Why is balancing apparently so difficult?

[lappee]

What are you even talking about. If there are 5 rogues who do 10, 11, 12, 13, 14k dps respectively, and rogue gets slapped with a -10% damage modifier (that is they do 90% of previous damage) then the new numbers will be 9, 9.9, 10.8, 11.7, 12.6k dps.

And if you take 5 mages who do 8, 9.5, 11, 12.5, 14k dps respectively you will only achieve "balance" at a certain point. Slapping a modifier only changes the point where they're equal (or grants a point where they are).

[Freedom]

Have you ever tried keeping 36 plates spinning on sticks at the same time, while also making sure the food on them doesn't fall off?

Didn't think so. (Every spec is a plate, and the food is PVP versus PVE).

[MooCowCat]

I think one thing blizzard could do is add an overall cap just so the guys who figure out crazy things that get there dps running real high compared to others or if a class is over tuned as well. more like a security cap but really most people wouldnt hit it unless it was the best players ect. Yes theat takes away from top parses lol but for those who dont care about that i guess it wouldnt matter to them.

[Lucetia]

This is solved pretty easily, actually.

You have a passive that, for example, increases all damage you do by 50%. Then if you want to nerf the class's damage by 10% you just make it increase the damage you do by 35% instead. No need to think about any sort of fancy interactions.

Except something like that poses just as many problems. By buffing with a flat passive by 50% you then run into the interactions from all the abilities. It could make ability more potent causing it to increase even further. Then you have trinket, artifacts, etc that interact throwing an ability even further ahead. So when you do 50% you could really be looking at 75% overall, or 63%, etc.

That actually is more complicated than it should be - if you really wanted the simplest way to nerf a spec by 10%, just slap a passive on that reduces damage done by 10% across the board for the spec, it doesn't matter what other modifiers the spec had previously.

As I mentioned above, something like that wouldn't work. It has a lot of side effects. It sounds like a simple fix, but if you did that then some abilities under perform causing other abilities to perform even worse that might rely on it and then item interactions.

Is it easier to develop a new spell instead of just having one ready to go on every single spec that you can modify as you please?

Of course, you can just have that spell increase damage done by 0% as a baseline and then change that. The point was that you should nerf the damage along the board with one single lever, not every single spell one by one, that makes no sense. That should be reserved for when one spell deals too much / little damage in comparison to other spells of the class. For example in AoE vs ST, or if a spell isn't worth casting when it should be.

Except one single lever is the worst ting possible in this case. This also takes into account everyone has the exact same talents, same rotation, same gear, same artifact talents, etc.

Such as? If you put a passive that reduces *all damage* by 10% after all other calculations have resolved, you will get -10%, exactly on the dot. If not, you missed the entire point of reducing 10% across the board.

Except you don't. Math isn't so simple in terms of game balancing. Someone with a lot more time and does this a lot could explain it in a bit more detail than I could.

If every class had 1 set of gear, 1 defined rotation, no talents, no rng, nothing that makes them different from another player then it could work. There are just too many variables at play that throws it completely off balance.

Luckily for us, there's no such thing as hardcoded base values like it was in previous expansions. Your base damage is utterly decided by ap/sp.

No, a lot abilities have a small flat damage portion and then a +X% AP/SP.

Overall while your ideas sound simplistic to do, the sad reality is that there is way too much going on in the background for it to be possible. If they truly wanted to do it then they could, but everyone would have exactly the same gear, talent, rng, etc. Which would destroy the whole point of an RPG like system. There'd be no point to do content more than once since you can't ever get different gear, etc.

[BannedForViews]

You're assuming things scale similarly for both. That is not the case in WoW.


Indeed, so nothing actually changes apart from the ilvl where Fire mage is equal to some other specs. They still scale differently thus you're not solving anything.

To be clear, I am referring to things that give a flat damage bonus, i.e. main stats/versatility. If you are implying that if a rogue goes from 5000 to 6000 agi and a hunter does likewise, that the rogue will somehow gain twice as much damage boost from this over the hunter, I don't buy it for a second. YOU would have to prove that, and if that were the case, then these people we pay a lot of money to are really fucking stupid. Now I will concede that crit/mastery can have more impact on different classes/specs, so as they gain more of that you will see a geometric increase to a degree, but this method of fixing things is still better than nothing. If fire mage does 250k in T19 and 300k in T20, but rogue does 200k and 210k, then rogue is still better off by giving them a 25% damage buff across the board so they do 250k and 262.5k. Even thought the problem persists again at T20, it is much better than the alternative. while other things are fixed (changing gear stats around for either rogues or mages, etc). or they could give rogues a 40% damage buff and make them do 294k at T20 and be overtuned for non-current content, that won't really bother anyone.

The reality though should be that most of the 36 classes will respond very well to a tuning knob like this. Don't harp on the outliers, they can be fixed in other ways.

- - - Updated - - -

And if you take 5 mages who do 8, 9.5, 11, 12.5, 14k dps respectively you will only achieve "balance" at a certain point. Slapping a modifier only changes the point where they're equal (or grants a point where they are).

What do you mean balance? In my example it was implied they did different damage because of user skill, whereby 14k was closest to optimal play. All in the class get helped/hurt the same way. If you are playing an OP class but can't do OP damage, you can't cry about a nerf. hell they are going through every step they can to make the rotations simple enough for a casual 9 year old.

- - - Updated - - -

Except you don't. Math isn't so simple in terms of game balancing. Someone with a lot more time and does this a lot could explain it in a bit more detail than I could.

If every class had 1 set of gear, 1 defined rotation, no talents, no rng, nothing that makes them different from another player then it could work. There are just too many variables at play that throws it completely off balance.



No, a lot abilities have a small flat damage portion and then a +X% AP/SP.

I'm sorry but you'll have to give a concrete example because this isn't correct. if you look at my parse and I'm doing 30k dps, and you reduce every skills damage by 10%, I will do 27k dps. it is that simple. multiplying by a constant is linear.

[lappee]

To be clear, I am referring to things that give a flat damage bonus, i.e. main stats/versatility. If you are implying that if a rogue goes from 5000 to 6000 agi and a hunter does likewise, that the rogue will somehow gain twice as much damage boost from this over the hunter, I don't buy it for a second. YOU would have to prove that, and if that were the case, then these people we pay a lot of money to are really fucking stupid. Now I will concede that crit/mastery can have more impact on different classes/specs, so as they gain more of that you will see a geometric increase to a degree, but this method of fixing things is still better than nothing. If fire mage does 250k in T19 and 300k in T20, but rogue does 200k and 210k, then rogue is still better off by giving them a 25% damage buff across the board so they do 250k and 262.5k. Even thought the problem persists again at T20, it is much better than the alternative. while other things are fixed (changing gear stats around for either rogues or mages, etc). or they could give rogues a 40% damage buff and make them do 294k at T20 and be overtuned for non-current content, that won't really bother anyone.

The reality though should be that most of the 36 classes will respond very well to a tuning knob like this. Don't harp on the outliers, they can be fixed in other ways.

Twice as much? no, highly unlikely. Will they gain different amount? yes, very likely. You can check simulation that gives a value to your stats and they will be different.

Giving a buff to that rogue indeed would solve its problem when it has that specific T19 gear but it would screw the balance in its favour at lower ilvl's.
As I said earlier, you would have to give different modifier at different gear sets - it will never happen.
Not to even mention we're (likely?) talking just single target dps here, what happens when you take all different kind of cleave and AoE situations in consideration as well?

- - - Updated - - -

What do you mean balance? In my example it was implied they did different damage because of user skill, whereby 14k was closest to optimal play. All in the class get helped/hurt the same way. If you are playing an OP class but can't do OP damage, you can't cry about a nerf. hell they are going through every step they can to make the rotations simple enough for a casual 9 year old.

Why would you even make an example with user skill? makes no sense. You should make the example with different ilvl's - which is the way I responded.

[Deleted]

So, we all know how it goes, every single time Blizzard changes something about the classes, we end up with wildly different dps numbers across all classes

I don't think there's any evidence that we end up with wild difference in numbers, especially if you compare performance in same performance bracket with similar sample sizes.

The variance between classes isn't huge and it's completely and utterly inconsequential unless you're in the race for the 1st.

[BannedForViews]

Twice as much? no, highly unlikely. Will they gain different amount? yes, very likely. You can check simulation that gives a value to your stats and they will be different.

Giving a buff to that rogue indeed would solve its problem when it has that specific T19 gear but it would screw the balance in its favour at lower ilvl's.
As I said earlier, you would have to give different modifier at different gear sets - it will never happen.
Not to even mention we're (likely?) talking just single target dps here, what happens when you take all different kind of cleave and AoE situations in consideration as well?

- - - Updated - - -


Why would you even make an example with user skill? makes no sense. You should make the example with different ilvl's - which is the way I responded.

I'll say it one last time. Are there values which change dps in a nonlinear way? Yes. Fire mage is a prime example, crit giving it more procs of a more powerful spell, altering the rotation in a way to that changes the overall dps of the spec. First, I think that is the exception not the norm. Most specs are going to have the same rotations, or similar rotations, through all gear levels, assuming they don't change base values (which they completely will). A modifier adjustment will PROBABLY be all that is required for MOST specs. It will also AT LEAST bring the specs that scale more exponentially closer to the ones that lag behind it. I am not, nor have I suggested, that all dps should be the same, and it won't nor has it ever been that way in WoW. They should be similar though, say within 5 or even 10% of each other. This is a way to narrow gaps, not a fix for perfect "balance" whatever that may be.

And yes, it is mostly in reference to single target. "Rotations" are really only used for single target fights. AoEs are generally 1, maybe 2 spells that you spam. If there is imbalance there it is probably easier to just change that skills actual value.

I'll repeat, I don't think it is as simple as the OP made it, but the way a modifier will work on any given fight at a fixed gear level is most certainly predictable. If I do 200k single and you do 250k, but I do 400k on a 5 mob AoE and you do 600k, buffing my dps by 25% will always have me doing 250k ST and 500k AoE. AoE is significantly harder to balance anyway, and not something that people get nearly as hung up on. Rogues are pure shit at it for example, but they'll still get a raid spot.

I don't think that everyone should be the same. That sounds boring, but I do think that everyone should be able to do very similar damage so as to not alienate classes/specs that people enjoy playing just because a raid leader thinks it will help them down a boss.

[lappee]

I'll repeat, I don't think it is as simple as the OP made it, but the way a modifier will work on any given fight at a fixed gear level is most certainly predictable. If I do 200k single and you do 250k, but I do 400k on a 5 mob AoE and you do 600k, buffing my dps by 25% will always have me doing 250k ST and 500k AoE. AoE is significantly harder to balance anyway, and not something that people get nearly as hung up on. Rogues are pure shit at it for example, but they'll still get a raid spot.

I don't think that everyone should be the same. That sounds boring, but I do think that everyone should be able to do very similar damage so as to not alienate classes/specs that people enjoy playing just because a raid leader thinks it will help them down a boss.

No one has denied that the goal should be close to equal dps. The problem with modifier proposal still exists - it will only work at a certain ilvl. If you only care about that certain ilvl, then great for you. However if you think of the larger scale of things, the whole player base, then you still havent fixed anything at all.

[Hctaz]

Sorry, that's mathematically impossible, unless past a certain threshold there's a hardcoded mechanic that modifies your modifiers to do differently before and after said threshold.

You are basically saying that, before stats reach x, they are multiplied by y, and after x, they are multiplied by z. Which is impossible, unless specifically hardcoded into the spec itself.

That's not how that works at all. We have secondary stats for a reason and the synergies between those make the ilvl gains not as linear as you'd like. Something like Frost dk who gains a lot of benefit from Crit in the beginning but then tapers off as it gets into the realm of "Now I have too many Killing Machine procs than I can spend and that's all crit is good for," so there's a point where crit no longer gives you that same increase in damage. Suddenly haste might take over but who knows how good haste will be in comparison to early levels of crit. Like take this hypothetical situation: Gaining crit chance at 0-15% chance to crit gives frost dks a 2.3% increase in damage per 1% crit chance. Then getting crit in the 15-25% range is only a 1.9% damage increase per percent and so on. Now you scale scale less with crit and you can't just give them a flat 5% damage increase to compensate because then frost dks in the 0-15% chance to crit range are getting stronger per point of gear than is intended.

[Sir Andy]

Because it's hard to balance when you can't ever make mages bad, but anything else is on the table.

[Dormie]

Blizz isn't THAT bad at class balance, they just need to make changes more often. Sometimes there are blatant problems with a class that go unaddressed for way too long.

[D3cadent]

No one has denied that the goal should be close to equal dps. The problem with modifier proposal still exists - it will only work at a certain ilvl. If you only care about that certain ilvl, then great for you. However if you think of the larger scale of things, the whole player base, then you still havent fixed anything at all.

This is why the threshold question is probably "What part of the game do you balance for?".

As much as some people don't want to hear this, personally I think they should just balance for the absolute top end of PvP and PvE, as everything before those points is made improved by playing better, getting more gear or a combination of both. If they take this stance, and balance tier to tier, it becomes a much easier exercise.

(Then, while they're at it, they could remove huge RNG aspects of the DPS game, like the RNG of Soul Cap procs during shadowdance vs outside shadowdance, Arcane Mages in WoD etc etc... but that's a pipe dream, they're adding more RNG!! QQ)

[lappee]

This is why the threshold question is probably "What part of the game do you balance for?".

As much as some people don't want to hear this, personally I think they should just balance for the absolute top end of PvP and PvE, as everything before those points is made improved by playing better, getting more gear or a combination of both. If they take this stance, and balance tier to tier, it becomes a much easier exercise.

And balancing for the top end PvE (PvP is already being taken care of) means balancing it at the top gear (+ couple ilvl's) of previous raid (or entry level gear with new expansions) as progression is what really matters in PvE.

[Deleted]

It has been shown for so many years already that Blizzard WoW developers don't really care all that much about fine tuning balance between different classes. 10% and higher differences between DPS specs across different fight types is already unacceptable, but it's really difficult to balance them because ilvl and gear stats/set bonuses alter the DPS so much that you can't just slap a flat % nerf on a certain spec because it will have a higher/lower impact depending on stat scaling. At least in Legion they don't have to worry about balancing PvP at the same time.

In order to keep things balanced they need a to have a dedicated team constantly monitoring class/spec performance and pushing out adjustments at a fast rate (and preferably also taking high end theorycrafter feedback actually into account...). Only doing balance changes in major patches and a few hotfixes inbetween just doesn't work, this is how we end up with crazy 30% or higher differences between top and bottom specs.

[Deleted]

It has been shown for so many years already that Blizzard WoW developers don't really care all that much about fine tuning balance between different classes. 10% and higher differences between DPS specs across different fight types is already unacceptable, but it's really difficult to balance them because ilvl and gear stats/set bonuses alter the DPS so much that you can't just slap a flat % nerf on a certain spec because it will have a higher/lower impact depending on stat scaling. At least in Legion they don't have to worry about balancing PvP at the same time.

In order to keep things balanced they need a to have a dedicated team constantly monitoring class/spec performance and pushing out adjustments at a fast rate (and preferably also taking high end theorycrafter feedback actually into account...). Only doing balance changes in major patches and a few hotfixes inbetween just doesn't work, this is how we end up with crazy 30% or higher differences between top and bottom specs.

Yeah I agree. Anyone who can add 1+1 could join them and with little ambition just balance the game. This would even make Lappe happy who nonstop envies how his mate got 2 titanforged items and got a lot more value out of the pure upgrade number value than he when he finally replaced the same slots and with same items & itemlevel jump.

NOIR was always meaningless.

[Notshauna]

Balancing WoW isn't just difficult, it's virtually impossible. You have to use the same base damage and the same scaling to try and balance out every differential. Imagine trying to balance all the DPS classes with their different abilities and rotations so they are equally viable with A) every gear level B) every talent and artifact combination C) in every encounter, beit player, patchwerk, cleave, AoE, Burst, council, high mobility, and everything in between and D) also make them reasonably balanced in terms of non-simplistic matters such as differing player skill levels, utility, and try and predict the changes that will come in the future that will shake up the balance. And even if you could do that I'm sure I missed some niche situation that will break the balance so all that work would ultimately not result in balance. But, the truth is true balance isn't even a goal, instead the goal is to make it so everyone is close enough.

[Smauldy]

Blizzard does a very impressive job balancing classes. If you've ever played in older MMOs, you'd really appreciate the job they do.

[Deathquoi]

Lets say we have 1 skill, and my 1 skill hits for 1000 and yours hits for 1100 at ilvl 850 lets say. Assuming that an ilvl is a 1% buff, that means at ilvl 900 I should do 1645 and you 1810. It should be a 1% buff for you as well as for me. It may in fact not be, but as Pospospos stated, the only way that mathematically makes sense is if it is hardcoded into skills to do more or less damage at different ilvls, i.e. not actually get buffed by the 1% it should have been buffed by. Now that may be the case, but it makes a hell of a lot more sense for it not to be and needlessly complicated if it is. These things should all be pretty linear.

Now you are right that it could throw things off at lower levels, but they could just as easily give the buff/debuff after level x, for example.

- - - Updated - - -



a damage buff IS a linear scaler, that is the point. Yes, fire mages get more from crit, others from haste or mastery, blah blah, but a pure damage buff or debuff is indeed linear. It will have no affect on those other mechanics. Fire mages will still get just as many procs as before, they will all just do uniformly more/less damage.

You're not getting it.

Here's some made up numbers to illustrate my point.

Say Fire mages at 700 ilvl are doing about 10,000 DPS. Obviously this is a made up number, but whatever, it doesn't matter for the purposes of this example. Say every increase of a point in ilvl is giving them +100 crit, increasing their DPS by 10. At 800 ilvl, they'll be doing 11000 DPS. At 900 ilvl they'll be doing 12000 DPS.

Now say Frost mages are doing about 9,500 DPS. They're not as fortunate as Fire mages, they don't scale as well with haste as fire does with crit, so we'll say that every ilvl only bumps them up 5 DPS. That means at 800 ilvl they'd be doing 10,000 DPS and at 900 ilvl they'd be doing 10,500 DPS.

Say Blizzard doesn't like how low Frost's DPS is, so they bump it up by 5%. Now, at 700 ilvl they're doing 9,975 DPS, so cool, they're competitive. But they haven't fixed the difference in how the two specs scale with their preferred stats, so at 800 ilvl Frost is doing 10,475 DPS and at 900 it's doing 10,975 DPS. A flat damage increase to Frost narrowed the gap at 700, but by 900 it's back to being way behind fire.

This is an extremely simplified example but that's why it's not as simple as just buffing things with a flat percentage. Your assumption that an increase of 1 ilvl benefits everyone by the same percentage is where your argument is flawed. Some specs get more of a benefit from an increase in ilvl than others.

The relationship between specs and stats is not linear, and it is not nearly the same between all classes. The only way to make it linear is to turn every ability into Frostbolt, a straight damage ability with no interesting mechanics, and the only way to make the relationship between stats and damage the same for every spec is to homogenize them all to the point where having individual classes is completely pointless because they're all the same.

[Pull My Finger]

I would say it not only is difficult, I'd call it practically impossible in a system with so many variables. It just becomes incredibly complex, having so many specs with so many asymmetrical abilities, effects and unique encounter situations. They're doing a pretty good job. Anything in the +/- 10% range should be viewed as "more or less balanced". You can't really do better unless you homogenize classes to the point where they become purely cosmetic choices.