scaling, dreadnought, and hellbringer changes are all m19, arcanist adjustments are m18. None of that was the intention of the thread which is why it seems out of place for conversation in the preview forums for something a full module away.
The only way those numbers are even remotely close to being true is perhaps if you include (catastrophically) failed runs. There is sure going to be a lot of those so they will have a huge impact. Training runs with one or two strong dps players the same. Also because most of those dps specs aren't accepted for runs, I doubt some of the numbers (dps GF is probably the best example) have any significance.
The numbers were for ToMM runs over the period specified and it is accurate data that has been normalized for better comparison. We also have the whisker plots to see the full range of every player over the time period and where they fall.
The point of showing the ToMM data is not to suggest that in all aspects of the game the classes are that close together. It was intended to show that 1) when top item levels and top skill levels combine, the classes potentials are a lot closer than players would typically expect and 2) ToMM is not exclusive to any sub group of classes, although it is certainly easier for some classes than others.
We don't use that particular data set for where our major class balance efforts are targetted, we focus more on the top 10% game wide, which does show bigger percentage differences between the paragon paths, and does drop Dreadnought down quite a bit more as I had mentioned in that previous post.
Since this data was taken from TOMM runs and normalised for better comparison so why make this confusing and lets stick with TOMM as an example. So from your findings, a wiz is apparently only doing 6% more than fighter ST dps and 1.4% more than Hellbringer?
What if i can prove to you that actually, a cleric can out dps a fighter, warlock and even barb and rogue and on par with rangers?
As regards scaling and caps @noworries#8859 I am curious if you able to offer any further design insight regarding caps where scaling does *not* apply. For example, in the case of LoMM vs ToMM vs MOD18 Dungeon, the gradual increase of critter strength means that a played properly geared for MOD18 Dungeon will, upon returning to LoMM, find herself seriously overcapped. Intuitively, one might assume that gear obtained from "ultimate" content remains superior; in this case, however, being overcapped for content will always be undesirable and lead to worse performance, because the compromises made for those caps (reducing Power, say, or removing % Damage gear) no longer offer any benefit.
Not wishing to ask leading questions here, but is the expectation that players will voluntarily swap to older gear for optimal gear? Or has there been some discussion regarding "softening" caps to make Power-stacking less mandatory?
The only way those numbers are even remotely close to being true is perhaps if you include (catastrophically) failed runs. There is sure going to be a lot of those so they will have a huge impact. Training runs with one or two strong dps players the same. Also because most of those dps specs aren't accepted for runs, I doubt some of the numbers (dps GF is probably the best example) have any significance.
The numbers were for ToMM runs over the period specified and it is accurate data that has been normalized for better comparison. We also have the whisker plots to see the full range of every player over the time period and where they fall.
The point of showing the ToMM data is not to suggest that in all aspects of the game the classes are that close together. It was intended to show that 1) when top item levels and top skill levels combine, the classes potentials are a lot closer than players would typically expect and 2) ToMM is not exclusive to any sub group of classes, although it is certainly easier for some classes than others.
We don't use that particular data set for where our major class balance efforts are targetted, we focus more on the top 10% game wide, which does show bigger percentage differences between the paragon paths, and does drop Dreadnought down quite a bit more as I had mentioned in that previous post.
Since this data was taken from TOMM runs and normalised for better comparison so why make this confusing and lets stick with TOMM as an example. So from your findings, a wiz is apparently only doing 6% more than fighter ST dps and 1.4% more than Hellbringer?
What if i can prove to you that actually, a cleric can out dps a fighter, warlock and even barb and rogue and on par with rangers?
I always enjoy watching top tier players show off their skills, so I would certainly enjoy seeing a cleric performing that well. For clarification, in game wide data the Arbiter does outperform Dreadnought already, with a larger sample size it wouldn't be surprising at all to see them also exceed fighter in ToMM. But it would certainly be fun to watch an Arbiter go toe to toe with a top tier ranger.
In the end though, that by itself wouldn't change our plans to bring Arbiter and Dreadnought up to the target range for balance as on average they're both below where we'd like them to be.
Yeah, never seen any hellbringer warlock nor an arbiter even come close to numbers within successful runs. I would love to see at least one screenshot of a warlock in a successful run put out that damage differential to a cw and rogue.
The hellbringer class mechanics and magnitudes at peak damage potential don't even come close to base rotation of arcanist, let alone within those numbers in tomm.
Try showing us statistics of successful runs....because those numbers being "normalized" appear to be code for skewed. Take data from hundreds of random group fails and handfull or successful runs, muddle the results together and what you get vs just comparing successful run data and you have vastly different results.
The only way those numbers are even remotely close to being true is perhaps if you include (catastrophically) failed runs. There is sure going to be a lot of those so they will have a huge impact. Training runs with one or two strong dps players the same. Also because most of those dps specs aren't accepted for runs, I doubt some of the numbers (dps GF is probably the best example) have any significance.
The numbers were for ToMM runs over the period specified and it is accurate data that has been normalized for better comparison. We also have the whisker plots to see the full range of every player over the time period and where they fall.
The point of showing the ToMM data is not to suggest that in all aspects of the game the classes are that close together. It was intended to show that 1) when top item levels and top skill levels combine, the classes potentials are a lot closer than players would typically expect and 2) ToMM is not exclusive to any sub group of classes, although it is certainly easier for some classes than others.
We don't use that particular data set for where our major class balance efforts are targetted, we focus more on the top 10% game wide, which does show bigger percentage differences between the paragon paths, and does drop Dreadnought down quite a bit more as I had mentioned in that previous post.
I didn't doubt the accuracy of the data itself, I doubted the preparation/preprocessing. If you just average damage charts for a class across all TOMM runs, the result has no meaningful value. At minimum, all failed runs should be removed during preprocessing. I'm not sure what specifically you meant by normalization in this context.
I have been part of all sorts of TOMM runs, from training to speedruns. I can tell you from experience (and have numerous act logs) that as a warden, I cannot even come close to the suggested ~98.5% of arcanist's damage. Have I done it before? Yes, of course. But every time I did I had some sort of an advantage -- better build, less res sickness stacks, being able to position better to maximize dps time etc. When "top item levels and skill levels combine" as you say, the differences are there. And they are much larger than your data suggests.
Thank you for your answer and the insight, I appreciate it.
Always a popular topic so let's give a very brief history to this point and talk about the work going on for M19 and beyond.
I joined the Neverwinter team as Cloaked Ascendancy shipped. At that time there wasn't any useful scaling, although the game did attempt to level scale players in certain situations. As I'm sure many people remember during that time, if a player was level scaled up, they just got crushed (Ghost Stories). If a player was level scaled down, they didn't really notice it and just one shot everything.
We wanted much better scaling than that in the game. Unfortunately, there was not any real base structure for this to be built on. We did some initial backend work that I think was first included in the m14 build (didn't go back to verify while writing this post). With M15 we used this as part of the Acq Inc structure for running the dungeons at minimum item level. While this was the first time there was any noticeable effect from scaling, it was still far from ideal and still not very effective. The main aspect at this point was that we could now scale enchantments, although the way that worked is it would actually make the enchantments a lower or higher rank to change the effects.
With M16 we were now tackling a lot of core and backend structure of the game to allow the game to grow in a lot more ways. This isn't directly evident with what players saw as so much of the work was structural and organizational to allow faster and better work in the future. One of those tasks was to create a much bigger backend for scaling. The result still wan't ideal scaling either for us or for players, but it was a lot closer. Scaling now had a tangible effect, even if it wasn't as accurate as it could be and scaling up is still problematic in some ways. The main method scaling now takes is adjusting the item level of the different parts of the character, as all of the stats are based off of that item level (a change that we made in M16).
With M19 we plan to take the next pass on scaling. It is too early to guarantee how it will turn out/function/etc., but we can discuss the basic goals. Now that the backend exists, we can refine it and improve it. This is going to have a major change that most players will like which is that players stats will scale relative to the content's stats. In other words, the goal of this scaling is that if the player is capped on ratings at their unscaled item level, whatever content they're scaled down or up to, they remain capped on those ratings because they are scaled proportionally to the content. That will solve the major issues of having to worry about changing gear as you are scaled due to scaling tackling each equipment piece individually resulting in some over/under scaling on certain parts of the player that we have now. There are other benefits coming with this next step of scaling that include even easier adjustments on content difficulty allowing us to tune faster and better.
the player base DOES NOT WANT scaling. this is ALL YOU. it is alienating your player base. please just delete it and put it back where it was when you signed on. that's what we want. just make all the dungeons level 80 other than the original levelling dungeons. no one is under level 80 for any significant period of time anyway. it just doesn't make sense.
> @thefiresidecat said: > (Quote) > the player base DOES NOT WANT scaling. this is ALL YOU. it is alienating your player base. please just delete it and put it back where it was when you signed on. that's what we want. just make all the dungeons level 80 other than the original levelling dungeons. no one is under level 80 for any significant period of time anyway. it just doesn't make sense.
Scaling has made me realize there is no reason to buy zen or upgrade enhancements and companions. I just run my master expeditions and log off. I will never feel powerful because scaling prevents it.
Scaling also is a direct affront to D&D which this game is supposed to be based on. If I fight a critter at level 50, when I see it again at level 80 it should be toast. I should not be automatically drug back down to it's level because I worked my bum off to get LOMM/TOMM stats. That is the whole freaking point of progression. SMDH
None of you have ever justified why you are scaling at all. Why? Because it makes content easier for us who spent the Zen/AD/100's or 1000's of hours to get where we are? Stop penalising your most dedicated players, please.
The numbers were for ToMM runs over the period specified and it is accurate data that has been normalized for better comparison. We also have the whisker plots to see the full range of every player over the time period and where they fall.
Thank you for giving insight into your collected data. Still, like others already mentioned here, I have problems to match those numbers to actual DPS results (out of my experience). But I know data preparation can have a huge impact in data interpretation. I don't doubt your numbers, but I still have some questions:
With "normalized" you mean you recalculated all data with a z-transformation using the Assassin as a baseline? So the mean Assassin does "0" damage and somebody doing one Assassin standard deviation more damage will get the score "1"?
Are those numbers a comparison of the means or medians out of the non-normalized data? Or are those the results of the normalized data? (I would also be interested in the sample sizes.)
Is there a chance to see a whisker plot/box plot?
Those questions are just out of curiosity. I kinda like statistics.
We agree that there is an imbalance in the DPS roles. In M19 the two main classes being adjusted for balance are the Dreadnought and the Hellbringer. When M18 comes to preview we will have information on a few adjustments to Arcanist feats/class mechanics to bring them more in line with the balance target (I realize saying this will make people think the worst, but Arcanist will still be powerful after those adjustments and the changes will be available as preview goes live for feedback).
All of that is important context for the following part of that discussion. First is that there will never be perfect balance across the classes, and there will always be some classes that are harder to play and therefore under-perform for a more casual player, but can potentially even over-perform for a particularly skilled player. We have created a lot of analytics on class balance since M16. These include normalized damage charts, whisker plots, and percentile graphs, which we can filter by time ranges, classes, and specific content. We are actively using this information for how to tackle class balance.
As a general point, Assassin in most charts is right around where we feel ideal balance should be right now. Since ToMM was brought up, let's take a look at the PC results from 1 Nov to this morning, and use Assassin as a baseline for where the other classes are at when running that content. I think players may be surprised at where some of the classes line up in this comparison.
Class -> Damage performance in ToMM +/- %
Arcanist +3%
Blademaster +1.8%
Warden +1.5%
Assassin --
Hellbringer -1.6%
Dreadnaught -3%
Arbiter -6%
It is clear there are outliers in Arcanist, Dreadnaught, and Arbiter. The others, however, are all very close together and in general would be considered all within an acceptable range of balance. In charts that include a wider range of content (or all content) there are larger percentage differences which shows there are more areas of balance to tackle than this one chart shows. ToMM is a useful example to see how the classes compare when played by top tier IL players, and hopefully also shows that ToMM is complete-able (and has been completed) by all classes in the game.
Changes from this chart compared to more broad charts show a larger positive differential for Arcanist and Warden, brings Blademaster below Assassin with Hellbringer right behind that and brings Arbiter above Dreadnought. You see a wider variance when including a larger selection of content as it adds a far greater percentage of the player base into the damage pool.
There were some paragon paths not listed there, such as Whisperknife, Hunter, and Thaumaturge. Whisperknife and Hunter are not performing where we'd like, and are paths we want to work on, however since those classes have very solid paths as their other choices, that puts the priority a bit further down the list on class work.
I think you're missing some filters. You can't make an average of everything dividing only per class, it's a too big approximation which is not descriptive of the reality. You can make a comparison ONLY with same base data input: - Same gear (or same amount of %damage bonus); - Same Power; - Same amount of buff/debuff active; - Same amount of revive sicknesses (more complicated); - Player skills (impossible).
While the first 4 are possible to take into account, it's not possible with the last one. And it's the one with the biggest random oscillation since there are many players with different skills levels. For the other points it's possible to make a filter to really see the damage output.
As stated from other people in many other posts in this forum, the DPS balance and damage output it's way far from what you're describing.
I'm happy to read that. In italian we say "meglio tardi che mai" which is "late is better than never". It's long time we're asking a better work with scaling, because actually it's a mess in most of the dungeons/skirmishes/trials. I hope in mod 19 it will be good, since we're waiting 3 mods to have a good scaling.
> @noworries#8859 said: > (Quote) > I always enjoy watching top tier players show off their skills, so I would certainly enjoy seeing a cleric performing that well. For clarification, in game wide data the Arbiter does outperform Dreadnought already, with a larger sample size it wouldn't be surprising at all to see them also exceed fighter in ToMM. But it would certainly be fun to watch an Arbiter go toe to toe with a top tier ranger. > > In the end though, that by itself wouldn't change our plans to bring Arbiter and Dreadnought up to the target range for balance as on average they're both below where we'd like them to be.
I really like, that you plan to balance Dreadnought (and Arbiter). But I hope that will happen before mod19 (and before mod18). Because mod17 is live. On console it will be live for a long time. And I as a Dreadnought want to have the chance to pass ToMM before mod18 or mod19 will go live. I also want to have the chance to enter a random group for ToMM and I don't want to get refused all time because I am a Dreadnought.
PS: sorry for off-topic. But I have that on my mind.
Can you tell us if all dps paragon will have 3 offensive slot. Right now Barb's and gf have 2 less. And tha data that barbarian are 1.2 % behind arcanist in single Target aka tomm runs with 2 offense slot less is not possible. As other are saying this data are way off to the one we see everyday when we run tomm.
> @coolgor28#5062 said: > Can you tell us if all dps paragon will have 3 offensive slot. Right now Barb's and gf have 2 less. And tha data that barbarian are 1.2 % behind arcanist in single Target aka tomm runs with 2 offense slot less is not possible. As other are saying this data are way off to the one we see everyday when we run tomm.
I think all dps classes should have 3 offensive slots... When using a defensive tank 3... You should be able to change your teammates depending on the build
I think all dps classes should have 3 offensive slots... When using a defensive tank 3... You should be able to change your teammates depending on the build
I am very happy that this is being discussed. I am an endgame barb and from what is my experience every single day those data do not correspond to reality. Arcanists, rangers but also thieves do much more damage for the same objects as my barbarian, especially on bosses. We need adjustments in many parts and it does not seem at all sufficient to me what I have read. I was hoping for a more incisive intervention, I feel that I will be disappointed and I believe that I will stop playing if mod 18 does not solve this huge gap. I really don't understand why you don't listen to what the players of a certain class are telling you, since it is under everyone's eyes. I wish I could make toom as a dps and feel necessary as an arcanist but instead I am seen as a doll barbie. What's the point of playing again if I can't run the final dungeon as a dps? I've never seen this in any game
Or why not let all classes have 1 utility slot, 1 offensive slot, 1 defensive slot, and 2 offensive/defensive slots (that allow for the slotting of offense or defense companion bonus)? With this, you address the issue of certain DPS classes being penalized and also provide greater flexibility to players in building their characters.
The only way those numbers are even remotely close to being true is perhaps if you include (catastrophically) failed runs. There is sure going to be a lot of those so they will have a huge impact. Training runs with one or two strong dps players the same. Also because most of those dps specs aren't accepted for runs, I doubt some of the numbers (dps GF is probably the best example) have any significance.
The numbers were for ToMM runs over the period specified and it is accurate data that has been normalized for better comparison. We also have the whisker plots to see the full range of every player over the time period and where they fall.
The point of showing the ToMM data is not to suggest that in all aspects of the game the classes are that close together. It was intended to show that 1) when top item levels and top skill levels combine, the classes potentials are a lot closer than players would typically expect and 2) ToMM is not exclusive to any sub group of classes, although it is certainly easier for some classes than others.
We don't use that particular data set for where our major class balance efforts are targetted, we focus more on the top 10% game wide, which does show bigger percentage differences between the paragon paths, and does drop Dreadnought down quite a bit more as I had mentioned in that previous post.
Since this data was taken from TOMM runs and normalised for better comparison so why make this confusing and lets stick with TOMM as an example. So from your findings, a wiz is apparently only doing 6% more than fighter ST dps and 1.4% more than Hellbringer?
What if i can prove to you that actually, a cleric can out dps a fighter, warlock and even barb and rogue and on par with rangers?
I always enjoy watching top tier players show off their skills, so I would certainly enjoy seeing a cleric performing that well. For clarification, in game wide data the Arbiter does outperform Dreadnought already, with a larger sample size it wouldn't be surprising at all to see them also exceed fighter in ToMM. But it would certainly be fun to watch an Arbiter go toe to toe with a top tier ranger.
In the end though, that by itself wouldn't change our plans to bring Arbiter and Dreadnought up to the target range for balance as on average they're both below where we'd like them to be.
Thank you, it is appreciated. I do not doubt your findings, like i said, i have seen barbs dish out 286k encpds over 3-4minutes dummy test (probably more for the one barb in my guild). I am sitting comfortably at 300encpds on a dummy with no lion heart and 193k power but i have another build which presumably can do much better but harder to play. What i do think makes the difference is how difficult the class is. Wizards, rangers and assassins are just too easy to play and can easily mimic a dummy test in TOMM than an arbiter or a blademaster can, so my dummy tests over 3-5minutes are too difficult to mimic in a real dungeon and that's where most arbiters in TOMM fail. That doesn't mean the class doesn't have the capacity. The reason why most of us are infuriated is because the select few that know their class thoroughly are judged because the majority of others are not able to play their class optimally and thus we are discriminated. I will look forward to arbiter adjustments as i have so much to share but making gear complement them and making their mistakes less unforgiving in the kit can be a way forward. I can't however say anything about hellbringer, but i have seen amazing hellbringers, however, TOMM favours wiz way too much so they are always too far ahead in dps. It is widely known that to play wiz optimally, you have to smash the board as fast as you can, whereas arbiter is all about timing, a well thought out class i must say but a little clunky right now though.
Be wary of the population numbers by class in your sample. Far fewer Arbiters, Dreadnoughts, Blademasters and Hellbringers complete ToMM than the other 3 DPS classes and they are likely to be the cream of the crop, typically alts of players who are already experienced runners of ToMM on another class. Furthermore, the population will also contain a far higher percentage of "weaker" Arcanists completing ToMM because it is the most sort after DPS class. Low populations can seriously skew outcomes and conclusions, thus the old adage Lies, Damned Lies and Statistics
15
gabrieldourdenMember, NW M9 PlaytestPosts: 1,212Arc User
While I am always hesitant to go off topic in a thread as it makes the information about the initial topic hard to find. In this case, at least so far, there doesn't seem to be any confusion about the information in the initial post, so I will briefly tackle the 2 main things being brought up here, but there likely won't be a lot of back and forth on those topics in this particular thread.
Class Balance
We agree that there is an imbalance in the DPS roles. In M19 the two main classes being adjusted for balance are the Dreadnought and the Hellbringer. When M18 comes to preview we will have information on a few adjustments to Arcanist feats/class mechanics to bring them more in line with the balance target (I realize saying this will make people think the worst, but Arcanist will still be powerful after those adjustments and the changes will be available as preview goes live for feedback).
All of that is important context for the following part of that discussion. First is that there will never be perfect balance across the classes, and there will always be some classes that are harder to play and therefore under-perform for a more casual player, but can potentially even over-perform for a particularly skilled player. We have created a lot of analytics on class balance since M16. These include normalized damage charts, whisker plots, and percentile graphs, which we can filter by time ranges, classes, and specific content. We are actively using this information for how to tackle class balance.
As a general point, Assassin in most charts is right around where we feel ideal balance should be right now. Since ToMM was brought up, let's take a look at the PC results from 1 Nov to this morning, and use Assassin as a baseline for where the other classes are at when running that content. I think players may be surprised at where some of the classes line up in this comparison.
Class -> Damage performance in ToMM +/- %
Arcanist +3%
Blademaster +1.8%
Warden +1.5%
Assassin --
Hellbringer -1.6%
Dreadnaught -3%
Arbiter -6%
It is clear there are outliers in Arcanist, Dreadnaught, and Arbiter. The others, however, are all very close together and in general would be considered all within an acceptable range of balance. In charts that include a wider range of content (or all content) there are larger percentage differences which shows there are more areas of balance to tackle than this one chart shows. ToMM is a useful example to see how the classes compare when played by top tier IL players, and hopefully also shows that ToMM is complete-able (and has been completed) by all classes in the game.
Changes from this chart compared to more broad charts show a larger positive differential for Arcanist and Warden, brings Blademaster below Assassin with Hellbringer right behind that and brings Arbiter above Dreadnought. You see a wider variance when including a larger selection of content as it adds a far greater percentage of the player base into the damage pool.
There were some paragon paths not listed there, such as Whisperknife, Hunter, and Thaumaturge. Whisperknife and Hunter are not performing where we'd like, and are paths we want to work on, however since those classes have very solid paths as their other choices, that puts the priority a bit further down the list on class work.
Thanks for the numbers. As an HR player I'd really like you to work on Hunter. Warden is performing well, but it is extremely boring, has basically one single build and transform all HRs in the game into cookie-cutter copies of each other. Most of the HR powers are completely useless.
Please do not focus only on numbers, but also on what is fun to play. A character with single power that does average damage and has an average cooldown will fulfill your damage requirements but will be horrible to play.
Le-Shan: HR level 80 (main)
Born of Black Wind: SW Level 80
Bingo! You cannot estimate skill ceiling, so in reality, taking a sample from a bigger population will actually result in a skewed result where the dps of that specific class will be lower. It's obvious that more wiz's are played in TOMM so their average dps will skew to being average whereas generally only the best of non-meta classes are chosen for TOMM. Although, i am a bit skeptical about the exact %performance difference, i do know that skill ceiling non-meta classes can really shock the majority of playerbase if they saw their true potential. We started with 50% dps difference between a barb and wiz but now people have adjusted it to 15-20%. The dev's consider it about 1.2%, which is definitely lower than i would put it at, especially in TOMM.
I do advise the dev's to stick with skill ceiling players only, especially for TOMM and balance ST depending on that. For other dungeons, the bigger the sample the better, in my opinion.
scaling, dreadnought, and hellbringer changes are all m19, arcanist adjustments are m18. None of that was the intention of the thread which is why it seems out of place for conversation in the preview forums for something a full module away.
damn another 6/7 months until gf and sw changes... sigh
Is the Hellbringer TOMM performance statistics including or excluding the period of time when a bug caused them to do millions of damage in split second?
Well, I suppose I should be happy that Dreadnaught is finally going to get some help, but, I do have to say that there is a considerable gap and I am skeptical that the changes will narrow this gap much, adding 3% or even 10% damage is not going to cut it, the Dreadnaught really is terrible compared to CW/TR/HR
During that 4 day sample size of ToMM there was probably 1000 Arcanist....of which 80% were likely bad players but were invited simply because they are wizards. Where as the dreadnaughts, Warlock's, arbiters, and blademasters probably accounted for 25 total players....all of who were invited because they are OUTSTANDING players.
... so, until you guys realize that something is wrong in that balance data, if something is wrong in that balance data, we will need wait, for example, one year at least?
but i have a simple question: what is the difference, in damage, between tanks/healers and that dps?
While I am always hesitant to go off topic in a thread as it makes the information about the initial topic hard to find. In this case, at least so far, there doesn't seem to be any confusion about the information in the initial post, so I will briefly tackle the 2 main things being brought up here, but there likely won't be a lot of back and forth on those topics in this particular thread.
Class Balance
We agree that there is an imbalance in the DPS roles. In M19 the two main classes being adjusted for balance are the Dreadnought and the Hellbringer. When M18 comes to preview we will have information on a few adjustments to Arcanist feats/class mechanics to bring them more in line with the balance target (I realize saying this will make people think the worst, but Arcanist will still be powerful after those adjustments and the changes will be available as preview goes live for feedback).
All of that is important context for the following part of that discussion. First is that there will never be perfect balance across the classes, and there will always be some classes that are harder to play and therefore under-perform for a more casual player, but can potentially even over-perform for a particularly skilled player. We have created a lot of analytics on class balance since M16. These include normalized damage charts, whisker plots, and percentile graphs, which we can filter by time ranges, classes, and specific content. We are actively using this information for how to tackle class balance.
As a general point, Assassin in most charts is right around where we feel ideal balance should be right now. Since ToMM was brought up, let's take a look at the PC results from 1 Nov to this morning, and use Assassin as a baseline for where the other classes are at when running that content. I think players may be surprised at where some of the classes line up in this comparison.
Class -> Damage performance in ToMM +/- %
Arcanist +3%
Blademaster +1.8%
Warden +1.5%
Assassin --
Hellbringer -1.6%
Dreadnaught -3%
Arbiter -6%
It is clear there are outliers in Arcanist, Dreadnaught, and Arbiter. The others, however, are all very close together and in general would be considered all within an acceptable range of balance. In charts that include a wider range of content (or all content) there are larger percentage differences which shows there are more areas of balance to tackle than this one chart shows. ToMM is a useful example to see how the classes compare when played by top tier IL players, and hopefully also shows that ToMM is complete-able (and has been completed) by all classes in the game.
Changes from this chart compared to more broad charts show a larger positive differential for Arcanist and Warden, brings Blademaster below Assassin with Hellbringer right behind that and brings Arbiter above Dreadnought. You see a wider variance when including a larger selection of content as it adds a far greater percentage of the player base into the damage pool.
There were some paragon paths not listed there, such as Whisperknife, Hunter, and Thaumaturge. Whisperknife and Hunter are not performing where we'd like, and are paths we want to work on, however since those classes have very solid paths as their other choices, that puts the priority a bit further down the list on class work.
First off, the scaling that was talked about for mod 19 sounds great. I'm tired of feeling like all the hard work, upgrades and farming to become BiS is thrown out the window when running scaled content. Players should not be punished for their progression and have to swap out gear depending on the content they run. I hope that this translates well and if so it will be a great implementation and make the experiences an overall better one.
As for the damage performances. No. Just no. I'm not sure where those numbers are drawn from but the majority of the playerbase know how each class performs, what might not had been taken into consideration with those numbers is what contributed to every class performance. There are a number of things that can alter dps at the moment: the broken Darkened Storyteller, Xuna companion, overall weaker players entering Tomm, no name a few. The top end players already finished Tomm and perhaps now we see the second tier of players entering the trial which evens out the dps if you have less experienced wizards running with fully maxed out fighters for an example. We all know that the flavor of this mod are the wizards, even if you aren't particularly knowledgable of the class you will perform well as opposed to other classes. You pick a fighter and expect to be somewhat keeping up with a wizard in dps? Good luck. We know you will stick to your data no matter what but the playerbase will say otherwise on this topic, but all in all the numbers don't mean much without context.
While I am always hesitant to go off topic in a thread as it makes the information about the initial topic hard to find. In this case, at least so far, there doesn't seem to be any confusion about the information in the initial post, so I will briefly tackle the 2 main things being brought up here, but there likely won't be a lot of back and forth on those topics in this particular thread.
Class Balance
We agree that there is an imbalance in the DPS roles. In M19 the two main classes being adjusted for balance are the Dreadnought and the Hellbringer. When M18 comes to preview we will have information on a few adjustments to Arcanist feats/class mechanics to bring them more in line with the balance target (I realize saying this will make people think the worst, but Arcanist will still be powerful after those adjustments and the changes will be available as preview goes live for feedback).
All of that is important context for the following part of that discussion. First is that there will never be perfect balance across the classes, and there will always be some classes that are harder to play and therefore under-perform for a more casual player, but can potentially even over-perform for a particularly skilled player. We have created a lot of analytics on class balance since M16. These include normalized damage charts, whisker plots, and percentile graphs, which we can filter by time ranges, classes, and specific content. We are actively using this information for how to tackle class balance.
As a general point, Assassin in most charts is right around where we feel ideal balance should be right now. Since ToMM was brought up, let's take a look at the PC results from 1 Nov to this morning, and use Assassin as a baseline for where the other classes are at when running that content. I think players may be surprised at where some of the classes line up in this comparison.
Class -> Damage performance in ToMM +/- %
Arcanist +3%
Blademaster +1.8%
Warden +1.5%
Assassin --
Hellbringer -1.6%
Dreadnaught -3%
Arbiter -6%
It is clear there are outliers in Arcanist, Dreadnaught, and Arbiter. The others, however, are all very close together and in general would be considered all within an acceptable range of balance. In charts that include a wider range of content (or all content) there are larger percentage differences which shows there are more areas of balance to tackle than this one chart shows. ToMM is a useful example to see how the classes compare when played by top tier IL players, and hopefully also shows that ToMM is complete-able (and has been completed) by all classes in the game.
Changes from this chart compared to more broad charts show a larger positive differential for Arcanist and Warden, brings Blademaster below Assassin with Hellbringer right behind that and brings Arbiter above Dreadnought. You see a wider variance when including a larger selection of content as it adds a far greater percentage of the player base into the damage pool.
There were some paragon paths not listed there, such as Whisperknife, Hunter, and Thaumaturge. Whisperknife and Hunter are not performing where we'd like, and are paths we want to work on, however since those classes have very solid paths as their other choices, that puts the priority a bit further down the list on class work.
May i ask what the topend difference between the classes is? Im speaking of the highest dps recorded of an arcanist compared to the highest dps recorded of e.g an Arbiter or Dreadnaught? Those numbers dont correspond at all with what i and the people i know see ingame when running Tower of the Mad Mage.
Comments
What if i can prove to you that actually, a cleric can out dps a fighter, warlock and even barb and rogue and on par with rangers?
As regards scaling and caps @noworries#8859 I am curious if you able to offer any further design insight regarding caps where scaling does *not* apply. For example, in the case of LoMM vs ToMM vs MOD18 Dungeon, the gradual increase of critter strength means that a played properly geared for MOD18 Dungeon will, upon returning to LoMM, find herself seriously overcapped. Intuitively, one might assume that gear obtained from "ultimate" content remains superior; in this case, however, being overcapped for content will always be undesirable and lead to worse performance, because the compromises made for those caps (reducing Power, say, or removing % Damage gear) no longer offer any benefit.
Not wishing to ask leading questions here, but is the expectation that players will voluntarily swap to older gear for optimal gear? Or has there been some discussion regarding "softening" caps to make Power-stacking less mandatory?
In the end though, that by itself wouldn't change our plans to bring Arbiter and Dreadnought up to the target range for balance as on average they're both below where we'd like them to be.
The hellbringer class mechanics and magnitudes at peak damage potential don't even come close to base rotation of arcanist, let alone within those numbers in tomm.
Try showing us statistics of successful runs....because those numbers being "normalized" appear to be code for skewed. Take data from hundreds of random group fails and handfull or successful runs, muddle the results together and what you get vs just comparing successful run data and you have vastly different results.
I have been part of all sorts of TOMM runs, from training to speedruns. I can tell you from experience (and have numerous act logs) that as a warden, I cannot even come close to the suggested ~98.5% of arcanist's damage. Have I done it before? Yes, of course. But every time I did I had some sort of an advantage -- better build, less res sickness stacks, being able to position better to maximize dps time etc. When "top item levels and skill levels combine" as you say, the differences are there. And they are much larger than your data suggests.
Thank you for your answer and the insight, I appreciate it.
> (Quote)
> the player base DOES NOT WANT scaling. this is ALL YOU. it is alienating your player base. please just delete it and put it back where it was when you signed on. that's what we want. just make all the dungeons level 80 other than the original levelling dungeons. no one is under level 80 for any significant period of time anyway. it just doesn't make sense.
Scaling has made me realize there is no reason to buy zen or upgrade enhancements and companions. I just run my master expeditions and log off. I will never feel powerful because scaling prevents it.
None of you have ever justified why you are scaling at all. Why? Because it makes content easier for us who spent the Zen/AD/100's or 1000's of hours to get where we are? Stop penalising your most dedicated players, please.
But I know data preparation can have a huge impact in data interpretation. I don't doubt your numbers, but I still have some questions:
With "normalized" you mean you recalculated all data with a z-transformation using the Assassin as a baseline? So the mean Assassin does "0" damage and somebody doing one Assassin standard deviation more damage will get the score "1"? Are those numbers a comparison of the means or medians out of the non-normalized data? Or are those the results of the normalized data? (I would also be interested in the sample sizes.)
Is there a chance to see a whisker plot/box plot?
Those questions are just out of curiosity. I kinda like statistics.
You can make a comparison ONLY with same base data input:
- Same gear (or same amount of %damage bonus);
- Same Power;
- Same amount of buff/debuff active;
- Same amount of revive sicknesses (more complicated);
- Player skills (impossible).
While the first 4 are possible to take into account, it's not possible with the last one. And it's the one with the biggest random oscillation since there are many players with different skills levels.
For the other points it's possible to make a filter to really see the damage output.
As stated from other people in many other posts in this forum, the DPS balance and damage output it's way far from what you're describing.
In italian we say "meglio tardi che mai" which is "late is better than never". It's long time we're asking a better work with scaling, because actually it's a mess in most of the dungeons/skirmishes/trials.
I hope in mod 19 it will be good, since we're waiting 3 mods to have a good scaling.
> (Quote)
> I always enjoy watching top tier players show off their skills, so I would certainly enjoy seeing a cleric performing that well. For clarification, in game wide data the Arbiter does outperform Dreadnought already, with a larger sample size it wouldn't be surprising at all to see them also exceed fighter in ToMM. But it would certainly be fun to watch an Arbiter go toe to toe with a top tier ranger.
>
> In the end though, that by itself wouldn't change our plans to bring Arbiter and Dreadnought up to the target range for balance as on average they're both below where we'd like them to be.
I really like, that you plan to balance Dreadnought (and Arbiter). But I hope that will happen before mod19 (and before mod18). Because mod17 is live. On console it will be live for a long time. And I as a Dreadnought want to have the chance to pass ToMM before mod18 or mod19 will go live. I also want to have the chance to enter a random group for ToMM and I don't want to get refused all time because I am a Dreadnought.
PS: sorry for off-topic. But I have that on my mind.
The imaginary Friends
Main Kingslayer.jr(barb)
> Can you tell us if all dps paragon will have 3 offensive slot. Right now Barb's and gf have 2 less. And tha data that barbarian are 1.2 % behind arcanist in single Target aka tomm runs with 2 offense slot less is not possible. As other are saying this data are way off to the one we see everyday when we run tomm.
I think all dps classes should have 3 offensive slots... When using a defensive tank 3... You should be able to change your teammates depending on the build
I am an endgame barb and from what is my experience every single day those data do not correspond to reality. Arcanists, rangers but also thieves do much more damage for the same objects as my barbarian, especially on bosses. We need adjustments in many parts and it does not seem at all sufficient to me what I have read. I was hoping for a more incisive intervention, I feel that I will be disappointed and I believe that I will stop playing if mod 18 does not solve this huge gap. I really don't understand why you don't listen to what the players of a certain class are telling you, since it is under everyone's eyes.
I wish I could make toom as a dps and feel necessary as an arcanist but instead I am seen as a doll barbie. What's the point of playing again if I can't run the final dungeon as a dps? I've never seen this in any game
Low populations can seriously skew outcomes and conclusions, thus the old adage Lies, Damned Lies and Statistics
Please do not focus only on numbers, but also on what is fun to play. A character with single power that does average damage and has an average cooldown will fulfill your damage requirements but will be horrible to play.
Born of Black Wind: SW Level 80
Bingo! You cannot estimate skill ceiling, so in reality, taking a sample from a bigger population will actually result in a skewed result where the dps of that specific class will be lower. It's obvious that more wiz's are played in TOMM so their average dps will skew to being average whereas generally only the best of non-meta classes are chosen for TOMM. Although, i am a bit skeptical about the exact %performance difference, i do know that skill ceiling non-meta classes can really shock the majority of playerbase if they saw their true potential. We started with 50% dps difference between a barb and wiz but now people have adjusted it to 15-20%. The dev's consider it about 1.2%, which is definitely lower than i would put it at, especially in TOMM.
I do advise the dev's to stick with skill ceiling players only, especially for TOMM and balance ST depending on that. For other dungeons, the bigger the sample the better, in my opinion.
During that 4 day sample size of ToMM there was probably 1000 Arcanist....of which 80% were likely bad players but were invited simply because they are wizards. Where as the dreadnaughts, Warlock's, arbiters, and blademasters probably accounted for 25 total players....all of who were invited because they are OUTSTANDING players.
but i have a simple question: what is the difference, in damage, between tanks/healers and that dps?
is possible decrease that a little?
First off, the scaling that was talked about for mod 19 sounds great. I'm tired of feeling like all the hard work, upgrades and farming to become BiS is thrown out the window when running scaled content. Players should not be punished for their progression and have to swap out gear depending on the content they run. I hope that this translates well and if so it will be a great implementation and make the experiences an overall better one.
As for the damage performances. No. Just no. I'm not sure where those numbers are drawn from but the majority of the playerbase know how each class performs, what might not had been taken into consideration with those numbers is what contributed to every class performance. There are a number of things that can alter dps at the moment: the broken Darkened Storyteller, Xuna companion, overall weaker players entering Tomm, no name a few. The top end players already finished Tomm and perhaps now we see the second tier of players entering the trial which evens out the dps if you have less experienced wizards running with fully maxed out fighters for an example. We all know that the flavor of this mod are the wizards, even if you aren't particularly knowledgable of the class you will perform well as opposed to other classes. You pick a fighter and expect to be somewhat keeping up with a wizard in dps? Good luck. We know you will stick to your data no matter what but the playerbase will say otherwise on this topic, but all in all the numbers don't mean much without context.