Can ZOS share internal DPS parses?

MLGProPlayer · September 2018

The recent class rep notes had a section on DPS parses. Given the huge discrepancy that exists between magicka warden and the other classes in terms of DPS, I challenge ZOS devs to share their internal parses for all the classes and convince me that they are all "relatively close to each other".

I eagerly await your responses!

Sparr0w · September 2018

ermm... they'll just use gilliam right?

Turelus · September 2018

I doubt they're going to do it, because players will work as hard as possible to find faults and scream "devs can't play!" everywhere.

MLGProPlayer · September 2018

Turelus wrote: »

I doubt they're going to do it, because players will work as hard as possible to find faults and scream "devs can't play!" everywhere.

They should be able to replicate all the top parses because that's what they should be balancing the game against. Otherwise, how can they gauge the impact of their changes? And if they can't hit top parses, they need to hire someone who can, asap.

Turelus · September 2018

I'm sure there are people within the company who can do good damage parses. Considering they have internal testers which clear all the trials before they go live they can't be that bad at the game.

As I said though, them releasing things like their internal parses is just going to fuel fires. It's the same as when people ask Devs to come to PvP, there is that element of people wanting to prove the development team doesn't play well so they can then use that in arguments.

MLGProPlayer · September 2018

Turelus wrote: »

I'm sure there are people within the company who can do good damage parses. Considering they have internal testers which clear all the trials before they go live they can't be that bad at the game.

As I said though, them releasing things like their internal parses is just going to fuel fires. It's the same as when people ask Devs to come to PvP, there is that element of people wanting to prove the development team doesn't play well so they can then use that in arguments.

The point is to be more transparent, not to show that they can hit the biggest numbers.

As in the case of the magden changes on PTS, it's evident these weren't play tested, at all. A simple DPS parse comparison against other classes would show the devs that something is very wrong.

They could thus be held accountable for their changes. If their parses show a 15% difference between one class and the rest, we could demand answers as to why that is the case.

This would also force them to actually play test their changes since sharing the parses would force them to explain any major discrepancies. Something tells me that we wouldn't see these discrepancies if they actually play tested their changes.

John_Falstaff · September 2018

@Turelus , I'm actually curious - wouldn't it make a valid argument if people were able to hit higher than devs? Provided that we'll define some basic limitations that would eliminate "cheesed up" parses (say, >= 16k health, one defensive skill slotted, 6mil dummy - maybe something else I omitted, like percentage of crits falling within certain margins?), it would be interesting to see how the developers' parses stack up against highest-parsing players. I'm not condoning the vague criticism ("anyone can do better!"), but transparency is a thing.

Besides, I would -very- much want to see if stamina parses the devs make are made on redguard characters. Now that would give them some flame, and well-deserved at that.

Turelus · September 2018

John_Falstaff wrote: »

Turelus , I'm actually curious - wouldn't it make a valid argument if people were able to hit higher than devs? Provided that we'll define some basic limitations that would eliminate "cheesed up" parses (say, >= 16k health, one defensive skill slotted, 6mil dummy - maybe something else I omitted, like percentage of crits falling within certain margins?), it would be interesting to see how the developers' parses stack up against highest-parsing players. I'm not condoning the vague criticism ("anyone can do better!"), but transparency is a thing.

Besides, I would -very- much want to see if stamina parses the devs make are made on redguard characters. Now that would give them some flame, and well-deserved at that.

If the community could take the data and not act like jerks sure. I know devs in other games have shown builds and parses when discussing changes so there is an idea where comments are coming from.

However let's be entirely honest and say that most people want that information to prove ZOS wrong. Maybe they are wrong and we can highlight issues, but I doubt it'll come off that way. Not to pick on MLGProPlayer but for example of key words in their posts.

MLGProPlayer wrote: »

The point is to be more transparent, not to show that they can hit the biggest numbers.

As in the case of the magden changes on PTS, it's evident these weren't play tested, at all. A simple DPS parse comparison against other classes would show the devs that something is very wrong.

They could thus be held accountable for their changes. If their parses show a 15% difference between one class and the rest, we could demand answers as to why that is the case.

This would also force them to actually play test their changes since sharing the parses would force them to explain any major discrepancies. Something tells me that we wouldn't see these discrepancies if they actually play tested their changes.

That doesn't to me sound like something where a respectable back and forth is going to happen.
It's going to be cries from people that because the devs don't parse 65k with their Nightblade they have no idea how to balance a Sorcerer.

Remember that ZOS also has access to all the live and PTS data, they could in theory look up all the parses done on dummies by the community during testing. They also have on hand players (and now former players) who can parse well for them if they needed it.

Lastly I am not against ZOS showing their parses. I would love to have ZOS come to the table with parses, charts and graphs, if they did post parses I wouldn't be annoyed I would love it, but... as my first reply to the thread stated I don't think they ever will because of how that information will be used by some in the community.

Beardimus · September 2018

MLGProPlayer wrote: »

Turelus wrote: »

I doubt they're going to do it, because players will work as hard as possible to find faults and scream "devs can't play!" everywhere.

They should be able to replicate all the top parses because that's what they should be balancing the game against. Otherwise, how can they gauge the impact of their changes? And if they can't hit top parses, they need to hire someone who can, asap.

honestly I don't think that's a true statement. first of all they can look at other players data. Devs don't need to be pro players.

John_Falstaff · September 2018

@Turelus , granted, it'll be a sine qua non that we'd have to be constructive about the data, and discuss it like we're all in for it to make the game better for everyone. I know there are hotheads, I'm often one myself, but I think a few hotheads isn't a reason for not having any dialogue at all (and dialogue is important - after all, it's largely thanks to dialogue that shield cast times were taken away).

I simply have large concerns after the story with Sload's on previous PTS. ZOS had all data indeed, all duels, all Cyrodiil encounters, all death recaps, CP levels, gear, everything... but Sload's went live and then was nerfed gradually over subsequent patches. That's an indication that ZOS may not be analyzing data they have available as rigorously as they should be, and hand on players didn't highlight the problems either until after release. So yes, in theory, they have the data and in-house testers. In practice... things often get overlooked just the same.

Turelus · September 2018

John_Falstaff wrote: »

Turelus , granted, it'll be a sine qua non that we'd have to be constructive about the data, and discuss it like we're all in for it to make the game better for everyone. I know there are hotheads, I'm often one myself, but I think a few hotheads isn't a reason for not having any dialogue at all (and dialogue is important - after all, it's largely thanks to dialogue that shield cast times were taken away).

I simply have large concerns after the story with Sload's on previous PTS. ZOS had all data indeed, all duels, all Cyrodiil encounters, all death recaps, CP levels, gear, everything... but Sload's went live and then was nerfed gradually over subsequent patches. That's an indication that ZOS may not be analyzing data they have available as rigorously as they should be, and hand on players didn't highlight the problems either until after release. So yes, in theory, they have the data and in-house testers. In practice... things often get overlooked just the same.

Agreed. It does seem (hopefully) that those mistakes are not being repeated though as the new in question set has already been looked at and is still under a watchful eye.

Maybe it's something as we get better communication between the class reps and ZOS we could have more information from ZOS side shared (such as parse data or sexy graph ***).

Sparr0w · September 2018

Turelus wrote: »

John_Falstaff wrote: »

Turelus , granted, it'll be a sine qua non that we'd have to be constructive about the data, and discuss it like we're all in for it to make the game better for everyone. I know there are hotheads, I'm often one myself, but I think a few hotheads isn't a reason for not having any dialogue at all (and dialogue is important - after all, it's largely thanks to dialogue that shield cast times were taken away).

I simply have large concerns after the story with Sload's on previous PTS. ZOS had all data indeed, all duels, all Cyrodiil encounters, all death recaps, CP levels, gear, everything... but Sload's went live and then was nerfed gradually over subsequent patches. That's an indication that ZOS may not be analyzing data they have available as rigorously as they should be, and hand on players didn't highlight the problems either until after release. So yes, in theory, they have the data and in-house testers. In practice... things often get overlooked just the same.

Agreed. It does seem (hopefully) that those mistakes are not being repeated though as the new in question set has already been looked at and is still under a watchful eye.

Maybe it's something as we get better communication between the class reps and ZOS we could have more information from ZOS side shared (such as parse data or sexy graph ***).

What's in the stars tho?...

Turelus · September 2018

Sparr0w wrote: »

Turelus wrote: »

John_Falstaff wrote: »

Turelus , granted, it'll be a sine qua non that we'd have to be constructive about the data, and discuss it like we're all in for it to make the game better for everyone. I know there are hotheads, I'm often one myself, but I think a few hotheads isn't a reason for not having any dialogue at all (and dialogue is important - after all, it's largely thanks to dialogue that shield cast times were taken away).

I simply have large concerns after the story with Sload's on previous PTS. ZOS had all data indeed, all duels, all Cyrodiil encounters, all death recaps, CP levels, gear, everything... but Sload's went live and then was nerfed gradually over subsequent patches. That's an indication that ZOS may not be analyzing data they have available as rigorously as they should be, and hand on players didn't highlight the problems either until after release. So yes, in theory, they have the data and in-house testers. In practice... things often get overlooked just the same.

Agreed. It does seem (hopefully) that those mistakes are not being repeated though as the new in question set has already been looked at and is still under a watchful eye.

Maybe it's something as we get better communication between the class reps and ZOS we could have more information from ZOS side shared (such as parse data or sexy graph ***).

What's in the stars tho?...

The common word for adult material, although spelt the lolz with with the r0 rather than the or. I didn't realise that word was filtered.

JinMori · September 2018

Sparr0w wrote: »

ermm... they'll just use gilliam right?

Obviously.

JinMori · September 2018

John_Falstaff wrote: »

@Turelus , I'm actually curious - wouldn't it make a valid argument if people were able to hit higher than devs? Provided that we'll define some basic limitations that would eliminate "cheesed up" parses (say, >= 16k health, one defensive skill slotted, 6mil dummy - maybe something else I omitted, like percentage of crits falling within certain margins?), it would be interesting to see how the developers' parses stack up against highest-parsing players. I'm not condoning the vague criticism ("anyone can do better!"), but transparency is a thing.

Besides, I would -very- much want to see if stamina parses the devs make are made on redguard characters. Now that would give them some flame, and well-deserved at that.

From what i heard there are only a couple of devs who can actually play the game at a high level, one of them is gilliam, another one was mentioned by alcast, but i don;t remember his name.

Anyway, if you are developing something, it should be expected that you are at least an above average player, i mean, being in contact with the game everyday, testing and stuff, you should be a pretty damn good player right? At least it should be expected, how can we trust you balancing when you can;t even play? when many players know more about a class than the developers do?

Turelus · September 2018

JinMori wrote: »

John_Falstaff wrote: »

Turelus , I'm actually curious - wouldn't it make a valid argument if people were able to hit higher than devs? Provided that we'll define some basic limitations that would eliminate "cheesed up" parses (say, >= 16k health, one defensive skill slotted, 6mil dummy - maybe something else I omitted, like percentage of crits falling within certain margins?), it would be interesting to see how the developers' parses stack up against highest-parsing players. I'm not condoning the vague criticism ("anyone can do better!"), but transparency is a thing.

Besides, I would -very- much want to see if stamina parses the devs make are made on redguard characters. Now that would give them some flame, and well-deserved at that.

From what i heard there are only a couple of devs who can actually play the game at a high level, one of them is gilliam, another one was mentioned by alcast, but i don;t remember his name.

Anyway, if you are developing something, it should be expected that you are at least an above average player, i mean, being in contact with the game everyday, testing and stuff, you should be a pretty damn good player right? At least it should be expected, how can we trust you balancing when you can;t even play? when many players know more about a class than the developers do?

Remember that there is also a difference between a designer and a tester in some regards. Someone might be fantastic at designing interesting challenges and content but not able to clear it themselves. This is where internal testing groups come into play.

For devs we actually know who play, we know that Rich and Finn are both pretty good at end game PvE content, Wrobel has been on ESO Live in groups and I think it's Bobby who is often on ESO Live in their groups as well?
PvP wise Wrobel and Brian have commented in the past about PvP content, I've also seen some of their non design or leads tweet their Battlegrounds scores in the past.
They probably have high end people in all areas, but we only know of the big names working at ZOS. Jeff the QA guy who tests dungeons all day isn't on our radar.

Edit: Bold part showing once again why if they show their information and it doesn't match player expectations it's going to end badly.

Merlin13KAGL · September 2018

Turelus wrote: »

I doubt they're going to do it, because players will work as hard as possible to find faults and scream "devs can't play!" everywhere.

Which is why you take the player completely out of the equation initially.

The could essentially code an NPC to do a 'perfect rotation.' Perfectly timed GCD's and LA's, perfectly timed potions and executes.

Give only baseline crit chance (taking RNG out of the equation, also). Parse after the fact could show the theoretical perfect minimum (0 crits), max (100% crits), and average (based on crit totals for given set/skill combinations.)

No mistakes, no lag, no human factor at all.

These would be the highest parses the game could ever allow and could be brute forced for any combination of skills/gear, then adjusted after-the-fact to show the same results at various CP levels.

From there, humans could try to reproduce the results: High level players, mid-tier players, and relatively newer players.

They could look at how each of these things varied, 'ceiling' to 'floor' and theoretical parse compared to live parse, and make better adjustments from there.

These could be repeated ad infinitum with each new patch, skill, and set introduction. No machine fatigue ever setting in, no human error initially in the mix.

Follow up with examination of the massive amounts of live data available at any given time, for any fight, any player, any build.

They own the kingdom already. They don't need the keys. Let the simulation do the initial work for them and all would benefit.

Turelus · September 2018

Merlin13KAGL wrote: »

Turelus wrote: »

I doubt they're going to do it, because players will work as hard as possible to find faults and scream "devs can't play!" everywhere.

Which is why you take the player completely out of the equation initially.

The could essentially code an NPC to do a 'perfect rotation.' Perfectly timed GCD's and LA's, perfectly timed potions and executes.

Give only baseline crit chance (taking RNG out of the equation, also). Parse after the fact could show the theoretical perfect minimum (0 crits), max (100% crits), and average (based on crit totals for given set/skill combinations.

No mistakes, no lag, no human factor at all.

These would be the highest parses the game could ever allow and could be brute forced for any combination of skills/gear, then adjusted after-the-fact to show the same results at various CP levels.

From there, humans could try to reproduce the results: High level players, mid-tier players, and relatively newer players.

They could look at how each of these things varied, top 1% down to lower 1% and theoretical parse compared to live parse, and make better adjustments from there.

These could be repeated ad infinitum with each new patch, skill, and set introduction. No machine fatigue ever setting in, no human error initially in the mix.

Balancing around that wouldn't be helpful still. Unless all you want is for every class to have the same max computerised value.
It doesn't take into account so many variables in regards to real actual gameplay values.

If all we want to balance is a computerised max value sure, but that won't stop stamina X being weaker in Y because of Z when it comes to an actual trial.

OrphanHelgen · September 2018

I don't mind parses being different, as long as the the classes with lower single target make up for it in another way. Either way more cleave dmg or healing etc.
The moment one class dominated on all places, is where balance is needed.
Right now I would personally like to see no more content coming out, but purely focusing on balancing and more creativity. I have been casting hail and caltrops for years now. It's not normal to have this little creativity in a subscription based mmo. It's many years ago since we first mentioned a 3rd morph for example. Or 150 sets that needs a rework. Nothing happens.

Merlin13KAGL · September 2018

Turelus wrote: »

Merlin13KAGL wrote: »

Turelus wrote: »

I doubt they're going to do it, because players will work as hard as possible to find faults and scream "devs can't play!" everywhere.

Which is why you take the player completely out of the equation initially.

The could essentially code an NPC to do a 'perfect rotation.' Perfectly timed GCD's and LA's, perfectly timed potions and executes.

Give only baseline crit chance (taking RNG out of the equation, also). Parse after the fact could show the theoretical perfect minimum (0 crits), max (100% crits), and average (based on crit totals for given set/skill combinations.

No mistakes, no lag, no human factor at all.

These would be the highest parses the game could ever allow and could be brute forced for any combination of skills/gear, then adjusted after-the-fact to show the same results at various CP levels.

From there, humans could try to reproduce the results: High level players, mid-tier players, and relatively newer players.

They could look at how each of these things varied, top 1% down to lower 1% and theoretical parse compared to live parse, and make better adjustments from there.

These could be repeated ad infinitum with each new patch, skill, and set introduction. No machine fatigue ever setting in, no human error initially in the mix.

Balancing around that wouldn't be helpful still. Unless all you want is for every class to have the same max computerised value.
It doesn't take into account so many variables in regards to real actual gameplay values.

If all we want to balance is a computerised max value sure, but that won't stop stamina X being weaker in Y because of Z when it comes to an actual trial.

It would merely be a starting point, and of course things would have to be adjusted for live human factors, everything from lag to reaction time to differences in hardware.

My point being, if done right, no one could ever go beyond the theoretical maximum, and if the theoretical max on one setup is 20k DPS more than the maximums on most other setups, then it would help point out some balance issues with sets, skills, or classes.

Major discrepencies (especially from high end players) between the theoretical max and live data could help point out issues with content mechanics or other in game aspects.

I'm not even calling for this data to be available anywhere but internally. Not looking for a FoTM builder from ZoS. I'm looking at a potential tool they could use to get things closer to where they need to be, floor, ceiling, in game mechanics and encounters, and everything in between.

And they have live data for days ~ everything that comes across their servers. What makes a Hodor run compare to a Progression run compared to a PuG run? HM SR ND's vs standard runs vs norms...

MLGProPlayer · September 2018

Turelus wrote: »

Merlin13KAGL wrote: »

Turelus wrote: »

I doubt they're going to do it, because players will work as hard as possible to find faults and scream "devs can't play!" everywhere.

Which is why you take the player completely out of the equation initially.

The could essentially code an NPC to do a 'perfect rotation.' Perfectly timed GCD's and LA's, perfectly timed potions and executes.

Give only baseline crit chance (taking RNG out of the equation, also). Parse after the fact could show the theoretical perfect minimum (0 crits), max (100% crits), and average (based on crit totals for given set/skill combinations.

No mistakes, no lag, no human factor at all.

These would be the highest parses the game could ever allow and could be brute forced for any combination of skills/gear, then adjusted after-the-fact to show the same results at various CP levels.

From there, humans could try to reproduce the results: High level players, mid-tier players, and relatively newer players.

They could look at how each of these things varied, top 1% down to lower 1% and theoretical parse compared to live parse, and make better adjustments from there.

These could be repeated ad infinitum with each new patch, skill, and set introduction. No machine fatigue ever setting in, no human error initially in the mix.

Balancing around that wouldn't be helpful still. Unless all you want is for every class to have the same max computerised value.
It doesn't take into account so many variables in regards to real actual gameplay values.

If all we want to balance is a computerised max value sure, but that won't stop stamina X being weaker in Y because of Z when it comes to an actual trial.

It would be an important first step towards balance and avoid the creation of grossly underpowered classes (like the current iteration of magden) or grossly overpowered.

Once you have roughly the same DPS for everyone, you can start looking at other a factors (like utility) and adjust DPS accordingly.

MLGProPlayer · September 2018

OrphanHelgen wrote: »

I don't mind parses being different, as long as the the classes with lower single target make up for it in another way. Either way more cleave dmg or healing etc.
The moment one class dominated on all places, is where balance is needed.
Right now I would personally like to see no more content coming out, but purely focusing on balancing and more creativity. I have been casting hail and caltrops for years now. It's not normal to have this little creativity in a subscription based mmo. It's many years ago since we first mentioned a 3rd morph for example. Or 150 sets that needs a rework. Nothing happens.

This is what ZOS fails at spectacularly. Taking the magden class for instance, they have the lowest DPS in the game with 0 utility (none of their DPS skills provide group buffs or target debuffs).