Seeing the Forest for the Talent Trees

We’ve gotten a lot of feedback on our crazy, exciting, and scary talent overhaul, for which we are enormously appreciative. For real and for true. We *want* your feedback on the new talents. That is why we are presenting so much detail so early. While we will continue to iterate on talent specifics, your feedback is an important part of that process. Don’t abstain because you’re convinced that things will change without you. Your input is one of our most important tools for improving the game.

We have seen a few consistent responses from players concerned or dismissive about the model, so we thought we’d take the opportunity to explain the philosophy behind some of our decisions, to provide a better framework upon which you can continue giving us feedback.

1.       "I have fewer choices."

This is the big one, and the truth is it is ultimately correct. You will have fewer choices. But you will have more choices that *matter*. One of the important philosophies of game design is that interesting choices are fun. The word ‘interesting’ is key. Choosing between a talent that grants 10% damage and one that grants 5% damage, all else being equal, isn't interesting (unless perhaps you’re a superstar role-player). Choosing between a talent that grants you 5% haste or 5% crit might be interesting, but more than likely there is still a right answer (and like most of us, you'll probably just ask someone else what the answer is.) Choosing between a talent that grants you a root or a snare can be interesting. Which does more damage? Hard to say. Which is better? It depends on the situation.

This is why we don't have a clear damage, tanking, and healing talent choice every tier. In the case of the old trees, choosing the talent you want from among the talents that don’t interest you isn’t an interesting decision -- it's a multiple choice test, and an easy one at that. Are you Ret? You probably want the damage option. But what if the Ret player had to choose from three healing talents and couldn't sacrifice healing for damage? Now it gets interesting. Worst case scenario is the player just picks one at random because he refuses to heal. However, he has the ability. Maybe he'll use it in some situation. Meanwhile, other players will be happy that they can benefit more from the hybrid nature of the paladin class without having to give up damage to do so.

2.       "There weren't cookie cutter builds."

You're wrong. Next!

To be fair, we did manage to engineer most of the Cataclysm talent trees to include a few legitimate choices. These typically occur when you need to spend enough points to get to the next tier of a tree to get the good stuff. Many specs had 1-4 points to spend wherever they want. That's a huge victory compared to pre-Cataclysm talent trees, but ultimately nothing to really brag about.

It is possible of course to strike a blow for individuality and use a non-cookie cutter build. Ninety-nine percent of the time, these builds are just going to be less effective. The remaining percent of the time, they will eventually become the new cookie-cutter. When players talk about their love of options, I think what they are really saying is they are in love with the idea of having dozens of interesting talents. We just don't think that will ever happen.

Look, we tried the talent tree model for seven years. We think it’s fundamentally flawed and unfixable. We know some of you have faith in us that someday we’ll eventually replace all of the boring +5% crit talents with interesting talents and give you 80 talent points that you can spend wherever, and that the game will still remain relatively balanced and fun. We greatly appreciate your faith, but we fear it is misplaced. It’s not a matter of coming up with enough fun mechanics, which is challenging but ultimately doable. The problem is the extreme number of combinations. When you have such a gigantic matrix, the chances of having unbeatable synergies, or combinations of talents that just don’t work together is really high. That’s not lazy design. That is recognizing how math works.

So given that we don’t think it’s humanly possible to have 40-50 fun, interesting and balanced talents in a tree, the alternative is to continue on with bloated trees that have a ton of inconsequential talents that you have to slog through to get to the fun stuff. A lot of you guys have stuck with us for years, continue to play regularly, and still love World of Warcraft. You are the reason we’re still making this game. We think you deserve better, and we think we can do better.

3.       "We'll still have cookie-cutter builds with the new design."

I am slightly amused by the number of comments that say "The theorycrafters will just math out which is the right talent and we'll all just pick that one." But the theorycrafters aren’t agreeing with those comments, because they know they won’t be able to.

Just to make sure, I chose several specs at random and researched their builds. Sure enough, even with the Cataclysm builds today, you see quotes like “spend the last two points wherever you want” or “choose X or Y at your discretion.” It is “easy” (which I put in quotes because theorycrafters devote a lot of time and neurons to it) to determine the value of a DPS talent like Incite or Ignite. It is hard to determine the DPS value of Improved Sprint or Lichborne. Most of the Mists talents are things like the latter. Now there are still some pure throughput (damage, healing, or tanking) talents in the trees. We expect there will sometimes be a right answer as to which talent to take for those roles. On a fight like Baelroc (one boss, no adds), Bladestorm and Shockwave probably aren't competitive with Avatar. We're okay with that, because on Beth'tilac (lots of adds) they definitely can be and it will depend a lot on your play style and the role you have in the fight. However, given that we know a player can only have one of those three talents and that the synergistic effects from those talents with other talents are limited, it is much easier for us to balance say the healing value of Archangel and Divine Star. Despite what you read on the forums, we actually have gotten better at balancing World of Warcraft over the years.

4.       "No rewards for leveling."

Once upon a time, you got a new talent point every level. That worked okay for a game with 60 levels. It works less well for a game with 90 levels. It probably is totally incomprehensible for a game with 150 levels, should we ever get there. We keep bumping the level cap because frankly it’s fun and we haven’t yet come up with a progression mechanism that will feel quite as good.

Leveling is pretty fast these days and fairly rewarding, in that you see lots of new content and get gear quickly, which is something we have trouble replicating at max level (though stay tuned for Mists of Pandaria). On top of that, you’ll still get lots of abilities as you level up. Instead of having to click Raging Blow, we’ll just give it to you, because frankly if you skip it, you’re making a mistake (or you’re RPing a Fury warrior who has taken too many blows to the head). There are gaps in getting new abilities, especially at high level, because we don’t want players to have to have four rows of action bars to play their character. Again, that is just the blessing and curse of having a game with so many levels.

Third, I’ll challenge the notion of just how interesting it is to get that second point in Pain and Suffering or Rule of Law while leveling. Do you really notice that you now kill a creature in 2.9 GCDs instead of 3 GCDs? (But see below for a bit more on this.) There are some game-challenging talents of course, like Shadowform, but as we just discussed, you'll still get those.

Finally, the reality is that for many players, WoW has become a game focused on max level. Back in the day, leveling a fleet of alts was really compelling gameplay, but for many of the old-timers, there just isn’t a ton of interest in making a second mage or whatever. Hopefully account-level achievements will help with that somewhat, but at the same time, I don’t think it’s realistic to expect all of our long-term players to have thirty or more characters at some point in the future. It’s a fair concern that the new talent system is geared more towards making max level exciting, but that’s also where players tend to spend most of their WoW-playing hours these days. We don’t know yet what we are going to do for players who want to play a monk but just can’t stomach the idea of hitting Hellfire Peninsula one more time, and how we solve the problem when you get a friend to try WoW, only to discover that your pal will need to spend several weeks or months getting up to max level before he or she is ready to join your Arena team or raid group. But these feel like problems we are going to have to solve at some point.

5.       "I like being better than noobs."

It was surprising and a bit disappointing at how frequently we saw this argument. The players in question fully admit that they don't experiment to find the best build. They accept the cookie cutter spec that is offered from a website, but then they use the fact that they knew the cookie cutter to mock players who don't. Intimate knowledge of game mechanics certainly is and should be a component of skill. But knowing how to Google "4.3 Shadow spec" doesn't automatically make you a better player. Sorry, but I’m just going to dismiss this one as an illegitimate concern.

6.       "The talents are all PvP choices."

We see this response from players who say “I don’t care about PvP,” or “raid bosses can’t be snared,” or even “I am a solo player, so I don’t need a defensive cooldown.”

First, a lot of players do care about PvP, and almost every choice in the new talent model will be interesting for them. We are also taking some steps with Mists to encourage more crossover between PvP and PvE as the game once had, so even if you don’t care for PvP now, maybe we can get you interested in the future.

Second, a lot of raid bosses can’t be snared, but their adds and trash sure can be. We don’t do a lot of Patchwerk fights these days. Crowd control, movement increases, and defensive cooldowns are all an important part of raid encounters these days. They are even a part of dungeon encounters until you overgear the content.

Now if you're a solo player or a fairly casual raider and you don't often find the need to use crowd control or hit a defensive cooldown, then maybe the choice isn't compelling. But we think that's a problem with the game. I think it’s a fair complaint that our outdoor world creatures have become a little monotonous over the years. Once upon a time, you could choose to take on that camp of gnolls, or you could try and handle the elite ogres, or you might get a patrolling kobold. While we don’t want outdoor leveling to be brutally difficult, that doesn’t mean that every situation needs to be solved with 3 Sinister Strikes. Imagine a cave full of weak spiders. You can choose to AE them all down, use a movement cooldown to get through the cave quickly, use a defensive cooldown to survive the damage, or use your heals to keep you up. When players use their full toolbox of abilities intelligently, they tend to feel good about their character and the game. But it is our responsibility to engineer more of those situations into the world.

7.       "Spec doesn’t matter."

This is a concern especially for warriors, priests, DKs and the pure classes (those characters who have multiple specs of the same role). What we have concluded is that many players want to choose their spec based on flavor (“I want to be the mage who uses Frost magic”) or rotation (“I like the fast gameplay of the Frost DK”). While the raid buff / debuff matrix and spec utility helps to encourage diversity among groups and discourage raid stacking, it’s also a little lame when the Affliction lock is asked to spec Demonology (against the player’s desire) in order to bring a specific buff. In Mists, we want players to have even more flexibility about which character they want to play. Asking a player to swap from damage to tanking for a couple of fights is acceptable to us. Asking someone to respec from Unholy to Frost just for the debuff is not.

There will still be some utility in the various specs, but less than we have today. You should pick a spec because you like the rotation or the kit. Fire is about crit, Hot Streak, and Ignite. Frost is about Shatter combos and the Water Elemental. Arcane is about mana management and clearing Arcane Blast stacks.

8.       “It must be new to be good.”

This is a tricky one. Specifically, the warlock and druid trees include a lot of new talent ideas simply because we felt like those classes needed them. While we want to make an effort to add some new mechanics every expansion just to keep things fresh, we don’t want to arbitrarily replace fun talents that have stood the test of time just in the name of change for change’s sake. Bladestorm is fun. Body and Soul is fun. Shadowstep is fun.

From a designer’s perspective, the half-life of a new spell or talent idea is fleetingly short. You know how when you buy a new car and drive it off the lot it immediately loses a huge chunk of its value? New game ideas are like that. Seeing something brand new is super exhilarating, but that thrill just doesn’t last. I suspect even by the time Mists launches, we will see a lot of comments along the lines of “When are druids going to get something new? We haven’t seen any new ideas since November!”

It isn’t our goal to come up with 18 new talents for every class. We want to come up with 18 fun talents, and that’s going to mean a mix of old and new. Try not to confuse “shiny” with “good,” and we’ll try on our end not to fall into that trap as well.

9.       “You overhaul talents every expansion. Please leave well enough alone.”

This is another tricky issue, because neither extreme (stagnation versus constant design churn) is appealing, and every individual player (and designer!) has a different definition of where those extremes lie. We changed talent trees in Cataclysm to try and fix some of the underlying problems the talent design had since its inception. We actually considered going to the Mists model for Cataclysm, but we were worried that the change would be too shocking to players, so we went with a more restrained design first. As often happens with compromises, it didn’t fix the underlying problems. Our hope is that this new design solves them once and for all. That isn’t a promise to not change talents for 6.0, 7.0, and beyond. But we hope that an overhaul this drastic isn’t necessary again for a long time to come.

MMOs are inherently living designs that are going to change over time. This is particularly true of subscription models, where players rightly expect to see something for their monthly payment. We don’t think it’s fair to cling to designs that aren’t working just because that’s the design we shipped with. As we have discussed a great deal lately, we will try to limit our big design changes to new expansions, but it’s just not in our DNA to leave something at a B- level if we think we can make it A+.

10.   “You’ve got your minds made up and don’t care about what we think.”

You’re wrong. Next!

As I have said a million times, good games (maybe good anything) can’t be designed by popular vote. Our design feedback process is about making informed decisions. The developers will make the decisions we feel are right for the game, but we’ll do that armed with the feedback from players about what is fun and not fun for them. If you want to provide the best feedback possible, try to be succinct (we get a lot of feedback), try to be specific (why don’t you like something), and don’t assume you speak for everyone (game design, like art, is often subjective). Don’t get upset if we don’t implement your idea -- that’s just not a realistic expectation. Don’t confuse the echo chamber phenomenon that can occur in forum discussions for consensus. Most importantly, try to remember what will be fun for everyone, and not just your character.

Soon -- TM.

One more thing to keep in mind: Playing with the new talent system in-game is really different from choosing talents on “paper.” Some of the decisions we made didn’t come about until we could get into the game and see how leveling and playing actually felt. Once we’re in alpha, many of you guys will be able to give us some more concrete feedback. We understand that, and we’re pushing for doing that just as soon as we can. In the meantime, enjoy the Hour of Twilight.

*Greg “Ghostcrawler” Street is the lead systems designer for World of Warcraft. He role-plays a Fury warrior who has taken too many blows to the head.

@Kohzere: No no no. You're doing it wrong! You don't want to complain, you want to provide constructive feedback about what you believe will work or doesn't seem to work and your ideas on what could be the solution from your viewpoint!

@Bräin: You're misinterpreting what his meaning is there. He's not saying you have to care about PvP, but we do care about it and know others do, so are trying to make the ability to go between PvE and PvP make more sense.