronger than skipping it altogether in parse #3).

Parse #4 shows that optimizing towards Direct Damage means that you're better off just spamming one skill than incorporating weak DoTs.

I will add a note to make sure that people are paying attention to total damage done and the actual mean damage of each cast in CMX, and also do another test with the loaded MaA CP while also using Rending Slashes.

I meant that 56/56 distribution is ideal in condition where DoT and Direct portion of DPS are equal. Speccing into Direct is gonna result in more DPS everytime if DoT portion of parse isnt higher.
So Parse 4 shows that optimizing CP around damage dealt results in higher DPS. Nothing more than that. It would result in DPS increase even with Rending being used (by replacing the one fighters guild skill).

As for Constellation, are you using manual input instead of parse import? I have doubts about the 'it will eventually optimize virtually every point out of Thaumaturge' since just poisons with relequen, hail, trap, DW DoT and possible maarselok will put your DoT portion in parse above 30%. Something parse import in constellation is being incorrect about.

If your DoT damage is 30% of your total damage, and Direct Damage is 60%, and your LA damage is 10%, you're going to end up putting the vast majority of your points in Master at Arms. About a 66/20 split IIRC. And then you'll do another parse and find that your ratio has changed and it will recommend moving a few more points, and so on.

That seems like bug. Using manual calculation of ratios my constellation seems to stabilize within one parse and it shouldnt move after that because its supposed to account for used CP distributon. Or better said, the calculation made on empty (without CP) parse will be nearly set in stone and have no reason to be changing. Getting anything less than 20 in Thaum should not be iany way feasible, not with guaranteed 1/3 of dps coming from DoTs no matter what.

WrathOfInnos

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

I miss 2 important tests here that would actually be telling the story better. One with precise CP distribution (56+56 when you do test where DoT is only 6.7% is very skewed) and one where you go all in on Master-at-Arms but you still use the DoT in rotation.

If not for sake of being correct then for sake of not showcasing data that can be misinterpreted. Like Test3 that shows 700 difference gain over Test2 even tho total damage dealt was 5k less

What is "precise CP distribution"? If you do full parses on a DD-focused setup like a stamblade, and use Constellations to optimize CP, it will eventually optimize virtually every point out of Thaumaturge. Maybe 3-5% or something at most.

Putting 56 points in Thaumaturge was being extremely generous to Rending Slashes in the first place.

And DPS is ... damage divided by time. There's no way to ensure that each parse takes exactly the same amount of time to complete. You can simply look at the actual parse values and extrapolate. Yes, parse #2 has more total damage than parse #3 (and if you look at the mean values of all the skills, that makes perfect sense; Rending Slashes in parse #2 is very slightly stronger than skipping it altogether in parse #3).

Parse #4 shows that optimizing towards Direct Damage means that you're better off just spamming one skill than incorporating weak DoTs.

I will add a note to make sure that people are paying attention to total damage done and the actual mean damage of each cast in CMX, and also do another test with the loaded MaA CP while also using Rending Slashes.

I meant that 56/56 distribution is ideal in condition where DoT and Direct portion of DPS are equal. Speccing into Direct is gonna result in more DPS everytime if DoT portion of parse isnt higher.
So Parse 4 shows that optimizing CP around damage dealt results in higher DPS. Nothing more than that. It would result in DPS increase even with Rending being used (by replacing the one fighters guild skill).

As for Constellation, are you using manual input instead of parse import? I have doubts about the 'it will eventually optimize virtually every point out of Thaumaturge' since just poisons with relequen, hail, trap, DW DoT and possible maarselok will put your DoT portion in parse above 30%. Something parse import in constellation is being incorrect about.

If your DoT damage is 30% of your total damage, and Direct Damage is 60%, and your LA damage is 10%, you're going to end up putting the vast majority of your points in Master at Arms. About a 66/20 split IIRC. And then you'll do another parse and find that your ratio has changed and it will recommend moving a few more points, and so on.

That seems like bug. Using manual calculation of ratios my constellation seems to stabilize within one parse and it shouldnt move after that because its supposed to account for used CP distributon. Or better said, the calculation made on empty (without CP) parse will be nearly set in stone and have no reason to be changing. Getting anything less than 20 in Thaum should not be iany way feasible, not with guaranteed 1/3 of dps coming from DoTs no matter what.

If 1/3 of damage is DoT and 2/3 is direct, then each 1% increase from Master at Arms is twice as valuable as a 1% increase from Thaumaturge. The jump from 20 to 28 Thaum is 3% DoT damage, or equivalent to 1% average damage for 8 points invested, so 0.125% gain per CP. The jump from 56 Master to 66 Master is 2% direct damage or 1.33% average damage for 10 points invested, yielding 0.133% gain per CP. The next jump in Master would be 72 for another 1% direct damage or 0.67% average damage costing 6 points or 0.11% per CP. So I agree that 66/20 is the correct split if it is a 1/3 DoT 2/3 direct parse.

Edit: If it’s a 70/30 split between direct and DoT then it’s actually worth going up to 72 Master at Arms before investing more than 20 in Thaumaturge.

SodanTok

WrathOfInnos wrote: »

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

I miss 2 important tests here that would actually be telling the story better. One with precise CP distribution (56+56 when you do test where DoT is only 6.7% is very skewed) and one where you go all in on Master-at-Arms but you still use the DoT in rotation.

If not for sake of being correct then for sake of not showcasing data that can be misinterpreted. Like Test3 that shows 700 difference gain over Test2 even tho total damage dealt was 5k less

What is "precise CP distribution"? If you do full parses on a DD-focused setup like a stamblade, and use Constellations to optimize CP, it will eventually optimize virtually every point out of Thaumaturge. Maybe 3-5% or something at most.

Putting 56 points in Thaumaturge was being extremely generous to Rending Slashes in the first place.

And DPS is ... damage divided by time. There's no way to ensure that each parse takes exactly the same amount of time to complete. You can simply look at the actual parse values and extrapolate. Yes, parse #2 has more total damage than parse #3 (and if you look at the mean values of all the skills, that makes perfect sense; Rending Slashes in parse #2 is very slightly stronger than skipping it altogether in parse #3).

Parse #4 shows that optimizing towards Direct Damage means that you're better off just spamming one skill than incorporating weak DoTs.

I will add a note to make sure that people are paying attention to total damage done and the actual mean damage of each cast in CMX, and also do another test with the loaded MaA CP while also using Rending Slashes.

I meant that 56/56 distribution is ideal in condition where DoT and Direct portion of DPS are equal. Speccing into Direct is gonna result in more DPS everytime if DoT portion of parse isnt higher.
So Parse 4 shows that optimizing CP around damage dealt results in higher DPS. Nothing more than that. It would result in DPS increase even with Rending being used (by replacing the one fighters guild skill).

As for Constellation, are you using manual input instead of parse import? I have doubts about the 'it will eventually optimize virtually every point out of Thaumaturge' since just poisons with relequen, hail, trap, DW DoT and possible maarselok will put your DoT portion in parse above 30%. Something parse import in constellation is being incorrect about.

If your DoT damage is 30% of your total damage, and Direct Damage is 60%, and your LA damage is 10%, you're going to end up putting the vast majority of your points in Master at Arms. About a 66/20 split IIRC. And then you'll do another parse and find that your ratio has changed and it will recommend moving a few more points, and so on.

That seems like bug. Using manual calculation of ratios my constellation seems to stabilize within one parse and it shouldnt move after that because its supposed to account for used CP distributon. Or better said, the calculation made on empty (without CP) parse will be nearly set in stone and have no reason to be changing. Getting anything less than 20 in Thaum should not be iany way feasible, not with guaranteed 1/3 of dps coming from DoTs no matter what.

If 1/3 of damage is DoT and 2/3 is direct, then each 1% increase from Master at Arms is twice as valuable as a 1% increase from Thaumaturge. The jump from 20 to 28 Thaum is 3% DoT damage, or equivalent to 1% average damage for 8 points invested, so 0.125% gain per CP. The jump from 56 Master to 66 Master is 2% direct damage or 1.33% average damage for 10 points invested, yielding 0.133% gain per CP. The next jump in Master would be 72 for another 1% direct damage or 0.67% average damage costing 6 points or 0.11% per CP. So I agree that 66/20 is the correct split if it is a 1/3 DoT 2/3 direct parse.

Yeah that what I was trying to say. It cant go lower than that. Jump from 23% to 24% M-A-A is 9 points of CP (5 after 20%) yet Thaum all the way to 16% costs only 3 CP per %.

Imryll

Davadin wrote: »

LiquidPony wrote: »

Conclusion: ...is not sufficient for DoTs to be marginally stronger than spammables; they need to be significantly stronger.

disagree.

and this is coming from a stamDK with 56-points in Thaum.

i want both "stack-and-forgot" and burst combo both be viables.

so looking at individual skills, i want DoT that ticks marginally weaker than spammable, because they are inherently less risk.

To me, dots seem riskier because I pay full price whether or not they run their full course. Also folks have more time to respond to them.

WrathOfInnos

SodanTok wrote: »

WrathOfInnos wrote: »

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

I miss 2 important tests here that would actually be telling the story better. One with precise CP distribution (56+56 when you do test where DoT is only 6.7% is very skewed) and one where you go all in on Master-at-Arms but you still use the DoT in rotation.

If not for sake of being correct then for sake of not showcasing data that can be misinterpreted. Like Test3 that shows 700 difference gain over Test2 even tho total damage dealt was 5k less

What is "precise CP distribution"? If you do full parses on a DD-focused setup like a stamblade, and use Constellations to optimize CP, it will eventually optimize virtually every point out of Thaumaturge. Maybe 3-5% or something at most.

Putting 56 points in Thaumaturge was being extremely generous to Rending Slashes in the first place.

And DPS is ... damage divided by time. There's no way to ensure that each parse takes exactly the same amount of time to complete. You can simply look at the actual parse values and extrapolate. Yes, parse #2 has more total damage than parse #3 (and if you look at the mean values of all the skills, that makes perfect sense; Rending Slashes in parse #2 is very slightly stronger than skipping it altogether in parse #3).

Parse #4 shows that optimizing towards Direct Damage means that you're better off just spamming one skill than incorporating weak DoTs.

I will add a note to make sure that people are paying attention to total damage done and the actual mean damage of each cast in CMX, and also do another test with the loaded MaA CP while also using Rending Slashes.

I meant that 56/56 distribution is ideal in condition where DoT and Direct portion of DPS are equal. Speccing into Direct is gonna result in more DPS everytime if DoT portion of parse isnt higher.
So Parse 4 shows that optimizing CP around damage dealt results in higher DPS. Nothing more than that. It would result in DPS increase even with Rending being used (by replacing the one fighters guild skill).

As for Constellation, are you using manual input instead of parse import? I have doubts about the 'it will eventually optimize virtually every point out of Thaumaturge' since just poisons with relequen, hail, trap, DW DoT and possible maarselok will put your DoT portion in parse above 30%. Something parse import in constellation is being incorrect about.

If your DoT damage is 30% of your total damage, and Direct Damage is 60%, and your LA damage is 10%, you're going to end up putting the vast majority of your points in Master at Arms. About a 66/20 split IIRC. And then you'll do another parse and find that your ratio has changed and it will recommend moving a few more points, and so on.

That seems like bug. Using manual calculation of ratios my constellation seems to stabilize within one parse and it shouldnt move after that because its supposed to account for used CP distributon. Or better said, the calculation made on empty (without CP) parse will be nearly set in stone and have no reason to be changing. Getting anything less than 20 in Thaum should not be iany way feasible, not with guaranteed 1/3 of dps coming from DoTs no matter what.

If 1/3 of damage is DoT and 2/3 is direct, then each 1% increase from Master at Arms is twice as valuable as a 1% increase from Thaumaturge. The jump from 20 to 28 Thaum is 3% DoT damage, or equivalent to 1% average damage for 8 points invested, so 0.125% gain per CP. The jump from 56 Master to 66 Master is 2% direct damage or 1.33% average damage for 10 points invested, yielding 0.133% gain per CP. The next jump in Master would be 72 for another 1% direct damage or 0.67% average damage costing 6 points or 0.11% per CP. So I agree that 66/20 is the correct split if it is a 1/3 DoT 2/3 direct parse.

Yeah that what I was trying to say. It cant go lower than that. Jump from 23% to 24% M-A-A is 9 points of CP (5 after 20%) yet Thaum all the way to 16% costs only 3 CP per %.

Yeah 81 Master for 24% direct is probably not worth it in 5.2.3 (it was on some builds in 5.2.0). Depending on investment into Staff Expert, Elfborn and Spell Erosion (or Stamina equivalents) you might be able to get to 72 Master and 40 Thaum. I’ve been putting a lot into Staff Expert myself, since my Light Attack DPS is about the same as my DoT DPS, and it ramps up quicker than Thaum.

LiquidPony

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

LiquidPony wrote: »

SodanTok wrote: »

I miss 2 important tests here that would actually be telling the story better. One with precise CP distribution (56+56 when you do test where DoT is only 6.7% is very skewed) and one where you go all in on Master-at-Arms but you still use the DoT in rotation.

If not for sake of being correct then for sake of not showcasing data that can be misinterpreted. Like Test3 that shows 700 difference gain over Test2 even tho total damage dealt was 5k less

What is "precise CP distribution"? If you do full parses on a DD-focused setup like a stamblade, and use Constellations to optimize CP, it will eventually optimize virtually every point out of Thaumaturge. Maybe 3-5% or something at most.

Putting 56 points in Thaumaturge was being extremely generous to Rending Slashes in the first place.

And DPS is ... damage divided by time. There's no way to ensure that each parse takes exactly the same amount of time to complete. You can simply look at the actual parse values and extrapolate. Yes, parse #2 has more total damage than parse #3 (and if you look at the mean values of all the skills, that makes perfect sense; Rending Slashes in parse #2 is very slightly stronger than skipping it altogether in parse #3).

Parse #4 shows that optimizing towards Direct Damage means that you're better off just spamming one skill than incorporating weak DoTs.

I will add a note to make sure that people are paying attention to total damage done and the actual mean damage of each cast in CMX, and also do another test with the loaded MaA CP while also using Rending Slashes.

I meant that 56/56 distribution is ideal in condition where DoT and Direct portion of DPS are equal. Speccing into Direct is gonna result in more DPS everytime if DoT portion of parse isnt higher.
So Parse 4 shows that optimizing CP around damage dealt results in higher DPS. Nothing more than that. It would result in DPS increase even with Rending being used (by replacing the one fighters guild skill).

As for Constellation, are you using manual input instead of parse import? I have doubts about the 'it will eventually optimize virtually every point out of Thaumaturge' since just poisons with relequen, hail, trap, DW DoT and possible maarselok will put your DoT portion in parse above 30%. Something parse import in constellation is being incorrect about.

If your DoT damage is 30% of your total damage, and Direct Damage is 60%, and your LA damage is 10%, you're going to end up putting the vast majority of your points in Master at Arms. About a 66/20 split IIRC. And then you'll do another parse and find that your ratio has changed and it will recommend moving a few more points, and so on.

That seems like bug. Using manual calculation of ratios my constellation seems to stabilize within one parse and it shouldnt move after that because its supposed to account for used CP distributon. Or better said, the calculation made on empty (without CP) parse will be nearly set in stone and have no reason to be changing. Getting anything less than 20 in Thaum should not be iany way feasible, not with guaranteed 1/3 of dps coming from DoTs no matter what.

~33% of your damage coming from DoTs depends on actually using DoTs and Maarselok and Poisons though. And those things aren't always guaranteed.

Relequen + Hail + Twin Blade & Blunt + Trap is about 27.5% on the parses I'm looking at.

Regardless you're still going to end up heavily favoring MaA on most stam builds.

And I should also note that it remains extremely amusing that Twin Blade & Blunt still does far more DPS than the DoT components of Rending Slashes or Poison Injection.

And again, it bears repeating: if DoTs are only marginally stronger than spammables, people won't use them. This is where ZOS's "nanosecond perfect rotations" actually comes into play. There's no point in bringing in the additional complexity and risk (i.e., fights where your DoTs fall off, like Dragons) of DoT-heavy rotations for an infinitesimal gain. This pattern is horrible. We're quite literally at the point where putting down a ground DoT and spamming one or two skills does 99% as much DPS as a complex, well-timed rotation.

Davadin

Imryll wrote: »

Davadin wrote: »

LiquidPony wrote: »

Conclusion: ...is not sufficient for DoTs to be marginally stronger than spammables; they need to be significantly stronger.

disagree.

and this is coming from a stamDK with 56-points in Thaum.

i want both "stack-and-forgot" and burst combo both be viables.

so looking at individual skills, i want DoT that ticks marginally weaker than spammable, because they are inherently less risk.

To me, dots seem riskier because I pay full price whether or not they run their full course. Also folks have more time to respond to them.

i meant less risk to my health.

dead people do 0 DPS.

but from a resource point of view, i agree. there's risk for wasting ur resource.

Inklings

Drako_Ei wrote: »

If dots are not worth casting... imagine aoe

this ^ this ^ this

What they did to AOEs massively killed diversity in magic builds and left us to play a copy and paste build lay out for every class like stam has had to do for so long do to lack of class stam skills. The main reason why we had so much diversity at least in magic builds when Elsweyr came out was cause of how high DPS was. This opened up so many more unique fun play styles for us and then they took that all away with their attempt to "balance" the game with this skill audit that no one was asking for.

They need to learn to change and balance combat from out side their own shoes and see the bigger picture of what makes this game fun and enjoyable to their player base. You can balance this game down to the smallest of %s but if the game is not fun or enjoyable to play all that math means jack squat.

sfpiesb14_ESO

Nothing to add other than this was a fascinating discussion and extremely insightful. OPs tests may be flawed but they highlight the complexity of the issue and the responses have been great. Not just blindly bashing the devs and it didn’t turn into a nerf sorc/cloak/temp thread or mindless name calling.

LiquidPony

sfpiesb14_ESO wrote: »

Nothing to add other than this was a fascinating discussion and extremely insightful. OPs tests may be flawed but they highlight the complexity of the issue and the responses have been great. Not just blindly bashing the devs and it didn’t turn into a nerf sorc/cloak/temp thread or mindless name calling.

The only people capable of running "perfect" tests in isolation work for ZOS. Certainly some of the data here could be extrapolated and analyzed to get at more exact values, but ... this isn't my job. And to be honest, if I'm going to have to spend that kind of time proving something that is so obvious to ZOS so they don't totally break their game ... well, I'll just go play something else.

ZOS has clearly failed to do the correct tests in isolation. Or at the very least their test expectations are entirely wrong.

But I'm totally bashing the devs. Just not blindly.