Hey!
I treat this thread as a fun thought experiment. Designed many board games in the past and I find pleasure in theorycrafting. I wish to share my 'vision of class balances' with you, without expecting or wishing for any change in the actual game. The below, long article describes how my idealized class balance state would look like, also explaining my decisions based on the underlying philosophy I followed.
My goal is to offer a fun read to anyone interested. Because of the purpose of this writing (or the lack of it), I chose the creative corner category instead of suggestions. The changes I mention are sometimes radical.
Please bear in mind that I did not play with all of the classes. In case I write about a class which is inherently foreign to me, I rely on people's forum feedback and the TWOW talent tree tool as a reference.
My underlying philosophy at game design is diversity, because it gives variability in gameplay. Let's say that you raid Naxx! You look around, 90% of the tanks are warriors, 7% are brave druids and 3% of people are memeing. My goal is to give diversity to these numbers, making raid experience more unique. Same builds at high level tend to do the same things which leads to repetition in gameplay. However, if you heal a warrior MT one week and a prot pala the next one, it would give a fresher experience of the same raid.
Because of this, the purposes of this theorycrafting is as stated below:
* Making overly uniformized roles more diverse. In this theorycrafting I count on 4 classes regularly MT-ing Naxx and 2 more being able to tank until high-end dungeons or low-end raids.
* Giving each build a viability in PVE raiding. In my opinion it can be boring if you go to MC and you see your usual frost mages, fury warriors etc each week. I'd make such changes after which raid composition becomes much more diverse, allowing fun and yet unexplored mechanics. These theoretical changes usually mean buffs. I think that the lack of world buffs at raids opens a window of build-buffing to a degree - it shouldn't be overdone but some builds receiving some love wouldn't break the game. I only regard things from the point of view of PVE raiding (maybe some PVE leveling). PVP is mostly ignored, unless a theoretical ability/buff/change would make a class absolutely disgusting in the PVP environment.
* I'll use a strength-utility scale meaning that classes providing less utility would be better at the things they do. However, the gap will not be that big.
This will be a long ride and I'll most probably be wrong at many instances. But I enjoy writing this. If you enjoy theorycrafting and you are not afraid of a huge wall of text, I welcome you.
Build changes
Mages
I think of mages as a very strong caster class with some added utility. Arcane goes more deeply into the utility and frankly speaking fire outdamages frost at lategame but mages just respec between these two based on the raid they go to. I'd keep fire mages as they are but would increase the utility of frost and arcane - however, in a different way. For arcane I'd boost their 'statistical utility' - their cabaility to buff up different numbers on party members. Here are some of my ideas:
I'd remove wand specialization from the tree (casters use their wands rarely). It would free up two talent slots. Would use them to create an improved arcane intellect/brilliance talent for extra int on raid members. Would also slightly buff brilliance aura (from 15% to 20-25% for example) making arcane mages better at refilling the party members' mana. Arcane mages would be sought after in raids for their better buffs and because of the mana refill ability.
I'd also make frost mage a more utility-based build but instead of 'statistical utility' I'd give them 'spatial utility' - an improvement in their ability to freeze down enemies and an ability to freeze a non-boss mob solid for a certain amount of time. I'd give them an improved frost nova which reliably freezes raid trash for at least five secs (to support aoe casters) and would grant them a deep-freeze ability which works similarly to polymorph but would have no race-restriction. Boss could not be deep-frozen and only one deep-freeze could be active from a mage at a time. With these additions freeze mage would be responsible of removing a nasty mob from a pack at pull or could help pinpointing a pack of trash mobs for aoe casters to defeat them.
Shamans
Restoration shamans are considered as wide healers. I'd further add to this by changing how healing way talent works, making it possible to proc from chain heals, too. As a last addition would buff mana-tide totem a bit (let's say 2 min shorter CD). Other than that, resto is in a very good shape.
Elemental shamans need some love though. They are utility-based casters but to work properly as a caster DPS they lack some talents. What they need in my opinion: a threat reduction spell similar to healing grace (or a slight buff to tranquil air totem), easier access to nature's guidance for added hit chance, buff to ancestral knowledge for a slightly bigger mana pool and storm reach affecting shock radius, too. A curse of elements-like ability for nature would also be welcome but I'd rather consider it for moonkins. In short I'd keep elementals mostly the way they are while buffing some numbers here and there. Slightly more mana, less threat, easier access to hit chance talent and the ability to shock from max range - also the possibility for a nature curse but it is coming from another class.
Enhancement shamans. Oh boy! Enhancement tree should look very similar to druid's feral tree, enabling a melee and a tank build at the same time. It does not. In my vision enhancement DPS is a melee utility spec with better damage and enh tank is a totally fine option even at Naxx level.
First I'd start with an itemization change at shamans - remove the added spirit from the tier items and add some agility and sometimes (rarely) a + (spell or melee) hit chance there! No shaman talent comboes with spirit and agility could help melee with crits and tank with dodges and crits.
Let's start with the tanking changes. Shaman tanks lack a taunt, an aoe threat ability, a general survivability and unlike warriors or bears they can get oom if a fight gets too long.
For the oom issue, the improved ancestral knowledge could somewhat help. For taunt a next level of earth shock could be added which taunts. For aoe threat my main idea is a whirlwind-like effect to enhancement. A shaman tank could hit many mobs with a rockbiter-ed weapon at the same time as long as the ability is not on CD. That would make a huge difference for threat management. For survivability, would buff spirit shield. Shamans should never wear plate but should be made viable even in mails. An interesting workaround is saying that a shield in a shaman's hand is more effective. I'd add one last change to tank shamans: would rewrite how stoneclaw totem works. Instead of giving a certain HP to the totem, I'd make it invulnerable but it would automatically despawn after a certain amount of time. It could work as an in-flavor aoe taunt for shamans. Also, the fact that shock CDs can be reduced to 5 sec via reverberation talent would mean that shamans can be extremely good at single target tanking (in contrast of paladins who'd excel at multi-target taunting; better threat-scaling at rockbiter would also help). Shaman tanks would also be viable but somewhat clunky in case of multi-target tanking. The nieche where a shaman tank would go is locking down one enemy until the tank's mana runs out - as an ideal scenario, but shaman, as all tanks should be generally viable for most scenarios. As a shield wall proxy for lategame raids, stoneclaw totem could be used to redirect the boss' attacks for those critical seconds.
For enhancement melee I'd change surprisingly few. They have low damage and getting a decent hit chance as a shaman is hard. They also want to DPS but they are forced to give the good CDs like bloodlust to others for optimal outcome. Hit chance gets better if nature's guidance is more easily purchased and if tier items have (altogether) 2-3% hit chance per full tier set. Would also buff weapon mastery talent from 2% to 3% damage per talent. Would change stormstrike, too so that it only buffs your next nature spell.
The most important change I'd make is regarding the shield talents. Improved lightning shield would be improved shields granting buff for each shield type AND the mana reduction from thunderhead for those types. Would remove thunderhead and would add a new ultimate to address the problem of 'buffing myself or the others'. The new talent would make certain spells (weapon enhancements, shields and bloodlust) dual-castable, meaning that if you cast them they are automatically casted on you and on the player you target at the same time. For example you cast a WF weapon and select the rogue. Rogue also gets it (only on 1 weapon). Same with shields and BL. In my opinion this is enough to make enhancement DPS an interesting choice at raids.
Lastly, would consider changing flametongue totem so that it gives spell damage or spell crit chance.
Would slightly change searing totem, too so that it only shoots at targets which are currently being attacked by a party member to avoid pulls and attacks on sheeps.
Druids
Druids are interesting. I think bears are 'almost there' while some other builds need more love. For bears I'd add barkskin even at bear form. Would maybe increase the CD of the bear-barkskin but would increase the damage reduction so it can function as a shield wall-effect.
For trees... I believe that druids are supposed to shapeshift into a specific form to be their best version at a specific role. For example no druid tanks without bear form or does melee DPS without cat form. I think that the base form of restoration druid should be the tree, making the alternative druid healing builds inferior, while also allowing some buffs to the tree druid. What the tree lacks, in my opinion is healing touch (with the 20% discount) and a normal rez. The first option would seriously increase the tree's healing output while the second is a huge quality of life improvement (if received early, way before tree form, just like other healers do). The nieche of the tree druid would be hots plus single target healing - a good tank healer also but not a match to a paladin who has enough spell crit chance.
For balance, I'd consider a better mana regen. Would make reflection talent easier to access. Would also add them a nature 'curse' like the one the warlocks have and also indoor roots. Would also consider buffing insect swarm - maybe removing the damage it deals but adding a crit chance reduction debuff dealt by the target next to the hit chance penalty. Would also consider a fearward-like ability for moonkins because fearwarding is highly needed but is only accessible by very few builds and an effect like 'Nature's Harmony' (let's call it this way) would absolutely be in flavor for a druid. Same ability might be accessible as a tree, too.
For cats, I honestly don't know. Tried them, failed at playing them and I am totally unfamiliar with them.
Heard a cool idea for aquatic form, though, an ability to catch fish with a 10 sec spell. Very in-flavor and a way to get some golds as a druid. I'm totally in favor of it.
Priests
TWOW Team has done a terrific job with shadow priest, I am totally happy with them. However, I'd make some changes to holy, to discipline and to fear ward. Currently every non-dwarven priest feels a bit wrong as soon as people enter the raiding phase.
First, I'dd a worse form of fear ward to each race as a generic priest ability. Extreme amount of mana, slow casting time but making the option to fearward possible for any priest. TWOW has done a similar thing with the Holy Champion talent but the extreme CD makes it impossible to fearward more than one person. I'd either compensate dwarves with a new racial or would just allow them to have a 2-3x more effective fearward as a compensation.
For Holy, I'd rework holy champion. Would remove armor and resistance buff to allow better champion buffs without making the ability too strong. Champion's grace is fine I think, this is what everyone is using right now. Champion's bond is different though. I'd change it in a way that up to 30% of the damage received by the priest would be redirected to the champion. Empower champion just needs a stat buff. Champion's resolve would prevent fear on the champion during the time of blessing (let's reduce it's duration to 2-3 minutes and CD should be 2h) while revive champion would allow one combat rezz on the target from the priest and then the blessing would be dispelled automatically.
Discipline priest is in an interesting spot though. We either make it a DPS priest (like a monk) or we make it a healer build with less raw healing power but added utility. I'd go with the boring option here and would make discipline priest a more paladin-like utility-healer. Imrpoved inner fire, instead of buffing the base talent would allow the discipline priest tio cast it on others (with a stat penalty which is fully removed if all 3 talents are spent here). Improved power world shield, instead of the added damage absorption would allow a fraction of the rage to be built up when the target is shielded. At max rank, rage is generated normally within the bubble. Unbreakable will's resistance percentage would be reduced to 2% from 3% but would give a party-wide, passive effect. These changes would make discipline priests a 'buff-focused' healer, giving away bubbles to thankful warriors and bears.
Warriors
I'd leave fury as it is. It is an excellent choice at lategame raiding. Would also touch protection only a little. I think that warriors being the best tanks in the game with the weakness of worse aoe tanking (though shouting and thunder clap helps) is a valid starting point. They are very good but not perfect so a slot is opened up for others. What they need at lategame is some added threat, preferably in relation to shields to overcome fury-prot which is a build created based on necessity. Giving threat to shield bash and increasing shield slam's threat modifier should do the trick in my opinion. Other than that, these two builds are totally fine if you ask me.
Arms is a different story though. It is a PVP build which was created to be as effective (alone) in different situations as possible. Problem is (regarding this theorycrafting) that at raids each build fills up a specific slot in which the build has to be an expert at. A jack of all trades, master of none gameplay doesn't fit raiding.
I'd make it possible to use sweeping strikes in each stance and would allow further 25 rage to be kept during stance changes via the max level weapon specialization talent if the weapon in question is equipped. Would also give a team utility to impale, adding for example an 5% additional damage taken debuff for 5 sec to a mob critted by an arms warrior, making them a good team utility. Would also move impale deeper into the talent tree.
Warlocks
Warlocks are in a very interesting position because they have 3 half-builds. They never go until the ultimate of a tree, they just choose two from the SM/DS/Ruin package.
In this theorycraft I change this despite thinking that these 3 builds are totally valid and viable. In my opinion they provide a way too similar gameplay, feeling like you play one build instead of 3.
Firstly, I'd increase the talents which are required in the talent tree for these 3 talents to force searching for alternatives. Either that or I'd remove a talent - did so with DS, see below. Then I'd switch the ultimate and I'd imagine 3 vastly different gameplay for the 3 builds.
In case of affliction, instead of dark pact, I'd add amplified curse as an ultimate and would allow it to affect CoS and CoE (50% amplification). Would also improve the shadow mastery buff from 2% to 3% to compensate for the missing DS talent. Would also change nightfall proc rate from 2-4% to 3-6% as a compensation. Lastly, would add a talent which enables the warlock to use 2 curses at a time (to consistently put up the main - improved - curses at raids even if only a fraction of warlocks is an affliction one).
Affliction warlocks would be your best bet to put up curses at raids and spamming shadow bolts. They are the most similar ones to the ones we see today.
For demonologists, I'd remove demonic sacrifice. I'd imagine them as warlocks, relying on their pets to a great effect. Sacrificing them goes against this, specially since later talents give you or to the pet buffs while the demon is alive.
Would move soul link to the place of demonic sacrifice and would add a new final talent, improved demons. It would boost the pets' HP and manabar and would upgrade the abilities of the pets. Bigger stamina buff from imp, more frequent interrupts and purges from the doggie. Would also make voidies torment a guaranteed taunt, making it possible for warlocks to tank dungeon contents and maybe raids. Would also buff some numbers in the talent tree to make demons more relevant.
* Improved Imp would give 15% buff per level
* Demonic embrace 5% stamina for 5% spirit per level
* Improved Health Funnel would give 20% per level
* Improved Succubus would give 15% buff per level instead of 10%
* Fel Stamina: 10% / level
* Demonic Aegis: 33% / level
* Unholy Power: 6% from 4% per level
* Master demonologist: imp threat reduction 5% from 4%, voidie physical damage 5% from 2% (!), succubus 3%, doggie 0,3 resistance per level
I make very drastic changes here to enable a very different gameplay, which is kinda hidden in this tree but was never fully executed. With these changes, during leveling the warlock can be a hunter-like, demon-dependent DPS or a tank. At raiding environment would probably use a succubus for 15% damage increase and would actively use the demon as a pet. Or he would tank. Unsure how viable it would become.
One important sidenote - the list of talents and abilities which could make warlock-tanking work (maybe even at higher levels, our dark magician got some nice armor, stamina and health regen buff, plus voidie's suffering can be used as a shieldwall proxy):
* Searing pain as a sunder armor-like threat generator
* Hellfire as aoe threat
* Voidie is single-target taunting
* Voidie is suffering in case of huge incoming damage
* Added stamina, armor and health regen help in tanking
* All they need is an aoe taunt of some sort - probably linked to improved voidwalker at last demonology talent
I'd make such radical changes here to open up a totally different gameplay for warlock and enable tanking for a class which is wearing cloth but has always desired to have the option to tank. My problem with the original setup (SM/DS/Ruin) is the uniformity of the gameplay. I want to theorycraft three different playstyles here for the builds. Affliction, your classic warlock (in a way), Demonology, your hunter who can also tank, and lastly, destruction, your damage-enthusiast.
In my vision, Destruction warlock would be a dark kind of fire mage. Would remove Improved Shadow Bolt from the Tree and would add a 5 step talent which reduces threat of fire spells casted by the warlock. Bane would reduce the cast time of immolate and soul fire (ignoring shadow bolt) but soul fire would be reduced by 0,6 sec per talent. Would change Shadowburn to be a fire spell (let's call it Flameburn). Improved Searing Pain would also decrease added threat from the spell, meaning that the fully learnt talent would give 10% extra crit chance and would fully remove the extra threat from the spell, making searing pain a spammable spell at raids.
Would also change Conflagrate a bit. Instead of consuming the Immolate spell and dealing some damage in the process, it could do something like this: let's say it is a 2 minute decurse which can only be casted on a target with Immolate on. If Conflag is casted, it deals no damage but gives all fire spells a chance (let's say 2%) to give an immolate debuff to the target, stacking up to 3. That would add an interesting raid utility while not making destruction lock even stronger at PVP.
One change I've forgot. If a warlock has the improved healthstone talent, soulwell should give improved ones to raiders.
Paladins
I think that holy palas are in a fine spot as the best single-target healers of the game, with the weakness of limited/no wide-healing. I'd keep them as they are.
For retribution I'd really like if crusader strike would be reconsidered. Would add it as the last retribution talent instead of Repentence. Would also make it a melee ability which buffs further melee damage on the target (as long as TWOW Team is not considering holy resistance as an idea). Would also consider making their buffs 15/30 min, instead of 5/15 min (would apply this to each build).
Protection is a build which deserves some love, though.
I'd change Ardent Defender, making it an ability which automatically procs if requirements are met but it would have a 3 min CD. So it could activate once every 3 min. New Ardent Defender would prevent 50% damage you face when you reach the 35% health treshhold (for 6 sec) and would automatically cast a divine shield on you. It could make it easy for the healers to patch you up. Could also work as a shield wall mechanic as long as you are not one-shot after a big attack. Would also give prot palas a castable spell for ranged single-target pulls. Paladin tank would be the aoe tank master in plate with an added survivability.
Hunters
I think Marksmanship is fine, we see a bunch of them and they provide decent damage and good utility with kiting, trueshot aura and handling frenzied beasts (would give kiting to Survivalists though). I'd do a similar things as I did with the warlock. Would choose one build (here the Marksmanship) which stays relatively close to the class concept we got used to and would make the other two trees to my liking. My goal here is to have a tanking tree under Beast Mastery and a valid melee DPS/trap utility tree under Survivalist. The only difference I'd make regarding Marksmanship hunters is that I'd add Tranquilizing Shot as a Marksmanship-only talent.
I imagine Survivalists as close-combat units who put down traps DURING fights (traps are on a global CD). Also our kiters. First they'd need a talent which allows them putting traps on a global CD during combat. Then I'd give all hunters an ability to taunt. As a survivalist it can be used mid-fight to make sure kiting doesn't fail and to lure mobs to a specific place, for example on top of a freezing trap. A survivalist hunter would join the fight, melee DPS then set up a trap of his liking and with a tauint he makes sure that the desired mob walks right into the trap. Once it's down, he fights again. The gameplay would be very interesting. A character which works as a classic melee until trap global CD is up, then he ignores the fight, sets up a trap if needed, lures a mob and traps it. The added parry and dodge talents would also help this build to mitigate the added damage after taunting. Would replace Wyvern Sting with an ability which reduces trap global CD and Survivalist would become a very versatile DPS/utility. He also needs some added damage to abilities like raptor strike (let's say to reach 80% of top performers' DPS) and once done, the class becomes a very viable and interesting choice. Would consider adding some custom traps for more interesting gameplay, too.
Beast Mastery hunter would be similar to Demonology warlock in some aspects. In my vision it could be played in two ways: either as a DPS mostly relying on the pet or as a tank. The hunter would also have a single-target taunt ability and they'd mostly rely on their mail armor and dodge plus parry chance to survive. Ultimately they could feign death if pet is second at the threat-meter, making it a valid strrategy to redirect damage. This tree would mostly buff up the pet for additional survivability and damage/threat, while also adding Intimidation, which is a good threat generation and stun ability. Since it is a very important ability for tanking, I'd reduce the cooldown of this ability drastically 1 min -> 10-20 sec.
Spirit Bond could also add survivability, while I'd give the option to tank based on the pet selection of the hunter. Unlike warlocks, hunters can tame a bunch of animals and each are unique. Would create some which are designed for tanking. If a Beast Mastery hunter has such an animal, he let's it charge to the enemy, and the beast builds up some threat. The hunter stays nearby and taunts if something gets loose, kites it back then runs away or if really needed, feigns death to transfer top threat to the pet. Would also allow Beast Masters to control their pet's ability while feigning death for around 5 sec, making it possible to tank while laying on the ground. The hunter itself can also withstand some blows, he can offtank for his pet, but not very long. The build's viability would mostly depend on the added animals and their abilities. For example a gorilla with an aoe threat shout, a taunt ability and a way to mitigate damage for some seconds can make the hunters very decent tanks. If pet abilities and stats are scaled alongside with raid progression (if it is possible) then hunters could even tank lategame raids this way.
Rogues
Here I am in the same boat as with feral cats. The concept of rogue is very foreign to me and the more in-depth opinion I write here, the stupidest things I come up with as a result.
Will give a very generic opinion about the things I imagined.
As far as I know, combat rogue is the one I see regularly and they are quite good but lack utility, besides stuns and lockpicking. I'd leave combat rogues as they are and would change Assassination and Subtlety. Giving very vague ideas because I'm not familiar with this class.
I'd make these two builds more utility-like. For Assasination I'd give further stun and poison options while also giving them the ability to apply beneficial poisons on friendly units. They are your poison-cooking assassins, they also know what's good for you. I'd imagine some buffs they could spread around this way. They could also remove poisons from friendly units. In exchange, lower damage output then a combat one. I also imagined a very specific role for Subtlety - deceiving the opponent. Let's say they receive a talent which allows them to select a friendly unit. For the next 20 sec the threat the rogue generates during combat is transferred to this unit. By lore I could imagine it as a character from the shadows deceiving the enemy, making it believe someone else attacked them. During gameplay, it could help tanks and offtanks keeping threat.