Please upgrade your browser for the best possible experience.

Chrome Firefox Internet Explorer
×

Formulaic Question

 

Kitru's Avatar


Kitru
12.24.2012 , 01:13 PM | #1
For those of you who aren't interested in math or the abstract concepts of game design and balance, you can probably just ignore this completely.

I working on a purely conceptual project dealing with building abilities and attack strings and bunch of other things only tangentially related to what's in game at the moment I'm curious if anyone else has already been able to answer a question that, otherwise, I would have to do a lot of work myself to arrive at. The question is remarkably simple but notoriously complex since you're dealing with a *lot* of specific abilities and I'm looking for a general guideline: what is the general balance formula for the additional power attributed to an ability based upon the presence and duration of a cooldown? For example, Healing Trance heals for more than Deliverance for a similar cast/activation time and cost because it has a CD. I'm looking for a basic formulaic guideline of how much more powerful an ability should be based upon the presence of a CD and the duration of said CD.

I have a general feeling (based upon my intuition and experience already dealing with similar formulas) that it's on something of a diminishing returns curve, but I'm just curious if anyone has already done anything of the kind?
Walls of Text? I *love* Walls of Text!
My New Class Idea
Shadow Class Rep - Suggest/Review Questions Here
Quote: Originally Posted by Fende View Post
Listen to Kitru. Kitru knows all.

RendValor's Avatar


RendValor
12.24.2012 , 07:33 PM | #2
That sounds like it would be relatively easy to at least get an aproximate value of.

Take all the instacast damaging abilities that don't have a cooldown and compare their damage-to-cost ratio, then take all instacast damaging abilities that have the same cooldown (ie all abilities that have a 6 sec cooldown, all abilities that have a 12 sec cooldown, all abilities with a 15 sec cooldown, etc. anazlyed separately) and compare their damage-to-cost-to-cooldown ratio and make a curve with that data
The Trayus Legacy - Prophecy of the Five

Kitru's Avatar


Kitru
12.25.2012 , 12:34 AM | #3
Quote: Originally Posted by RendValor View Post
That sounds like it would be relatively easy to at least get an aproximate value of.
Eh, it's bit harder than that because there are a lot of weird variables involved, like whether the attack is AoE or ST, has any stuns involved, requires special conditions, what its range is, whether it's a DoT, variable CDs (HiB has a 15 sec CD, but it can be reduced to ~6-7.5 for some specs based on procs), etc. Common outliers like Project (with exceptionally high cost and low default damage) coupled with the general smallness of the actual number of attacks (compared to similar systems I've dealt with like what City of Heroes/Villains operated with) in game means that simple averaging is going to have some really skewed results or have a very small data set.

I've kind of figured that I'm probably going to have to just compile most of the information myself. It was an outside hope that someone else would've already done it or something similar by now (thank the Force that I don't have to deal with family around this time of year, lol).
Walls of Text? I *love* Walls of Text!
My New Class Idea
Shadow Class Rep - Suggest/Review Questions Here
Quote: Originally Posted by Fende View Post
Listen to Kitru. Kitru knows all.

RendValor's Avatar


RendValor
12.25.2012 , 08:21 AM | #4
Quote: Originally Posted by Kitru View Post
(compared to similar systems I've dealt with like what City of Heroes/Villains operated with)
Off topic: thanks a lot you jerk, you just made me cry
The Trayus Legacy - Prophecy of the Five

Cleet_Xia's Avatar


Cleet_Xia
12.28.2012 , 11:23 AM | #5
I think coming at from the direction of setting the CD as the first step in the ability design process, is to come at it from a more difficult perspective, that won't necessarily produce a "finished" product any faster or with less effort.

The best bet is....
1) defining what abilties need to be there - the utilitiy tools (interrupts, stuns, cc, knockback, anything not a dps or heal)
2) creating passive abilities that provide a concept to the class
3) creating active abilties that fit the concept of the class design <--- mostly just about animation design & range
4) set the costs of the active abilities based on their base effect relative to baseline abilities in the game.
5) adjust the cost/ CD/ channel time/ effect ~ to achieve something that may be balanced. (I ordered them according their "resolution" of adjustment, cost/ rough<--> effect/ fine)
6) play test
7) cross fingers, launch, and be ready to make adjustments based on how the class effects player experience

* always ask yourself, does this break the game?

I think you may be looking for a corelation there that isn't actually used in developement. Not that a formula for calculating these values couldn't be created. I just think you'd find that such a formula wouldn't actually describe the "balance" between the classes. The reason why I say this, is that while casting time is mostly just a variable that is being used to set the "feel" of play for a particular ability, by influencing player mobility. CD I think has been set mostly for balance in an after the fact fashion. I'm saying HT doesn't heal for more because it has a CD- it has a CD because it heals for more. You could set the effect OR the CD first, and either (or both) could be altered to adjust balance. It's not really the length of the CD that matters - but how many of those casts can you cram into the interval of the longest CD in the game. Classes are not balanced on a cast by cast basis, you have to involve a longer time interval than that. Disturbance illustrates what I'm talking about, because it's a cast that can be shorter than a GCD, just to make it immobile, but it has no CD.

There are a lot of 60 second CDs. So HT isn't an ability that can be cast every 6 seconds ~ it's an ability that can be cast 10 times in 60 seconds. Vs. Deliverance that can be cast 24 times in the same interval. The 2 piece PvE set bonus makes HT a 13.333/sec ability ~ a 30% improvement. HT can deliver healing before being interrupted, at the cost of resource efficiency if it is interrupted. Deliverance on the other hand, doesn't use resource if it is interrupted. The resource efficiency of HT is dependent on the cast being completed in it's entirety. HT also doesn't benefit from force potency in the same way. The relative importance of those unquantified differences depends entirely on play style ~ it's purely subjective. IMHO, the average efficiency of (HT & interrupted HT casts) vs (Deliverance & interrupted Deliverance casts) is nearly the same. Healing without using HT at all, is more than doable, it's just very "bursty", and with Force Potency, Deliverance may actually be more resource efficient. I find myself doing this all the time, simply to conserve casting time when my force bar is full and I'm taking an agressive position on the field ~ and switching to HT and a more safer position on the field when I need to conserve resource. The choice being made based on how likely I am to be interrupted based on where I've chosen to position my toon. That's just how I play. But other players are going to default to always casting HT on CD, may not use force potency for healing, and probably don't even concern themselves with the small difference in casting time. I wouldn't say that they are playing the abilities wrong, or that I'm playing them right, I'd just say that we're both playing how we like to play.

As far as balance is concerned, it's pretty easy to compare the abilities of the ranged classes to each other and see that they were developed from the same starting place. Consular/ Trooper/ Inquisitor/ Bounty Hunter all started from the same template, and that was derived from the same template as all the classes. That template is just the design limitations of the engine. (going to drop the mirror classes now, as the animations are the biggest difference) High Impact Bolt/ Project are essentially the same ability. The same could be said of TK Throw and Full Auto. There are a whole pasel of abilities that are functionally identical between all of the classes, that were added to provide minimal utility functions. The differences in these exist just specifically to differentiate the classes. Trooper was originally denied a ranged interrupt & given a AoE stealth detection ability, for example. Just as sort of a way to create a "selling point" for the class. Later on, players demonstrated that the lack of a ranged interrupt was major nerf to a ranged class, and the stealth detect was pretty ~ meh ~ when any AoE serves the same purpose, stealth isn't often used by MOBS in PvE, and troopers have a wide array of AoE abilities.

Some abilties were added simply as fluff to fill in GCDs between cds. Those are the abilities that are the less efficient abilities in terms of cost/ GCDs/ and the amount of damage done, such as Mind Crush & Charged Bolts. Being a DoT, or having the damage applied at the front of the cast, just makes two abilities feel different ~ eventhough they have a similar overall effect. CDs of abilities, and being set up as DoTs serve only to limit the frequency of use. Weaken Mind is a great example, the CD is lower than it's duration, so it's lack of a CD is actually meaningless. In order for it to be used efficiently, it functionally has a 15 second "CD". The closest Trooper analog is probably Sticky Grenade, which also has a 15 second CD. Both are also instant, and thus mobile casts, though Sticky Grenade involves a short stun that works in PvE to mitigate incoming damage. While Weaken Mind does a higher amount of raw damage, shortening the period of time over which the mob inflicts damage, thus mitigating damage. The numbers look quite different, but the two abilities have almost the same effect on the solo game experience, in practice. Having a variety of these abilities gives players the option of playing to their own whim, and making choices in how they flesh out their roations betwen their bread and butter abilities that are on CD or cast on proc. Force Armor is another instant with a messed up CD- because the 20 second lockout debuff it applies ~ it's useful 6 times over the 120 second CD of Reactive Shield, but given the differences in passive damage mitigation between the classes, they're almost the same ability if used on CD during a 120 second battle.

As for AoEs, Commando was given a pile of these with long CDs vs the single AoE of the base Sage build. The devs opted to let dps players decide if they were interested in using more or less AoE, through spec choices. While AoE feels like it works different than other abilities. Its is just a sum amount of dps spread across multiple targets. Those multiple targets could be killed with single target casts in a slightly longer time frame, but that is why AoE is typically less energy efficient dps - because it's more casting time efficient, and it's also mitigating damage. AoE suffers from the limitation that it's output is situationally dependent. If used on a single target, it becomes horribly inefficient. While some AoE has a maximum number of targets defined, simply to prevent the ability from becoming too efficient to balance. Any AoE that is very easy to use efficiently from a resource standpoint, is coupled with a long CD. And short CD AoE comes at the cost of resource or casting time efficiency.

Just staying in the vein of Commando vs Sage as ranged dps classes, you can see how the damage of Commando single target abilities was adjusted downward to compensate for the large amount of "easy" AoE, and the passive effect of armor on damage mitigation. The middle dps trees of both ACs give players access to abilities that compensate for the shortcomings of the base AC & give the 2 classes similar capacities at endgame. Gunnery grants single target abilties while Telekinetics grants AoE. Both specs use on proc casts to boost dps. Both single target and AoE roations end up having a very similar feel. And both specs are considered flat out boring by many players, for exactly the same reason ~ the rotations are boring, and feel almost identical.

I think really all you need to consider if you want to "value" abilities for the sake of comparing them between classes, or designing new abilities, is whether or not the ability CAN BE balanced and whether or not players will enjoy the game play of the balanced ability. Base damage values (before stats are applied), CD, casting time, debuff/buff length, damage type, proc relationship, passive ability modifiers, can all be adjusted to "balance" the ability & give it a unique feel. And on the engine side of, at least this game, I really think all the abilities work exactly the same. They apply a buff/ debuff with a duration and effect, invoke certain graphics files/ animation sequences, and that's about it. Abilities that have all their damage applied at the front, simply have a very short duration defined, and don't invoke a debuff icon. Class abilities & tech tree abilities have no expiration timer for their buffs, but don't invoke buff icons or animations. Speeders inovke a buff and an animation, without a visible duration. Although I imagine that the duration for all passive abilities is actually 9999.9 seconds, and the timers are being reset during loading screens. I suppose it could be tested, by sitting somewhere for 2 hrs 47 min ~ without loading a new zone by logging out, using a door, elevator, or taxi (this would explain the loading screen we sometimes get after riding a taxi). And since there is a cap on the total number of abilities that a toon can learn ~ players who have hit the cap with pets and speeders may be able to learn 41 more if they reset their spec. (passive class abilities have already been shown to contribute to the total number of abilities per toon ~ different classes hit the speeder/pet cap at different numbers - considering that all toons have a maximum of 41 skill points. RotHC will make it obvious if tech tree abilities coount toward the total.)

If you ever gave any class any ability that is high output/ low cost/ with too short a CD - you've made a one or two button class. ~~ Mario~~ Ironically enough, Every class has abilities that either actively, or on proc, let the player feel that jolt in capacity every so often. Force Potency serves the same purpose for a TK sage as Tech Override does for a Gunnery Commando. Both generate bonus dps, one by creating 2 crits, the other by creating time for another cast. Clicky relics and adrenals, do the same thing.

Formula:
Total output per 60 seconds of casts/ (casting time + (cost - passive regen during the cast)) = rotation value

NOTES:
1) For this comparison to mean anything, you have to convert the resource of each class to a common denomination.
2) Any abilities that require a proc to be used, or consume a proc during their cast ~ cannot be directly "valued" as a stand alone ability. The relative value must be inferred by subtracting the values of the other abilities that enable them to be used, after calculating the value of the whole "string" and adjusting the derived value by the ratio of it's frequency.
3) The 1.5 GCD is used for the casting time of an instant.
4) The regen rate used, has to assume that the ability is being used while regen is at the baseline value. As soon as you throw in variable regen modifiers like the concentration proc, regen dubuffs, emergency regen abilities, or being at a low regen level for trooper/ scoundrel ~ these values just go all kinds of wonky. You end up with sepperate values for every ability used with different stacks of buff on self, debuff on target, etc. The base regen value used for the knight has to estimated based on how much focus is generated in an average 60 second period of mixed ability usage.
5) You almost have to assume that DoTs are refreshed at the optimal rate, and not cast on multiple targets on CD.
6) Mitigation effects from abilities have to be quantified somehow, even if it's just the dps that is eliminated by a KB against a melee target. The good news is that you could actually compare the value of a stun to a dps ability or a heal.

In short, either a lot of variables have to be assumed, or an astronomical number of calculations have to be performed. You could build a formula to accomodate all of the variables ~ but it'd be enormous, without accounting for the unquantified properties of the ability that are situationally dependent.....it's a real mess. But I think you would find that most of the "rock star" (abilities from spec choices), actually have pretty low rotation values ~ and that the classes are actually balanced mostly on the filler abilities..... at least in TOR.

If you want to get some broad idea about class balance, the passive mitigation abilities of tanks and armor proficiencies must be somehow estimated as a form of healing. The amount mitigated must be based on the length of the sum casting time multiplied by an average amount of dps. You can't ignore the effects of the passive abilities on class balance.

The biggest variable is always an unknown quantity ~ the player. Environmental variables also play a huge role. Target resistance to specific damage types completely rebalance the worth of many abilities. The sum total, of the incredible number of variables used to define a toon, and in the calculation of the ability effects results in emergent properties. It's just the result of stacking so many layers of variables. Create enough layers of variables/mathematical operation and emergent properties always result. Predicting the behavior of systems with emergent properties, takes a whole lot more computational ponies than it does to build a MMO server farm. You wouldn't get one descrete value for an ability anyway, you'd get a whole series of clouds. And you'd still have to compare how the clouds overlapped with certain variables assumed. I doubt very seriously any MMO class has ever been designed on paper, that made it into a finished product unchanged. That's the whole reason why they have play testing. That's why no one can really predict how changes will be recieved by the players, or what the players will do with those changes in the meta game. And it's why MMO "balance" is always an ongoing process.
~Master Telagtun Telag of Lord Calypho~

Kitru's Avatar


Kitru
12.28.2012 , 12:19 PM | #6
Quote: Originally Posted by Cleet_Xia View Post
I think coming at from the direction of setting the CD as the first step in the ability design process, is to come at it from a more difficult perspective, that won't necessarily produce a "finished" product any faster or with less effort.
I'm actually not using the CD assignment as the first step. I'm simply looking for a rough formula which can be used to determine what a pure damage ability should deal in a vague sense as a guideline for setting damage values while designing the intended attack strings in a holistic manner. The closest I've gotten is that all pure damage attacks start off with a baseline 1.0 modifier (equal to the base damage dealt by your basic attack without factoring in procs) and a percentage of a further 1.0 modifier based upon the percentage of the CD compared to a 15 second "generic" CD (6 second CD, which is 40% of 15 seconds, would equate to *roughly* 1.4 "basic attacks" worth of damage). This is a *very* vague guideline since it assumes that the damage is dealt roughly equally across most attacks (so it works for classes like Consulars and Knights but not as well for Troopers which tend to deal a *lot* of damage within a small number of attacks and rely upon their base attack more), but it works to some degree.

Quote:
The reason why I say this, is that while casting time is mostly just a variable that is being used to set the "feel" of play for a particular ability, by influencing player mobility. CD I think has been set mostly for balance in an after the fact fashion.
I'm actually not paying attention to casting time since it's more of an abstract applied to lengthen the GCD and moderate damage. When looking at abilities, I'm actually looking at damage per GCD and only within actual practical application (so I'm not looking at TkT for Shadows except for Kinetic Shadows and, only then, when you have 2 stacks of Harnessed Shadows; since it's a 3 sec cast, I'm functionally dividing the damage by 2 to represent the fact that it consumes 2 GCDs).

I also recognize that it's a holistic assignment in the first place, so it's not like CD is deterministic for damage or vice versa. You have to consider the whole of the attack rotation/priority rather than the individual abilities themselves. I'm simply looking at a guideline for a baseline,

Quote:
High Impact Bolt/ Project are essentially the same ability. The same could be said of TK Throw and Full Auto.
I'm not really seeing the similarity between HiB and Project: Project costs a crapton to do mediocre damage and a little stun. HiB deals a lot of damage without the stun and costs comparatively less. TkT and Full Auto are roughly similar however.

There are some "standard" archetypes for attacks that all classes use: Ion Pulse, Double Strike, Slash, and the like that are baseline consumers that deal more damage than the basic attack and have no CD; Stockstrike, Blaster Whip, and Blade Storm that are on CDs but deal substantially more damage for their cost to make up for it.

Quote:
Those are the abilities that are the less efficient abilities in terms of cost/ GCDs/ and the amount of damage done, such as Mind Crush & Charged Bolts.
Charged Bolts actually has more in common with Disturbance for Sages, Double Strike for Shadows, Ion Pulse for VGs, Quick Shot for Scoundrels, and Charged Burst for Gunslingers than it does with Mind Crush. Mind Crush is a "special" attack (generally reserved for proc'd effects making it effective) whereas Charged Bolts is a baseline consumer (since it tends to proc other attacks rather than getting improved by various in class procs).

Quote:
1) For this comparison to mean anything, you have to convert the resource of each class to a common denomination.
2) Any abilities that require a proc to be used, or consume a proc during their cast ~ cannot be directly "valued" as a stand alone ability. The relative value must be inferred by subtracting the values of the other abilities that enable them to be used, after calculating the value of the whole "string" and adjusting the derived value by the ratio of it's frequency.
3) The 1.5 GCD is used for the casting time of an instant.
4) The regen rate used, has to assume that the ability is being used while regen is at the baseline value. As soon as you throw in variable regen modifiers like the concentration proc, regen dubuffs, emergency regen abilities, or being at a low regen level for trooper/ scoundrel ~ these values just go all kinds of wonky. You end up with sepperate values for every ability used with different stacks of buff on self, debuff on target, etc. The base regen value used for the knight has to estimated based on how much focus is generated in an average 60 second period of mixed ability usage.
5) You almost have to assume that DoTs are refreshed at the optimal rate, and not cast on multiple targets on CD.
6) Mitigation effects from abilities have to be quantified somehow, even if it's just the dps that is eliminated by a KB against a melee target. The good news is that you could actually compare the value of a stun to a dps ability or a heal.
I've actually done all of this already except for the last. Since the mitigation effects of certain attacks only really take place in tank specs, such benefits are factored in as part of the survivability calculation based upon the use rate of the given attack rather than as an abstract attached to the damage/threat balance functionality of the attacks. As such, rather than trying to find some way to balance the 5% acc debuff on Force Breach for Force Sweep for Shadow and Guard tanks, it's simply considered to be a static benefit attached to the ability in a mechanically arbitrary manner: it wouldn't matter which ability it was tied to as long as it was tied to *some* ability such that it could maintain the same uptime. The quantification of non-direct mitigation benefits (such as KBs and conditional stuns) are less significant than you might imagine, only really being present for flavor (no one really cares that Project, Force Sweep, or Full Auto ends up stunning the weak and standard targets since those targets are largely unimportant and can be beaten by just slamming them with your basic attack ad nauseum; it's just a tangential secondary benefit that makes some degree of sense to have on certain attacks without really making a substantial balance difference; the "real" stuns are less concerned with their damage and actually follow pretty standard formulas for duration, range, damage, and cost based upon the PvP balance equations).

Quote:
Target resistance to specific damage types completely rebalance the worth of many abilities.
The variation in damage type and attack type is actually relatively simple to quantify: using the standard ~35% K/E DR that ops bosses get allows you to "increase" all I/E damage by a factor of ~1.5 to reach the damage equivalence; attack types follow a similar formula insofar as you can roughly quantify the chance to miss (or itemization lost to make up for said losses) and apply a roughly 5% higher damage to those attacks to reach vague parity.

Quote:
I doubt very seriously any MMO class has ever been designed on paper, that made it into a finished product unchanged. That's the whole reason why they have play testing. That's why no one can really predict how changes will be recieved by the players, or what the players will do with those changes in the meta game. And it's why MMO "balance" is always an ongoing process.
And I highly doubt that any MMO class was designed purely in the abstract without ever first being built upon a foundation of paper. The pure theory represents a baseline of performance and provides a structure for the remainder of the class to grow through the organic process of playtesting and simulation. The quality of said pure theory development has pretty substantial long term effects: just look at Guardian tanks compared to VG tanks or Shadow tanks; it's readily apparently that Guardian tanks didn't really have much in the way of on paper development because most of their functionality is just thrown together from ability archetypes in previous games and then modified later on to achieve numerical balance without much thought towards how it actually plays. Well thought out classes are readily apparent based upon their elegance of design. The intent would be to design a class with abilities that work together in a manner such that the end result would work together seamlessly without the rough hammering that would be required otherwise.
Walls of Text? I *love* Walls of Text!
My New Class Idea
Shadow Class Rep - Suggest/Review Questions Here
Quote: Originally Posted by Fende View Post
Listen to Kitru. Kitru knows all.

Cleet_Xia's Avatar


Cleet_Xia
12.28.2012 , 03:29 PM | #7
Quote: Originally Posted by Kitru View Post
I'm actually not using the CD assignment as the first step. I'm simply looking for a rough formula which can be used to determine what a pure damage ability should deal in a vague sense as a guideline for setting damage values while designing the intended attack strings in a holistic manner. The closest I've gotten is that all pure damage attacks start off with a baseline 1.0 modifier (equal to the base damage dealt by your basic attack without factoring in procs) and a percentage of a further 1.0 modifier based upon the percentage of the CD compared to a 15 second "generic" CD (6 second CD, which is 40% of 15 seconds, would equate to *roughly* 1.4 "basic attacks" worth of damage). This is a *very* vague guideline since it assumes that the damage is dealt roughly equally across most attacks (so it works for classes like Consulars and Knights but not as well for Troopers which tend to deal a *lot* of damage within a small number of attacks and rely upon their base attack more), but it works to some degree.
This is the whole reason why I suggest that you need to pull more time into the calculation to determine the "value". Without including a long period of time into the calculation, you don't really see what the overall effect of the ability is. If it's a 15 second CD - it's a 4 cast per minute. And while it might do signifigant damage per use, it doesn't influence the dps much over the course of a typical battle. DPS abilities with long CDs are intersting to use, but don't necessarily effect balance very much.

Quote: Originally Posted by Kitru View Post
I'm actually not paying attention to casting time since it's more of an abstract applied to lengthen the GCD and moderate damage. When looking at abilities, I'm actually looking at damage per GCD and only within actual practical application (so I'm not looking at TkT for Shadows except for Kinetic Shadows and, only then, when you have 2 stacks of Harnessed Shadows; since it's a 3 sec cast, I'm functionally dividing the damage by 2 to represent the fact that it consumes 2 GCDs).

I also recognize that it's a holistic assignment in the first place, so it's not like CD is deterministic for damage or vice versa. You have to consider the whole of the attack rotation/priority rather than the individual abilities themselves. I'm simply looking at a guideline for a baseline,
Maybe the basline you're looking for is just an average dps number from a parse ~ uninfluenced by AC, gear, or spec choices. Alternatively you could just average together a large number of "naked" parses from different ACs. You could probably just assume that TOR actually works in clock time units of 0.1 seconds, and ignore the hundreths. The engine may not actually be using true hundreths anyway, considering that would be above the resolution of the human eye. So if you just convert everything to dpTs ~ everything fits nicely. Divide longer casts by 10X the cast duration, and instants by 15. At least for damage you get a nice number, for quick comparison sake. But I still think you need to consider how frequently an ability can be used in a practical sense.

Quote: Originally Posted by Kitru View Post
I'm not really seeing the similarity between HiB and Project: Project costs a crapton to do mediocre damage and a little stun. HiB deals a lot of damage without the stun and costs comparatively less. TkT and Full Auto are roughly similar however.
The differences in damage & cost - are just balance changes. The reason why I compared them, is that they both benefit from increased dps output as a result of spec choices in the dps trees.

Quote: Originally Posted by Kitru View Post
There are some "standard" archetypes for attacks that all classes use: Ion Pulse, Double Strike, Slash, and the like that are baseline consumers that deal more damage than the basic attack and have no CD; Stockstrike, Blaster Whip, and Blade Storm that are on CDs but deal substantially more damage for their cost to make up for it.

Charged Bolts actually has more in common with Disturbance for Sages, Double Strike for Shadows, Ion Pulse for VGs, Quick Shot for Scoundrels, and Charged Burst for Gunslingers than it does with Mind Crush. Mind Crush is a "special" attack (generally reserved for proc'd effects making it effective) whereas Charged Bolts is a baseline consumer (since it tends to proc other attacks rather than getting improved by various in class procs).
Exactly - in truth, at the engine level almost all of the instant cast abilities are actually the same ability ~ it's just copy and paste. The only way they differ is in their ranges,damages, and animations. Cetain abilities are given casting times, which makes players choose between using them, or moving their toon. CDs and cost just limit what portion of a rotation they can fill. Sometimes the abilities were given associated buffs, just to add flavor to the game play. Charged Bolts & Disturbance probably originated as "the purple button".

Quote: Originally Posted by Kitru View Post
I've actually done all of this already except for the last. Since the mitigation effects of certain attacks only really take place in tank specs, such benefits are factored in as part of the survivability calculation based upon the use rate of the given attack rather than as an abstract attached to the damage/threat balance functionality of the attacks. As such, rather than trying to find some way to balance the 5% acc debuff on Force Breach for Force Sweep for Shadow and Guard tanks, it's simply considered to be a static benefit attached to the ability in a mechanically arbitrary manner: it wouldn't matter which ability it was tied to as long as it was tied to *some* ability such that it could maintain the same uptime. The quantification of non-direct mitigation benefits (such as KBs and conditional stuns) are less significant than you might imagine, only really being present for flavor (no one really cares that Project, Force Sweep, or Full Auto ends up stunning the weak and standard targets since those targets are largely unimportant and can be beaten by just slamming them with your basic attack ad nauseum; it's just a tangential secondary benefit that makes some degree of sense to have on certain attacks without really making a substantial balance difference; the "real" stuns are less concerned with their damage and actually follow pretty standard formulas for duration, range, damage, and cost based upon the PvP balance equations).
Threat doesn't really need to be considered when it comes to balance, so long as the dps of the tank specs is balanced ~ threat will be by default. I agree with you about the stat boosting passive abilities ~ you used accuracy as an example. The tech trees are full of these, with a crit increase/damage increase /cost reduction for all sorts of active abilities. You're also right about the mitigation effects that only occur against weak mobs in PvE - those effects that only work in PvE vs mobs ~ they don't have to balance ~ no one would notice if they didn't. But not all of those abilities are fringe benefits, the snare effects from Tk Throw or Weaken Mind for example. Both of which do have an effect on PvP balance.

The way I look at it is this.

(Mob damage output - player mitigation - player healing done) - (player damage output - Mob mitigation - Mob healing) = X

If X = 0 the encounter is balanced. Otherwise a victory condition is defined before the event occurs.

Passive mitigation of any type can be a huge part of this. It could be a tanking spec choice, a class ability, a stun/ root/ snare, a buff from an ability used, or just using range against a melee oponent. And in well balanced content, X is going to be close to 0. We consider helaing and mitigation to be sepperate, because we're used to the "trinity" structure. But there never really was a trinity, its always been just "doing damage" and "offseting damage" all along. What makes mitigation different is that it's largely passive, while healing is active. Force armor is considered a heal, because it's spammable on large group of friendly targets, but it's got more in common with a defensive CD.

Quote: Originally Posted by Kitru View Post
The variation in damage type and attack type is actually relatively simple to quantify: using the standard ~35% K/E DR that ops bosses get allows you to "increase" all I/E damage by a factor of ~1.5 to reach the damage equivalence; attack types follow a similar formula insofar as you can roughly quantify the chance to miss (or itemization lost to make up for said losses) and apply a roughly 5% higher damage to those attacks to reach vague parity.
So striike that variable... the work has been done.

Quote: Originally Posted by Kitru View Post
And I highly doubt that any MMO class was designed purely in the abstract without ever first being built upon a foundation of paper. The pure theory represents a baseline of performance and provides a structure for the remainder of the class to grow through the organic process of playtesting and simulation. The quality of said pure theory development has pretty substantial long term effects: just look at Guardian tanks compared to VG tanks or Shadow tanks; it's readily apparently that Guardian tanks didn't really have much in the way of on paper development because most of their functionality is just thrown together from ability archetypes in previous games and then modified later on to achieve numerical balance without much thought towards how it actually plays. Well thought out classes are readily apparent based upon their elegance of design. The intent would be to design a class with abilities that work together in a manner such that the end result would work together seamlessly without the rough hammering that would be required otherwise.
Design on paper - is abstract... but we're on the same page here. Knight was something that was hashed out from the same basic template of abilities, and given a unique resource mechanic just specifically to make it different. The class could possibly feel a lot better if the focus mechanic were changed to a large pool instead of rough blocks. Just to open up the ability choices a little bit more, and allow for a little better tuning of the focus generating abilities. I don't think it was based explicitly on content from another game intentionaly. But I agree that a clear vision didn't exist when they created the class. And I'm sure the more "elegant" classes served as the baseline that knights were balanced to fit with. (it felt like that when I leveled my knight) That's the real obstacle to creating any new class to fit with an existing design for another. And that obstacle gets larger with every class and active ability that is added to the game.

As far as avoiding the "rough hammering" ~ I don't think that you can predict whether or not it's going to be necessary without just shooting for blandly average. An interesting class is going to feel different because it IS different. And TOR has failed to produce a different feel for all of their possible specs. I don't consider it to be something to dislike the game over, but when you have 4 classes, 8 ACs, 24 specs ~ 3 ea heals/ tanks, HOW are you supposed to make 18 (pure) dps specs feel unique? The shared trees are going to feel different ~ but if you take those away that still leaves you with 14 dps specs. The PvE content is balanced around players in a 1:1:2 ratio ~ but the specs are in 1:1:4.6, and the problem gets worse when you consider the viable hybrid dps specs. This game really needs more tanking & healing specs IMHO. You have to be willing to drop the idea and start over if you're having to hammer on it too much, or you risk reinventing the knight. But each level increase give you a little more room to "hammer" out the dents in the base classes. RotHC is a chance for BW to fix some of the knight's problems, and make the other classes feel even more distinct.
~Master Telagtun Telag of Lord Calypho~

Kitru's Avatar


Kitru
12.28.2012 , 04:13 PM | #8
Quote: Originally Posted by Cleet_Xia View Post
Knight was something that was hashed out from the same basic template of abilities, and given a unique resource mechanic just specifically to make it different. The class could possibly feel a lot better if the focus mechanic were changed to a large pool instead of rough blocks.
Guardian tanks (which is the specific spec that I refer to) isn't clucky because of Focus management. If that were true, all knights would feel clunky rather than just the tank spec specifically. Guardian tanks are clucky because they have a lot of abilities with conflicting CDs: Vigilance works quite well because it explicitly operates within the confines of the 4.5 second Sundering Strike cycle (Sunder>2 GCDs>Sunder>2 GCDs, with all but one of the various attacks using a CD that syncs exactly up with that 4.5 second cycle), not to mention that it relies upon multiple abilities with excessively long CDs (considering they're pretty much intended by design to be included in default attack strings) such as Hilt Strike, Force Stasis, and Combat Focus to be resource neutral while still generating decent threat, not to mention Cyclonic Sweeps which reduces the CD on Combat Focus based upon the use of 2 attacks that almost never see actual use by Guardian tanks (Slash and Cyclonic Sweeps) thanks to their high cost, low damage, and the fact that Guardians tanks are already resource starved such that they couldn't really afford to burn that much Focus on largely redundant attacks. Guardian tanks, rather than being designed with a specific attack string in mind, are simply a mash-up of various archetypal attacks balanced roughly based upon formulaic manipulation; unless the CDs are explicitly tweaked, they're not really going to represent the same elegance of design that most of the other specs manage (presumably, the developers are going to address this with the skill tree tweaks in the xpac).

It's *because* of this functional elegance that pre-playtest theoretical design is so valuable. If the developers sat down and thought about what they wanted the attack string to look like rather than just designing abilities arbitrarily, it would've manage a more functional design. As such, I really see CDs as being more intrinsically tied to the construction of the attack string in the abstract and the damage values more tied to the need to apply weighting to the various abilities to encourage following the intended attack string. You build the attack string first, using CDs (both hard CDs by setting a CD and soft CDs by forcing the ability to require a proc or instant-reset when another ability is used) as the method of design and assign the damage values later, using generic guidelines of damage/CD/cost to arrive at something effective, both values of which can be tweaked based upon response via playtest (CDs can be tweaked if the intended attack string isn't effective, but damage values are much more likely to be tweaked to justify the attack string).
Walls of Text? I *love* Walls of Text!
My New Class Idea
Shadow Class Rep - Suggest/Review Questions Here
Quote: Originally Posted by Fende View Post
Listen to Kitru. Kitru knows all.

Cleet_Xia's Avatar


Cleet_Xia
12.28.2012 , 11:09 PM | #9
You and I must play Guardian tank a bit differently. I've never felt like my CD's "gnashed", & I use cyclone slash for generating AoE threat on multiple targets as I make the pull. Not saying I'm playing it well, but I don't feel like I suck at it.

What I was talking about with going to a large pool, is that the focus generating part of the string would vary moar better over the long term, and by tweaking the ability cost the string goes more solidly from
A>B>C>A>B>C>A>B>C>A>B>C>A>B>C>A ~ to ~ E>A>B>C>A>B>C>D>E>A>B>C>A>B>C>D>

A & E being focus generating abilities B,C,D being GCDs (1,2,1,2..... Vs. 2,2,1,3,2,2,1,3)

As it sits right now, if you stick with a stable rotation you either starve or "max out". So a resource nuetral string for Guardian involves what I'd call mini burn phases. It's got a weird "incomplete" feeling.

This brings up some important points -
1) if you tweak the CDs for the defense spec - you wreck the vigilance spec. - but you have identified what makes the Vigilance spec fun to play. And it's that it gives the player options without messing up their play waiting on CDs. They can keep a string of casts tight, without actually planning ahead ~ it's a twitch gamer's ideal spec.
2) people can use these abilties differently, sometimes wildly so.
3) we all like our choices
4) an "intended attack string" - is what ended up making Tk Sage & Gunnery Commando boring. Because if an ideal string exists, much deviation sacrifices too much dps. Abilities on proc add some very interesting qualities to game play, but the proc still needs to let players "roam". I kind of feel like the 10 second duration of some procs is too short, but then again a longer duration might be too easy. If a proc is activated & consumed once every 10 seconds, it eats at least 30% of those 10 seconds.
5) Multiples of the GCD are relevant when setting CDs, casting times, durations, channel lengths. I never really thought about why the 2 piece set bonus for a 1.5 second reduction in HT CD was chosen ~ it's because you wouldn't notice anything less, and 3 seconds would have been impossible to balance.

Now I understand why you are so interested in going at it from a "CD first" perspective.

But I still say CD & casting time are being used to accomplish the same goal - the difference being that casting time can be influenced by alacrity, interrupts, and pushback. A small amount of alacrity may not influence a orderly string of abilities at all, but enough alacrity can create space enough for an extra ability to be cast out of turn. But casting times break the rigid structure to a degree. Mind Crush for example, doesn't interchange well with any other ability in the Sage toolbox, except for maybe Salvation. But neither of those abilites fit particularly well anywhere, and Salvation on rotation is a very boring way to heal. While Weaken Mind interchanges quite well, and even has a 12 second duration ~ for optimal refresh timing.

It's pretty obvious that BW thought the way to go about it was to make sure that a class always has abilities that share the same CD or other abilites whose CDs total the longer CD, so if nothing else, they have forks in the road. Too many long CDs - creates a bursty class, that must spam a basic attack. These are also difficult to balance for PvP, because of their potent "opener". While too many different CDs creates a class with no real rythem to it. Which is why we see the multiple of 3 repeated so frequently with abilities. It's also why many debuffs & buffs on proc work in multiples that don't divide by 3 ~ essentially forcing players to refresh early, less than optimally, or let them expire.

Other multiples of 1.5 give you 4.5, 7.5, 9, 10.5, 13.5, 18, 22.5, 27, 31.5, 36 ~ seems like I've seen 4.5, 7.5, 9, & 18 in the game, as for the rest... I'm not sure. Any of those could be used to design abilities that played differently. But an ability that would balance with a weird "off GCD" cast time (such as 1.6, 3.1, 4.6, 6.1), could easily become imbalanced by alacrity.

Imagine a pair of 2,4,6 second dps casts as a class' abilities - add the talents to reduce the casting times to 1.6,3.3, 5 & then some alacrity gear to make those 1.5, 3, 4.5 second casts ~ the same abilities become real jaw breakers that are used in rotations with instant casts in a completely new way. A few talents could be used to make 3 specs in the same AC require completely different gear. It could be what makes the distinction between a heal spec and a dps spec for another example. Without the alacrity, the pair of abilites can be used efficiently and interchangeably in strings only at every fork- but with the right talents and alacrity... they could fit right anywhere. Power/Alacrity Vs. Crit/ Surge. I'd be surprised if this isn't already in the game somewhere, on one of the pure dps AC's I havn't played.

I Example of the base ability:
1) a 2 second casts fits tightly as part of a 10 second string, with any class design that has 10 second procs or debuffs.
2>1.5>1.5>2>1.5>1.5> ~ or ~ 2>1.5>2>1.5>1.5>1.5>
2) a 4 second casts also fit well 4>1.5>1.5>1.5>1.5> ~ or ~ 4>2>4> ~ or ~ 4>2>2>1.5> ~ or 4>1.5>4>
3) a 6 second cast ~ really only fits in once per 10 second string 6>2>2> ~or ~ 6>4 ~ or ~ 6>2>1.5>
*9 options that make sense

IIOn alacrity/power spec these same abilites become interchangeable with instants at the expense of creating the tightest possible strings that are just right to refresh 10 second procs. Tthe spec also gains an extra GCD worth of space in it's string roughly every 15 seconds, at the expense of passive resource regen during these abilities.
1) 1.5>1.5>1.5>1.5>1.5>1.5> = 9 seconds
2) 3>1.5>1.5>3> ~or~ 3>1.5>1.5>1.5>1.5> = 9 seconds
3) 4.5>4.5 ~ or ~ 4.5>3>1.5> ~or~ 4.5>1.5>1.5>1.5 = 9 seconds
*6 options that make sense

IIIif it's only the 4 second cast that benefits from the alacrity talents & stats:
1) 2>1.5>1.5>2>1.5>1.5> ~ or ~ 2>1.5>2>1.5>1.5>1.5>
2) 3>1.5>1.5>3> ~or~ 3>1.5>1.5>1.5>1.5> = 9 seconds / 3>1.5>3>2> ~ or ~ 3>3>1.5>2 ~ or ~ 3>3>2>1.5 = 9.5 seconds
3>2>3>2> ~ or ~ 3>3>2>2 = 10 seconds
3) 6>3> ~ or 6>1.5>1.5> = 9 seconds / 6>2>1.5> ~ or ~ 6>1.5>2> = 9.5 seconds / 6>2>2> = 10 seconds
*14 options that make sense

It's an overly simplified example, since it assumes the the ability that generates the proc is cast in the same "position" every 10 seconds, But ~ which of these looks like more fun to play?

0.5 seconds of time can dissapear over 10 seconds of casts, just due to the tiniest hesitation between abilities. Pushback can also eat up casting time. 12 second procs would work tighly for the same abilities in either spec. In play "perfect 10's or 12's" may actually be unplayable with less than 11% alacrity. This could actually be used as a mechanism to bar the alacrity/ power spec from using the procs ~ they could be reserved for the crit/surge spec. The alacrity stat required could be tuned down to a level like 5%, so that some of the effects became a reality while leveling, or so that a hybrid build/gearing became viable with endgame gear.
~Master Telagtun Telag of Lord Calypho~