My theory to balancing AOC
Voxtrium
Member
Balancing Ashes of Creation
https://ashesofcreation.wiki/Roles “ 1v1 matchups will have a rock-paper-scissors dynamic, where one class will be superior to another.
Assumption - placeholder number for Total available spec points per class = 150
Every level gained will give the player 2 points for specs until level 25, from level 25-50 each level will give the player 4 spec points for a total of 150.
• Tank will lose to Support
• Support will lose to DPS
• DPS will lose to tank
Each of the 8 classes will need to be categorized into its trinity role within its class. Each class would have spec trees aligned to their gameplay type. However, each class would have a tech tree that allows for partial specs into a class that shares similar traits. For example, a tank could have 3 spec trees, 2 of which offer the ability for full spec into tank related abilities and passives totaling 75 points per spec tree, the 3rd allows for spec into fighter related passives/abilities with only 33 available spec points.
However, to preserve class identity it would be important to create a classification of abilities that falls into class specific only.
Once level 25 is reached and a 2ndary class is chosen and an additional skill tree would become available with 50 total spec points. This skill tree would not have any abilities and would provide added status affects to the abilities held by the original class: DOT, duration, decreased ability mana/stamina cost etc, this would be related to the 2ndary class chosen. For example, Rogue could give crit chance and DOT augments etc, this would also need to change ability animations to preserve the feeling of the 2ndary class chosen.
From above we know that each class will have the ability to spec into 233 points of abilities/passives/augments etc. The player will only ever have 150 spec points. Each tree should have the ability to be fully utilized without increasing or ramping point usage. If you ramp the points needed to reach the end of any spec tree than a meta will form around those abilities as you will need to independently balance those final abilities to account for the points spent. This would result in classes being built around the way those abilities work instead of how the class works.
A theoretical example of class with their spec trees,
Tank – 2 spec trees for tank specific abilities/passives, 1 reduced one for fighter/summoner
Summoner – 2 spec trees for summoner abilities/passives, 1 reduced one for tank/mage/fighter
Rogue -- 2 spec trees for rogue abilities/passives, 1 reduced one for ranger/mage/fighter
Ranger -- 2 spec trees for ranger abilities/passives, 1 reduced one for mage/rogue
Mage -- 2 spec trees for mage abilities/passives, 1 reduced one for ranger/summoner
Fighter -- 2 spec trees for fighter abilities/passives, 1 reduced one for tank/summoner
Cleric -- 2 spec trees for cleric abilities/passives, 1 reduced one for bard/mage
Bard -- 2 spec trees for bard abilities/passives, 1 reduced one for cleric/ranger
If this partial spec tree was given regardless of 2ndary class, it would be essential to not allow for any abilities that provide uniqueness to the class to be shared. For example, if a rogue has a stealth ability nobody else has then the ranger cannot spec into that rogue ability within their own spec tree but perhaps, they could spec into the ability to hide their nameplate or perform a trap check etc. This design would also allow for more diverse raid/dungeon groups.
Damage types
• Magic
• Physical
• Holy/Radiant damage
• Elemental (Assumption = Air, Fire, Water, Lightning, Darkness, Light, Planar)
• True damage (non-mitigated damage)
Defensive Stats
• Magical
• Physical
Armor types
• Light – More mobile, better against magical damage mitigation
• Medium – general damage mitigation and possible 2ndary stats like critical chance
• Heavy - More hp and physical damage reduction
Mobility
• Auto attacks and most spells can be cast while moving however suffering a movement speed penalty.
Resources
• Stamina
• Mana
Ability types per class
Damage (1 Value)
• Direct damage
• Aimed damage
Tank (1 Value)
• Mitigation
• Threat Generation (Threat accumulation abilities for aggro management)
Support (1 Value)
• Heal
• Resistance
• Damage
Crowd Control (CC)
• Slow (1.25)
• Disarm (1.33)
• Snare (1.4)
• Root (1.5)
• Silence (1.33)
• Sleep (1.75)
• Stun (1.85)
Ability type subcategories
• AOE (Area of effect) -- (Square root radius r (r^1/2))
• DOT (Damage over time) -- (Square root time t (t^1/2)
• Physical (1)
• Magical (1)
• Holy/Radiant (1)
• Elemental (1)
• True damage (1.5)
• Dash (((Length^1/2)/(Duration^2))) * (1/Cast time) -- (L^1/2 / D^2) / (1/C) The higher cast time and duration lowers the value of the dash, higher the length, higher the value)
• Blink (Distance^1/2/cast time^2) – (D^1/2/C^2)
• Jump ((Length^1/2)/(Duration^2)) (L^1/2/D^2)
Now to balance out the classes.
Each ability will have its basic stat, like direct damage, which can be coupled with a CC ability and up to 2 ability subcategories.
Each modification to an ability carries a value. For example, a direct damage ability, augmented with a slow and AOE with radius 5 IGU (in game units) would have a value of
(1 + 1.25 + (5^1/2) = 4.486) regardless of damage.
Now we have a way to arbitrarily value every ability combo in the game before damage and mana/stamina drain are considered. Damage values are nearly meaningless by themselves so you would need to create an additional relationship between damage / cool downs and resource usage and the ability value I just talked about.
To assign appropriate mana/cool down and damage values we need to create an inverse relationship between the arbitrary value already given. Assign CDR (cool down reduction), mana use and damage a value between 0.1 and 9.99 and multiple it by the original arbitrary value. As the value nears 0 it is more resource intensive and as it nears 9.99 less resource intensive. For example, the ability above had a value of 4.486, the goal would be to achieve an arbitrary value of 1.
So, with said ability to achieve a value of 1 we could
• (Cool down (CD)) (4.486*.75)
• Mana (3.364*.35)
• Damage (1.177*.85 = 1.000)
Now to quantify what the values between 0.1 and 9.99 mean is more difficult but we will use percentages because number are kind of pointless here otherwise.
First you want to set your maximum and minimum values for your ability cool down, damage and mana use. Let’s just say cool down is between 5 and 60 seconds, mana is 5%-100%, and damage (% max hp on a player target without any resistances or armor at max level) 5%-50%. Now you have your minimum value 0.1 = 5 seconds, 5% mana and 5% max hp and your maximum value 9.99 = 60 seconds, 100% mana and 50% hp.
(Edited minimum CD from 2->5 seconds)
To be honest this last part I could not figure out, does this need to be exponential with 1 being perfectly in the middle and the rate of change (regarding CD/Mana/Dmg) increasing the further from 1 you go, or should it be linear? Probably exponential but I am truly unsure. I am unsure if it matters either way, I feel like it does but can’t put my finger on why.
Finally, when adding penetration to things you can add a subsect value to damage that is treated the same way damage was. Basically, if you have physical pen add in a number relational to the max pen and min pen available and multiple the damage modifier by that number. Make sure to include 0.1 through 9.99 so that you can have negative effects applied easily. (Like equipping the sword of blah blah increase physical damage taken by 20% etc)
Now you have a way to value every ability in the game and scale it to match every other ability in the game. The key here is that everything equals 1, even if the abilities are completely different, they all have a basic formula and when something is out of whack it should be out of whack across the board and thus changing the original formula value would result in the desired change.
I did not account for movement speed differences between classes or the trinity, but again if you assign a formula to this and create an arbitrary value you can scale each class to be different but in line with another by having the final value equal 1.
I understand that I may have made several fundamental errors when evaluating this however I hope I didn’t and that the base logic stands up. The base logic being that if you separate every ability and its defining factors, assign a value to that then add it up and multiply by its constraints (CD/Dmg/Mana) and make every ability equal the same number you should be able to have an easy way to balance the game forever, since anytime something is out of whack it merely means the arbitrary value assigned to that ability doesn’t correlate with its true value and thus needs to be changed. It will mean that balancing jump/etc for ranger will change jump/etc for all classes, however that should be the case with this balancing system.
Also, I was really bored today and really enjoyed making this post, sorry if it is nonsensical or illogical, I did try and no I don’t expect you to read it, but if you did, tell me what you really thought about it, I’d love some feedback or to change it to be more accurate if possible. Man I sure would love to intern at IS
https://ashesofcreation.wiki/Roles “ 1v1 matchups will have a rock-paper-scissors dynamic, where one class will be superior to another.
Assumption - placeholder number for Total available spec points per class = 150
Every level gained will give the player 2 points for specs until level 25, from level 25-50 each level will give the player 4 spec points for a total of 150.
• Tank will lose to Support
• Support will lose to DPS
• DPS will lose to tank
Each of the 8 classes will need to be categorized into its trinity role within its class. Each class would have spec trees aligned to their gameplay type. However, each class would have a tech tree that allows for partial specs into a class that shares similar traits. For example, a tank could have 3 spec trees, 2 of which offer the ability for full spec into tank related abilities and passives totaling 75 points per spec tree, the 3rd allows for spec into fighter related passives/abilities with only 33 available spec points.
However, to preserve class identity it would be important to create a classification of abilities that falls into class specific only.
Once level 25 is reached and a 2ndary class is chosen and an additional skill tree would become available with 50 total spec points. This skill tree would not have any abilities and would provide added status affects to the abilities held by the original class: DOT, duration, decreased ability mana/stamina cost etc, this would be related to the 2ndary class chosen. For example, Rogue could give crit chance and DOT augments etc, this would also need to change ability animations to preserve the feeling of the 2ndary class chosen.
From above we know that each class will have the ability to spec into 233 points of abilities/passives/augments etc. The player will only ever have 150 spec points. Each tree should have the ability to be fully utilized without increasing or ramping point usage. If you ramp the points needed to reach the end of any spec tree than a meta will form around those abilities as you will need to independently balance those final abilities to account for the points spent. This would result in classes being built around the way those abilities work instead of how the class works.
A theoretical example of class with their spec trees,
Tank – 2 spec trees for tank specific abilities/passives, 1 reduced one for fighter/summoner
Summoner – 2 spec trees for summoner abilities/passives, 1 reduced one for tank/mage/fighter
Rogue -- 2 spec trees for rogue abilities/passives, 1 reduced one for ranger/mage/fighter
Ranger -- 2 spec trees for ranger abilities/passives, 1 reduced one for mage/rogue
Mage -- 2 spec trees for mage abilities/passives, 1 reduced one for ranger/summoner
Fighter -- 2 spec trees for fighter abilities/passives, 1 reduced one for tank/summoner
Cleric -- 2 spec trees for cleric abilities/passives, 1 reduced one for bard/mage
Bard -- 2 spec trees for bard abilities/passives, 1 reduced one for cleric/ranger
If this partial spec tree was given regardless of 2ndary class, it would be essential to not allow for any abilities that provide uniqueness to the class to be shared. For example, if a rogue has a stealth ability nobody else has then the ranger cannot spec into that rogue ability within their own spec tree but perhaps, they could spec into the ability to hide their nameplate or perform a trap check etc. This design would also allow for more diverse raid/dungeon groups.
Damage types
• Magic
• Physical
• Holy/Radiant damage
• Elemental (Assumption = Air, Fire, Water, Lightning, Darkness, Light, Planar)
• True damage (non-mitigated damage)
Defensive Stats
• Magical
• Physical
Armor types
• Light – More mobile, better against magical damage mitigation
• Medium – general damage mitigation and possible 2ndary stats like critical chance
• Heavy - More hp and physical damage reduction
Mobility
• Auto attacks and most spells can be cast while moving however suffering a movement speed penalty.
Resources
• Stamina
• Mana
Ability types per class
Damage (1 Value)
• Direct damage
• Aimed damage
Tank (1 Value)
• Mitigation
• Threat Generation (Threat accumulation abilities for aggro management)
Support (1 Value)
• Heal
• Resistance
• Damage
Crowd Control (CC)
• Slow (1.25)
• Disarm (1.33)
• Snare (1.4)
• Root (1.5)
• Silence (1.33)
• Sleep (1.75)
• Stun (1.85)
Ability type subcategories
• AOE (Area of effect) -- (Square root radius r (r^1/2))
• DOT (Damage over time) -- (Square root time t (t^1/2)
• Physical (1)
• Magical (1)
• Holy/Radiant (1)
• Elemental (1)
• True damage (1.5)
• Dash (((Length^1/2)/(Duration^2))) * (1/Cast time) -- (L^1/2 / D^2) / (1/C) The higher cast time and duration lowers the value of the dash, higher the length, higher the value)
• Blink (Distance^1/2/cast time^2) – (D^1/2/C^2)
• Jump ((Length^1/2)/(Duration^2)) (L^1/2/D^2)
Now to balance out the classes.
Each ability will have its basic stat, like direct damage, which can be coupled with a CC ability and up to 2 ability subcategories.
Each modification to an ability carries a value. For example, a direct damage ability, augmented with a slow and AOE with radius 5 IGU (in game units) would have a value of
(1 + 1.25 + (5^1/2) = 4.486) regardless of damage.
Now we have a way to arbitrarily value every ability combo in the game before damage and mana/stamina drain are considered. Damage values are nearly meaningless by themselves so you would need to create an additional relationship between damage / cool downs and resource usage and the ability value I just talked about.
To assign appropriate mana/cool down and damage values we need to create an inverse relationship between the arbitrary value already given. Assign CDR (cool down reduction), mana use and damage a value between 0.1 and 9.99 and multiple it by the original arbitrary value. As the value nears 0 it is more resource intensive and as it nears 9.99 less resource intensive. For example, the ability above had a value of 4.486, the goal would be to achieve an arbitrary value of 1.
So, with said ability to achieve a value of 1 we could
• (Cool down (CD)) (4.486*.75)
• Mana (3.364*.35)
• Damage (1.177*.85 = 1.000)
Now to quantify what the values between 0.1 and 9.99 mean is more difficult but we will use percentages because number are kind of pointless here otherwise.
First you want to set your maximum and minimum values for your ability cool down, damage and mana use. Let’s just say cool down is between 5 and 60 seconds, mana is 5%-100%, and damage (% max hp on a player target without any resistances or armor at max level) 5%-50%. Now you have your minimum value 0.1 = 5 seconds, 5% mana and 5% max hp and your maximum value 9.99 = 60 seconds, 100% mana and 50% hp.
(Edited minimum CD from 2->5 seconds)
To be honest this last part I could not figure out, does this need to be exponential with 1 being perfectly in the middle and the rate of change (regarding CD/Mana/Dmg) increasing the further from 1 you go, or should it be linear? Probably exponential but I am truly unsure. I am unsure if it matters either way, I feel like it does but can’t put my finger on why.
Finally, when adding penetration to things you can add a subsect value to damage that is treated the same way damage was. Basically, if you have physical pen add in a number relational to the max pen and min pen available and multiple the damage modifier by that number. Make sure to include 0.1 through 9.99 so that you can have negative effects applied easily. (Like equipping the sword of blah blah increase physical damage taken by 20% etc)
Now you have a way to value every ability in the game and scale it to match every other ability in the game. The key here is that everything equals 1, even if the abilities are completely different, they all have a basic formula and when something is out of whack it should be out of whack across the board and thus changing the original formula value would result in the desired change.
I did not account for movement speed differences between classes or the trinity, but again if you assign a formula to this and create an arbitrary value you can scale each class to be different but in line with another by having the final value equal 1.
I understand that I may have made several fundamental errors when evaluating this however I hope I didn’t and that the base logic stands up. The base logic being that if you separate every ability and its defining factors, assign a value to that then add it up and multiply by its constraints (CD/Dmg/Mana) and make every ability equal the same number you should be able to have an easy way to balance the game forever, since anytime something is out of whack it merely means the arbitrary value assigned to that ability doesn’t correlate with its true value and thus needs to be changed. It will mean that balancing jump/etc for ranger will change jump/etc for all classes, however that should be the case with this balancing system.
Also, I was really bored today and really enjoyed making this post, sorry if it is nonsensical or illogical, I did try and no I don’t expect you to read it, but if you did, tell me what you really thought about it, I’d love some feedback or to change it to be more accurate if possible. Man I sure would love to intern at IS
2
Comments
idk if you value my experience much, but at the very least, you can bounce stuff off me.
If you want to 'consider balance for Ashes' though, as we understand it to be, you would have to drop the concept of assigning a balance-able value to CDR because of the way that works in group play, and start from the 'outcomes desired' side for that one.
Similarly, remember that even if you turned their Augments system into a pure skill-point thing (which at the moment, it actually is explicitly NOT), you would have to account for Social Orgs and similar Augment rewards, which doesn't work as cleanly. Right now, being able to think of it more like 'slots' is the 'safer' method for the 'Augment'/'subclass' part of the spec.
Finally, there would need to be a diminishing return or rising cost on the concept of assigning the same debuff type to every ability in your kit, even if it didn't stack for this to work as an MMO balance scalar. This isn't usually obvious because MMOs are actually the ONLY competitive game type that usually seem to require this due to their high number of available abilities.
Overall, (in all arrogance) you have a solid understanding of the starting points you'd need to get good balance on something, but you're lacking some fundamentals that take a while to learn without getting to actually test your concepts on a live 'audience'.
I am not sure why social orgs and augments can't have the same arbitrary value application given them? A player can only be in 1 social organization at a time so you make the correlating arbitrary value for each social organization equal each other.
Before I try to explain this with way too many examples OR way too much prefacing...
Have you played any MOBAs? Particularly 'things other than League'?
As for the social orgs thing, it's actually an extension of something you already noticed.
To be honest this last part I could not figure out, does this need to be exponential with 1 being perfectly in the middle and the rate of change (regarding CD/Mana/Dmg) increasing the further from 1 you go, or should it be linear? Probably exponential but I am truly unsure. I am unsure if it matters either way, I feel like it does but can’t put my finger on why.
The WAY one applies the Augment will matter if there isn't a hard-set 'order' for which way the multiplication is done, but as soon as you apply that hard-set order, you'll end up limiting the Org augments themselves. Whereas you could get a stronger thematic concept out of those generally. They would also 'break' any limitations you have on which ability types are allowed to apply which things, or upset players when their RP target and their Abilities don't gel (a common problem across many games that work like this) leading to an annoying type of meta.
But I suggest you attempt to crunch such numbers yourself, don't take my word for it here. I'm just letting you know that my experience from balancing games LIKE this leads me to give this advice type.
Your method solves that in a 1v1 or 2v2 game, certainly (except card games because in card games you're moreso 'controlling a collection')
CDR Balance
The reason CDR has to be balanced differently in a group game is that the 'secondary cooldowns' do not usually matter in a GvG situation. Here is why.
Any ability with a short cooldown is almost always just healing, knockback or DPS. Shortening your cooldown for this ability is equivalent to simply increasing your damage output, simple enough.
The Majority of abilities with LONG cooldowns are situational and are set up for their success (or buffs with an intended ratio of downtime/uptime, in which case, usually, see above). They may or may not be 'very powerful' in other situations. They're usually powerful enough for the situation they intend. Otherwise they're definitionally just bad.
When playing 1v1, the 'purpose' of a cooldown on a situational 'powerful' ability is almost always 'to limit you from using that ability every time the situation comes up'. Otherwise it might as well have no cooldown at all. This is where the problem arises in GvG situations and why teamfights in MOBAs tend to devolve into randomness and ability spam at low levels of play-skill.
If player A in any 'lane' did not need to use their longer cooldown ability in a 1v1 before the teamfight, then the cooldown does not matter. It will almost always be available. The fight, even if it lasts long, would have to contain 'a second instance of them wanting to use this ability, but being unable to', for the CDR to be helpful. Either it 'would be helpful because this often happens' (the situation is easy to set up) or it 'would be unnecessary because this doesn't happen more times than the player has the ability available'. That's just 'facts' though, it isn't the answer to why you have to do it separately, so bear with me a bit longer.
The specific thing I was 'concerned about' though, is 'assigning a balance-able value to CDR'. It isn't that I think there should be no CDR, or even that I think CDR is either 'always useful' or 'always useless'. It's that CDR is 'tuned' differently than the rest of the benefits you can give your character, in group play. If you outnumber your opponent, you don't need it usually, they die. If you are outnumbered, or suddenly put in a bad situation, you might benefit a LOT from it, or you might benefit precisely NOTHING from it.
Min-maxers (such as myself, though it's not my PLAYSTYLE, just analytic style) will look at how to tune their CDR to the ratio/frequency of the encounter, in a MOBA (basically change it according to enemy team comp in a match). In an MMO you would not have this option. CDR is very binary, in that sense for longer abilities. Either it is 'helping a LOT' or 'not helping at all' and the difference between those things can be any arbitrary amount of additional CDR, depending on the situation.
Therefore you can't reduce it to a SCALE, because if I pay 2 points to get X amount of CDR and it is not enough on average (and this curve is not a normal distribution), then I am paying 2 points for 'nothing'. I then pay 1 more point for Y CDR and suddenly due to team comp or whatever I am getting FULL benefit. Similarly if I gain enough CDR to do a 'killing amount' more damage through any means, I am always getting full benefit from THAT even if I don't actually succeed at the kill (you could tune it to be equivalent to damage output from similar choices, but then you'd still have to figure out how to factor for the longer Cooldown Stuff).
You also can't reduce it to a MULTIPLIER, because if I have to pay more points to reach a threshold and then 'giving the ability a different effect with a higher multiplier' puts it out of my build reach, then every ability's cost has to be balanced around that precise potential for min-maxing and a meta will form, (people will just pick the least 'counterable' option, which is what they always do when you increase the cost on something good).
This applies regardless of how exactly the CDR works (though percentage based CDR is generally 'worse' and more swingy because the efficacy distribution is basically a Poisson distribution, on the 'did the situation happen' side which is then sorta 'multiplied by' a binomial I THINK don't quote me on this one it's an aside and I haven't had to do this one in a WHILE).
Obv I'm making a lot of assumptions here and drawing purely from my own experiences, but hopefully I am making all the same ones you are based on what I read from your post. Happy to address something like a more specific skill+augment suggestion/concept if you have one.
All who see flaws or mistakes, please point them out.
2nd if the TTK is 30-60 seconds long doesn't that tell us the appropriate value regarding a powerful abilities cool down? Like you said if I only need to use an ability with a 60 second cd once then the cd was pointless which is why a 1v1 doesn't matter and IS wants to balance around 8v8, so in an 8v8 the fight may last 5 minutes where the usage of that ability may be significant enough that having the ability on a 30 second vs 45 second cd matters?
I do also want you specific input on my statement regarding the general application of what I was saying in regards to the balance.
"The base logic being that if you separate every ability and its defining factors, assign a value to that then add it up and multiply by its constraints (CD/Dmg/Mana) and make every ability equal the same number you should be able to have an easy way to balance the game forever"
Can this be true? Since the values behind the arbitrary value don't really matter that much as long as you understand the correlation you want to have between abilities and the result you want to have at the end.
What I mean is that as long as you understand that you want an ability to have more "value" based on the idea that it does more damage or has a longer range or shorter cast time etc, as long as that correlation is decided upon than balancing is relatively simple?
@Azherae
I believe that IS has already done the balancing on Cooldowns, and I would support a version of the game with no CDR at all, due to their other designs.
Your base logic is entirely sound.
"The base logic being that if you separate every ability and its defining factors, assign a value to that then add it up and multiply by its constraints (CD/Dmg/Mana) and make every ability equal the same number you should be able to have an easy way to balance the game forever"
The difficulties come in three points in most genres, and MMOs add a 4th one MOST of the time.
1. Players with a specific intention will min-max and then not go back because the psychological profile of most min-maxers will cause them to stay the same 'out of fear of losing when they go back on their build'.
2. Players experience their own concepts of balance very differently whether they are winning or losing (which basically means that a perfectly balanced game does not FEEL perfectly balanced to a LARGE subset of its players)
3. Without a lot of experience with the 'edge cases' of why certain numbers don't work, you will either make a lot of mistakes (resulting in imbalances) or end up under or overcomplicating the numbers to the point for your specific game, to the point where you're just rebalancing and upsetting (most of) your players all the time.
4. MMOs in particular tend to use RNG, and RNG is streaky even with true randomness involved. It is also extremely difficult to find bugs in your RNG quickly, and this balloons into a huge problem in complex, memory-optimized games (new languages and frameworks do not help here, because optimization takes place at the machine code/assembly level, and BOY is that fun and good at causing fail conditions). On top of that, RNG affects all three of the first things I noted, and in the case of #3, it then throws a wrench in all your balancing of numbers forcing you to do actual statistics just to determine if the numbers are right for actual balance.
But the base methodology you mention is how this 'should be done', and as a result, you can actually learn to 'see' the underlying values in new competitive games by trying to figure out the scales they are using, for practice. You can even 'learn to predict which things will need to be fixed and patched' or 'which things will be OP' before the game even goes live, and watch to see if you were right or not.
It's pretty good practice. I'm currently working on Predecessor (MOBA) but it's really hard to do the second part because their balance is so good. However that makes it a good starting point for people who want to learn to do it, because you wouldn't end up spending as much time 'wondering why this number looks wrong' and then having to wait 4 months for the Devs to gather enough statistical data to prove you right. You can just 'learn it with fewer wrong numbers'.
I, in my arrogance 'already have the numbers for Ashes' based on what they SEEM to want, and yours are very close to right, but note that I don't actually know what Steven/Combat Lead REALLY wants for a specific thing because I haven't seen their PvE, and to me, their base stats are flawed (basically it's a sort of copy-paste from L2, applied to a game that doesn't use the same structure, so I'm hoping it changes).
Excellent points on everything.
In particular, I want to highlight what you said regarding RNG.
This is one of the reasons why I stated my dislike for % block and % evasion in another thread.
% critical hit chance is acceptable because it is the most relevant when used on builds that want to deal damage through multiple hits. Builds that rely on single hits to deal large amounts of their damage are going to want sources of guaranteed critical hits that they can manually apply to specific skills or ones that are automatically applied once (a) certain condition(s) are met.
My thought was that I would like the idea of allowing each tank/support/dps role to have at least one sub class that can counter a sub class from any other role. For example, a tank has one sub class that can counter a dps, and another sub class that can counter a support- and same thing for support and dps sub-classes. That way things aren't so rigid in terms of which overall roles are good/bad in a given matchup (like if tank always beats dps, support always beats tank, dps always beats support), because your role would have a certian spec that could beat the sub-class you are matched up against- which gives more freedom to stick to a desired role without feeling helpless or forced to switch.
As far as actual balancing- its not neccessary that the values for each classes' end-output are exactly the same, due to the rock/paper/scissors design approach- but it is important that when comparing 2 classes in a given matchup, with one being better/worse in that matchup, that skill-difference is weighted much higher than the inherent difference in end-output.
To expand on this (for Voxtrium moreso), I'll repeat a specific thing I always say. In this case I will frame it as the REASON why the R/P/S Voxtrium outlines is correct.
Tanks lose to Support because Support tend to have DoT and Attrition style damage that 'ignores most things about the Tank other than their HP pool', while buffing themselves against the Tank's relatively limited damage /mobility.
'Support' lose to 'DPS' because they tend to be less HP and defenses and are quickly overwhelmed and can't escape.
DPS lose to 'Tanks' because they apply all their energy and don't manage the kill due to whatever mitigation is built into the Tank that allows them to be considered Tanks at all.
It is my constantly-repeated perception that this is really a 'Mitigation/Attrition/Burst' form of 'R/P/S' and even though a Tank will almost always be Mitigation first, they could choose to either double down on that, add Attrition, or add Burst Damage, to help them against their 'natural predator'.
In GvG, the distributions of these, combined with the understanding and cooperation styles of the players, would then tend to balance it out if you could adapt fast enough. So, while 'Support >> Tank >> DPS' is generally correct, it's correct because of something underlying that can be used to make a more interesting game overall without having to be concerned about 'overall balance problems'.
Basically, at some point, if a Mitigation Archetype decided to throw ALL their effort into becoming a direct Attrition character, the balance would be maintained because they would 'change form', they would give up so much Mitigation that they would go from Mitigation-Primary With Attrition-Secondary to Attrition-Primary With Mitigation-Secondary, but this wouldn't necessarily indicate a failure of balance. There would definitely be a person who might complain 'I use stacked DoT I'm supposed to be able to kill Tanks before they kill me' when they find one SPECIFIC Tank that kills them first, but this wouldn't imply that the R/P/S balancing was broken.
We normally end up arguing about 'whether or not we will be able to conceal our builds' though, so Intrepid might unfortunately be forced to 'make all Tanks only be able to function as Mitigation Primary' just so that you can never be 'fooled'.
Still, if they get past about 30-40% 'freedom' in build options (almost guaranteed given their gear system, but who knows), it should always be enough to at least make the fight 'even'.
Actually I disagree here. The reason being that while yes its a trinity system, but if each ability equals an abstract value that is the same across the board then the ability should be balanced against other abilities, so you don't end up with crap like League of Legends where bruisers have run supreme for seasons because of a support nerf seasons ago. The whole point of this arbitrary value assigned to each ability is the idea that as long as the developer took into account every aspect of a spell then they could balance all spells quite easily.
To create a proper trinity system within the bounds I specified would be as easy as creating a theme for each trinity role then properly applying ability types. CC naturally counters fragile targets like DPS, DOT-True damage, pen naturally counters tanks and burst naturally counters sustain. Like what @Azherae said, pretty much R/P/S. This doesn't actually change anything though, the application of an arbitrary value behind the scenes makes sense from a balancing standpoint.
I think you are possibly misunderstanding what I am trying to communicate, because I already held the same opinion that Azherae explained, which aligns just fine with the points I made. My point isn't neccessarily relevant to the mathematical method of creating balance that you suggested. It was moreso in regards to the philosophy behind what "balance" actually is- in regards to what end values you are striving to achieve with each class and their maximum effectiveness.
Don't get me wrong, I fully understand the point you are making, the purpose behind it, and how it works- my point is that the classes can be inherently unbalanced, while still having an overarching system that perfectly balances those unbalanced classes, as I will explain.
If the gameplay just boils down to sticking to one endgame class build and having those endgame builds be perfectly balanced, then the "trinity system" becomes more arbitrary/gimmicky, which kind of lessens the impact of the "distinct roles/class fantasies". If it were a fighting game then yeah I would agree, but its an rpg, so its important to have "good balance while actually having real tangible differences between classes". Which means an actually impactful rock/paper/scissors system. Its okay if things are not "perfectly balanced" in 1v1, as long as you allow players to switch their specs somewhat freely to empower them to use counter-play to prepare/adapt to different situations/opponents. And for the times that they are wrong in their spec choice strategy, they should be at a disadvantage within that rock/paper/scissors balancing, but not to the extent that skill difference cannot make up for that disadvantage. This can actually become "perfectly balanced" in a sense, through that easily accessible counterplay- but enhances the layers of strategy involved without diminishing the importance of actual combat skill, which is a good thing for those who value the importance of skill/strategy being prioritized.
I think that intrepid's claim that they are "not balancing for 1v1" but are "balancing for groups" is a bit misleading to what they are actually doing. If base classes have counterplay then so will team compositions. So really they are just saying that all of the classes aren't balanced 1:1 at an individual or group level. This, however does not mean things are unbalanced, it moreso just means that the balance is through counterplay, allowing players to actively choose strategies for different group compositions. So all I am saying is that this just needs to translate into the 1v1 setting by having that same strategy and freedom of counterplay at a 1v1 level with spec choices, to have balance in both 1v1 and group play, along with distinct classes.
It's also worth noting that skills that may be strong in group play may not be as effective in 1v1s, and vice versa.
In my oppinion to balance properly the R/P/S, the result would ensure that all classes die if they play in an ideal way which maximizes their survival time.
But balancing can still be difficult because there might be situational cases like amplified damage (backstab) or player preference and habits which may not change as developers expect.
Also different damage type may be countered better by different armor which takes time to be replaced. A balancing change creates an initial impact onto players which changes over time, some being more affected than others.
what about support vs support, tank vs tank and dps vs dps?
3 classes can beat 1 but lose to 4, 1 can beat 4 but lose to 3 and 4 can beat 3 but lose to 1? o-o
the game will be balanced around 8 men parties, not 1v1 btw. naturally, some classes / builds will have advantages vs others.
also 150 points is too much...each class will have around 30 skills...so u could get all thee skills at level 3 and still have 60 points to augment some skills. its better to separate the resources needed for these 2 systems.
and whats wrong with metas? and u cant really avoid that unless everything is the same and equal, then everything would be boring. classes are classes because of their skills. that's what makes them a class. nothing wrong with building around certain skills. you arent gonna pick a fire skill to fight fire resistant monsters. you would pick a water skill. then, you can reset and change it when you decide farming there isn't worth it anymore.
Work hard, play hard
It's quite impressive and interesting reading through posts with such meticulous thought put into them!
you spend 1 skill point to acquire one ability at level one, then you can level it to 2 or 3 by spending 1 or 2 points. i guess they could always make it so that skills cost more, or more the more you get..or just reduce the amount of points you get, or use a different resource from augment.
i still think a different resource is better so that both things can be tweaked separately
I actually am designing my own game, not that balance is ever going to be an issue, but this thought crossed my head and I ran with it, but by the time I actually sat down to put it into words it only took an hour or two. Still most of the time I feel silly writing about something like this, I always feel like I don't understand something, and when you think about someone sitting down and spending an hour or two writing a forum post that has no basis or fills any need, then it feels a lot sillier. Thank you very the kind response though.
Relatable. Nothing wrong with letting others know your thoughts and ideas, especially in a place where others have similar interests.
Examine PvE and PvP stats, and impact of items/gear.
If archetype X or class Y is unfairly distributed at the top of the PvP ranking (by unfairly, I mean if tanks represent 10% of all players but tanks are 50% of the top 100, and their winrate is at 65%, then they should be nerfed). Or a certain gear/item is OP as per pvp winrate.It also has to be examined closely in terms of the different comps being run; is it a class thats OP or a class combination of several players?
PvE is a little harder, but you could calculate in "time it takes to reach max level" or overall gear level, gold levels, pick %, something else. Quite complex as well and I can already see a ton of people fuming at this idea lol.
All classes should be buffed/nerfed so that they have roughly average 50% winrate in PvP. High skill cap classes will usualy translate to a higher difference between best/worst average win rates, i.e. larger standard deviation.
Its hard to manage and a game will never be perfectly balanced but if OP classes are just left untouch, it could ruin the game by making the Meta boring and uncompetitive.
Please no- winrate/usage balancing is so bad. There are so many factors that go into that other than the actual balance of just the classes in a vacuum. All that does is throw a wrench in your balancing by adding player skill/preferences into which classes are over or under performing.
No matter how hard you try to keep players in tiers of skill or whatever you try to do to have an "all else equal" comparison, you can never really get that so its just sub-optimal, rather than just balancng and tuning from the ground up, accounting for the numbers/skill-checks/combat capabilities of each class at a fundamental level, and letting the chips fall where they may in performance. Results should be skewed, because player differences are skewed- so if thats not the case then your classes are unbalanced to be able to reach complete equity in winrate performance.
A huge problem in game development, or really anything, is when people aren't able to recognize the difference between causation/correlation.
The reason I lean towards statistlcal help for balancing is that its impossible to balance different classes against eachother right out of the gate. You need testing for this. This is where the real players and real stats help you with the testing.
Of course a lot of other factors come into consideration, which is why being on the balancing team is a hard task.
I understand that take, but just because the goal isn't for 1v1 balance, doesn't mean the method of balance can't be balancing each class individually from the ground up in a vacuum. Even if they want advantageous matchups, this type of balancing would actually make it easier to achieve that goal, because they can then isolate and tune those balance factors in a way that lets certain classes perform differently against others.
Yeah- I absolutely agree that testing is definitely a necessary tool, but just as long as they don't take the results at face value, without taking the necesary steps to distinguish between the correlating and causal factors of those results. Definitely a tough job for sure, but very important to get right.