How we determine balance and its flaws.

[Vendalwind]

Disclaimer. I may like certain Classes. I don't deny my bias and i know many external factors effect balance (mobility, defensive and healing utility etc.) This isn't some end all statement.

My conclusion: The maximum rdps output of a utility rdps class should be higher than that of a selfish dps class for all classes to be desired. Note I'm talking 90th percentile and up performances.

Constantly on these forums i see people suggesting selfish dps should output the most Rdps. Please take significant notice of what I say here as Rdps is not Personal dps.

I want to present some simple math for this discussion. I've provided it before but never in an unheated or abstracted environment. I've also done much more complex calculations on this but i am trying to keep things as easy to follow as possible.

Imagine a party of 8 players
With MAXIMUM dmg as the values below.
Each dps deals about 14000
Each tank deals about 9000
Each healer deals about 6000
Total damage 86000

*side note reducing tank and healer contribution also reduces the effectiveness of utility dmg dps classes

Now lets add dps specific utility to some of those classes while reducing damage.

Also we must accept that by adding utility dps the difficulty the entire group handles increases marginally as they try to time boost placements. AND the utility rdps boost is reliant on all 8 members performing at maximum efficiency.
I choose 1% for simplicity buff to start, but also because much higher and balance for 24 man and 4 man parties gets disrupted more seriously.

3 dps deal 14000
1 deals 13000 +1% total group dmg (14x3+13+18+12) +850= 13850
Total group dps=85850

Already its obvious that bringing the utility dps at this rate is a loss.

This is aggravated by adding more supports ignoring compounding bonuses (though likely those compounding don't make up the loss anyways as they also normally amplify by a factor no more than 10% to the boosted dmg) as each additional support not only loses more standard dps but takes away from the Rdps of the other supports

2 deal 14000
2 deal 13000+(84000x.01)=13840
Total group = 85680

Considering a possible compound bonus to the buffed dmg of 10% an extra 84×2 at maximum is added
=85848 still slightly worse than 1 support.

With balance as above any Rdps class lowers the effectiveness of the group.

In the above balance scenario the most optimal group AND the most reliable dps is the pure selfish personal DPS units.

So how do we make it so both selfish dps and raid utility dps are desired?

There are a few paths forward all have other adverse effects on game balance. Optimally we want a scenario where 1 selfish and 3 utiliy dps yield the best results by a mild margin over 2 and 2 Why? Because that way the raid dps NEED at least one selfish dps to see their maximum, BUT that maximum is harder to obtain. (More reliant on coordination) and there are less selfish dps than there are dps with dmg utility in the game. Groups that run 4 selfish dps would still be very viable and the most dependable in pugs because lack of coordinated timing will effect damage output minimally.

So we would need to boost the buff while reducing base dps of the raid utility dps classes until we found that optimal point of balance. Alternatively like said above reducing the contribution of healer and tank also can reach the optimal point. Both options have adverse effect tho. Option 1 would make 4 mans and 24 mans less balanced as the buffs do more or less and would make solo play for utility dps classes more difficult. Option 2 would make solo play for healers and tanks more boring and difficult.

I found that optimal point using an excel spreadsheet and goal seek once and will within the week use it to find it again (but i am on my phone at the moment doing napkin math versions of what i had done months ago) just to show what I'm talking about has a solution and to invite discussion.

I would in the meantime invite people to consider this and do the math themselves whether in excel or by hand (goal seek in excel is much better than by hand btw) but my point is this: selfish dps should not do the most Rdps at high skill percentiles. If they do all utility dps classes will be undesirable across all skill levels

The game optimally should be balanced in a way that at high percentile play 3 utility dps and one selfish is optimal and as skill levels decrease more selfish dps perform the best rdps. Where that percentile line falls where 4 selfish should out do having a support tho is a matter of opinion. This is not to exclude them (selfish dps focused classes) from playing in meta content but to include as best as possible every class in meta content.

I frequently hear people ask for #### selfish class to be buffed so its rdps is the highest within its group. This in fact seems to be the normal opinion. And my point is that is not actually healthy for the balance of the game. Asking for a selfish dps to be buffed for other various reasons is fine. But selfish dps do not need to be the top RDPS and asking for it as if its the expected norm should not be accepted as a valid reason.

[Vendalwind]

Do note that if you extend my initial example to 4 dps with utility their compounding buffs outdo the 4 selfish dps scenario. This is intentional to again show a bad side of this balance where mixed parties are the worst. This shouldn't happen either and the solutions i propose above can in fact yield a scenario where even including compounding buffs 4 support dps is not as optimal as 2 and 2. Just gotta play with the numbers enough.

This thread also was intended to combat people that constantly complain about balance
Balance is a very very fluid and living thing in a game like this, especially as sub stats scale differently for different classes in terms of effectiveness cycles into each patch tier (crit becomes god, but some classes get even more added benefit from crit and group crit)
The above is just one of the reasons, and one of the things developers consider and math out constantly to theoretically balance a game.
And doing this while maintaining class diversity without homogenizing? Is real real tough. Put on top of that making the class fun to play and raid fights balanced to the classes mechanics... it gets complicated.

Personally I think my Main class deserves a bit more rDPS within melee tier (Ninja) BUT i wouldn't even ask for it right now. Because I am quite happy with how close to balanced regardless the game currently is with the only current minor flaw in my eyes being ranged lagging jsut a bit too far behind. And of course wanting a class changed for fun purposes or playability purposes (like how monk still needs to be changed and many Ninjas like myself are unhappy with the shift away from blade play) is a whole different beast, and when changes like that are done the above damage output balancing has to be reconsidered as well. Thats my opinion and my bias, but i hope you can abstract that bias and see it separately from my initial post.

Thanks if you manage to muscle through all of this.

[Kalise]

Optimally we want a scenario where 1 selfish and 3 utiliy dps yield the best results by a mild margin over 2 and 2

So... You basically want StB balance back? Where every raid was Bard + Dragoon + Ninja + [Selfish DPS] (Or often, SMN because they did pDPS equivalent to that of selfish DPS but also had utility and rDPS boosts too... Like they still do!)

The "Ideal" scenario is where any 4 jobs can be used together and be close enough to not matter in terms of DPS.

This means, 4 selfish DPS = 3 selfish DPS + 1 support = 2 selfish DPS + 2 support = 1 selfish DPS + 3 support = 4 support.

Why? Since that means that people can play the job that they like the most without having to adjust based on what other people in their raid are playing/want to play.

It however, can be tricky to balance around since the more support DPS that are in a party, the bigger factor that Tank + Healer DPS makes, which can be unhealthy for the overall game balance (See: Healers complaining about being "Green DPS" instead of Healers)

Furthermore, there's something to be said about smaller groups too. Dungeons exist and form a significant part of the game (Due to MSQ as well as Tomestone/Materia grinding). The more you lower support jobs pDPS in favour of making it up with rDPS in 8 man parties, the worse they perform in 4 man content (As well as in PuG content where you find duplicate jobs whose rDPS boosts won't stack)

Such additional aspects to job balancing are why ShB has skewed more towards removing much of the rDPS boosts that existed in the game (Though, many still exist and with some still being on "Selfish DPS" jobs like SMN, DRG, MNK and RDM not just the typical "Support" jobs like BRD, DNC, NIN... Though, this largely depends on your definition of "Support" vs "Selfish" and thus whether you think that only BLM, MCH and SAM are "Selfish" DPS due to their complete lack of utility/rDPS boosts)

[Vendalwind]

So... You basically want StB balance back? Where every raid was Bard + Dragoon + Ninja + [Selfish DPS] (Or often, SMN because they did pDPS equivalent to that of selfish DPS but also had utility and rDPS boosts too... Like they still do!)

No.

The "Ideal" scenario is where any 4 jobs can be used together and be close enough to not matter in terms of DPS.

Yes, but with caveats being that proper group diverse composition yields some rewards and having all roles also yields rewards. Last patch for example if not for the role bonuses and duplicate class reductions a team of 4 SMNS probs would have been the best and safest to play in any fight. Thats unhealthy. A meta will always exist. but the nature of the meta and the balance as i suggest above will yield your ideal scenario most effectively.

You assumed much beyond the scope of my post. I spoke of maximized dps. not viability.

[Kalise]

No.

Well, that was the "Optimal DPS is 1 selfish + 3 support"

Where max DPS was using BLM or SAM plus the aforementioned 3 supports, it just required a really good BLM/SAM to work and was far easier to do near enough DPS with SMN instead (While having an extra Swiftcast Raise to boot)

A meta will always exist. but the nature of the meta and the balance as i suggest above will yield your ideal scenario most effectively.

It won't. Your way of balancing will push people towards the aforementioned StB meta, where people will use the comp with the most bang for the buck, not the max DPS one but the one that provides the highest DPS without requiring as much skill (With the max DPS comps being used only by those who do speed clears)

You assumed much beyond the scope of my post. I spoke of maximized dps. not viability.

Maximized DPS and viability go hand in hand. Since viability is determined by the players on which jobs fit into the "Meta" which is determined by maximized DPS. For example, situations where people would decline BLM's, RDM's, MCH's and SAM's for raids during StB because they weren't "Good Enough" (It wasn't universal, but it did happen). Could you clear all content with non-"Meta" jobs? Yes. But some people didn't care about that, in favour of "Max DPS".

[Vendalwind]

[

Well, that was the "Optimal DPS is 1 selfish + 3 support"

Where max DPS was using BLM or SAM plus the aforementioned 3 supports, it just required a really good BLM/SAM to work and was far easier to do near enough DPS with SMN instead (While having an extra Swiftcast Raise to boot)

Read what I said again. I will not continue this conversation if you try to assume things tangential to my post.
I said 1 selfish 3 support
IE
1(SAM,BLM,MCH) 3 (ANY COMBINATION OF EVERYTHING ELSE of which there are 7) So proportionally. 1 or 2 slots to the selfish dps that make up 30% of the dps is fair with 1 slot being the closest mark.
and to quote myself
"This is not to exclude them (selfish dps focused classes) from playing in meta content but to include as best as possible every class in meta content."
"Disclaimer. I may like certain Classes. I don't deny my bias and i know many external factors effect balance (mobility, defensive and healing utility etc.) This isn't some end all statement."
of course there are other factors like the ridiculousness caused by tank stances and shadewalker.

I am not saying this because i want ninja buffed. Please do what I asked:"abstract that bias and see it separately from my initial post." or at least the numbers in my post.

I also talked about this ONLY looking at the damage to conclude ONLY one thing. That selfish dps should not in a vacuum contribute the most rDPS. if that occurs mathematically they will be the most chosen class when people pick classes with damage in mind. especially since most utility classes literally offer only utility in the form of... more dmg. because consequently if selfish DPS have the best RDPS at max percentile. it follows immediately that they will have the best dps at ANY percentile. while that does not hold true mathematically when an rDPS focused class has the highest rDPS contribution at maximum percentile.

It won't. Your way of balancing will push people towards the aforementioned StB meta, where people will use the comp with the most bang for the buck, not the max DPS one but the one that provides the highest DPS without requiring as much skill (With the max DPS comps being used only by those who do speed clears)

Why wont it? Prove it then. Give me some numbers that say otherwise. You are guessing at something without any backing to your statement just because you disagree with some portion of mine. Why does what i say irk you so? This is exactly what I'm talking about People blindly making statements that just don't actually have any bearing. I left things un mathed completely to encourage people to math things for themselves. go try to prove your point.

Maximized DPS and viability go hand in hand. Since viability is determined by the players on which jobs fit into the "Meta" which is determined by maximized DPS. For example, situations where people would decline BLM's, RDM's, MCH's and SAM's for raids during StB because they weren't "Good Enough" (It wasn't universal, but it did happen). Could you clear all content with non-"Meta" jobs? Yes. But some people didn't care about that, in favour of "Max DPS".

That's just entirely incorrect. Viability and Meta are absolutely separate Ideas. Do you know what Meta means?
Maximized DPS is talking about a team of all 100 percentile players. Viability generally includes teams of grey parsing 25th or lower players. When i talk about perfect class play you have to look at the numbers. And when i talk about balancing you have to look at the effect skill levels have on the changes of value rDPS focused classes provide.

[Kalise]

Read what I said again. I will not continue this conversation if you try to assume things tangential to my post.

It is not assumed. This is fact. This is what it was like in Stormblood.

The max DPS WAS 1 selfish and 3 support. But people had better overall success with 4 support because it was less skill reliant (Since, co-ordinating is easy, you just push your 60s/90s CD's every 60s/90s. Optimizing the selfish DPS's uptime especially during burst windows was hard and ensuring that said selfish player NEVER made any mistakes in the content because that would mean death and weakness from a raise)

It just so happens that due to the way that the different supports provided support in different ways, it ended up being the jobs that synergized with each others supports that became meta. I.e. Bard was the best rDPS job, so you wanted one of them and DRG had Piercing buff that helped the Bard's pDPS and then NIN was always there because Trick Attack was the most broken skill in the game.

At best you occasionally saw speed clears swap the DRG for a SMN for the extra magical damage debuff to support the BLM

I said 1 selfish 3 support
IE
1(SAM,BLM,MCH) 3 (ANY COMBINATION OF EVERYTHING ELSE of which there are 7) So proportionally. 1 or 2 slots to the selfish dps that make up 30% of the dps is fair with 1 slot being the closest mark.
and to quote myself

But you have to remember that not all "Support" jobs support in equivalent ways.

In no world would you want to run a BLM with MNK, RDM and DNC as your supports for example. Since MNK and RDM only buff PHYSICAL damage while DNC's pDPS relies on using many weaponskills/spells which BLM's slow casts do not fit well with.

Hence, we would go back to a meta of the jobs that synergize best, much like the old DRG + BRD + NIN combo.

I also talked about this ONLY looking at the damage to conclude ONLY one thing. That selfish dps should not in a vacuum contribute the most rDPS

This particular point does not need to look at compositions at all though.

It's common sense that jobs that sacrifice pDPS for rDPS should not thus have to deal less overall rDPS than jobs who's rDPS is equal to their pDPS.

The only time composition would matter for such argument, is when you try and look at what rDPS is being provided by the support jobs depending on the composition.

I.e. If a "Selfish" DPS does 15000 DPS and a "Support" DPS does 13000 DPS and buffs everyone else by 10% with Tanks that deal 10000 DPS and healers that deal 6000 DPS

Then a Selfish DPS's rDPS is simply a flat 15000.

However, a support DPS's rDPS can range from 20700 (With 3x Selfish DPS) to 20100 (With 3x Support) with this range increasing with a larger gap between selfish and support pDPS as well as increases in the boost they provide to other party members.

Though, it is not always the case that jobs lose pDPS for rDPS, as people have noted jobs like RDM have had pDPS penalties for utility such as "Vercure/Verraise" despite their lack of actual worth in most parties.

Why wont it? Prove it then.

Look at every expansion prior to ShB.

That's just entirely incorrect. Viability and Meta are absolutely separate Ideas. Do you know what Meta means?

Viability and Meta are 99.99% of cases the same thing. Outside of some terrible balancing that makes a job literally unable to perform in content (I.e. ARR Warrior before it got buffed).

Viability is being able to complete content.

Meta is the player controlled "Ideal" composition. Which, some players, will hold as a set in stone viability scale. If a job is not meta, it is not considered viable by these people and thus would get excluded from parties (Even if it is still possible to clear the content using said non-meta job)

Maximized DPS is talking about a team of all 100 percentile players.

Maximized DPS is only compositions.

Since, all talk of balance tends to use a situation where players play perfectly. Since that is the only tangible data to use. It is impossible to calculate DPS losses from "Lower skill levels" and apply them to data to get meaningful results.

This is where the divergence of "Max DPS" compositions and "Meta" compositions comes in. In perfect scenarios, there will be some compositions that will push out the highest numbers. However, these don't usually become meta due to the vast majority of people's inability to perform perfectly. Even the highest skill players will often shy away from the Max DPS compositions until they've practiced the encounters hundreds of times, instead favouring the "Meta" compositions that are easier to play and deal good numbers with.

And when i talk about balancing you have to look at the effect skill levels have on the changes of value rDPS focused classes provide.

Skill levels affect the value of pDPS to the same extent or more than rDPS. Selfish classes often do far worse than support classes at lower skill levels, due to selfish classes having to perform all of their DPS themselves. While support classes still get a portion of their rDPS done even if they themselves are not performing well (So long as you can "Co-ordinate" pressing your CD's on CD)

It is why jobs that can do "Max DPS" are often not widely used (Typically, this is seen with BLM which requires intimate knowledge of an encounter to be able to position correctly so as to have good uptime. Without this, a BLM's damage can become absolute garbage)