Double Buffered

A Programmer’s View of Game Design, Development, and Culture

The Singular Design of the World of Warcraft Talent Tree

Posted by Ben Zeigler on February 16, 2010

I recently started playing World of Warcraft again, for the first time in about 4 years. I managed to make it up to 77 after 200 hours of play before giving up, which is a lot better than last time.  The only reason I stopped is that I ran out of compelling goals to work towards, as I knew I wasn’t going to hang around long enough to be a serious raider. I was also having a hard time assigning my Talent Tree points, largely out of indifference. Then I realized how impressive that was: for 77 out of 80 levels I was compelled to level up almost solely by WoW’s Talent Trees.

My 77 is a Tauren Druid, so I’ll be using the Druid talent tree as my example. I’m only directly familiar with Druids and Hunters, but I suspect my conclusions will hold just as well for the other classes. Yell at me in the comments if you disagree. So, what is it about the structure of WoW’s Talent Trees that makes them the most successful character development system in the history of gaming?

Directed Goals

As soon as you hit Level 10 in WoW, you gain convenient access to information about 90% of your character’s development choices. You can mouse over the highest-level talent in the game, see its requirements, and learn what it does. It may not be initially obvious how an ability works, but simply by virtue of it being at the bottom of the list and having the highest requirements, you know you want it. This is the single most important component of character development as it pertains to keeping players involved: it gives them a compelling goal.

Luckilly, WoW satisfies this goal by making each bottom-level ability worth the effort. Even better, along the way to the bottom of the tree are a variety of enticing sub goals. In the druid example, Moonkin Form and Tree of Life immediately jumped out as things I knew I wanted. After identifying the goals I wanted to hit, I would then use them as guides for picking my powers as I leveled up. Nearly every point you buy in WoW is working towards several goals at once, via direct dependencies or point requirements. On the other hand, a wide open character system like the one in Champions Online lacks any clear long term goals. Without those goals, a player has nothing to work towards.

Concrete Rewards

The long term goals of a character development system give you a future, but a game needs something to keep players going in the present. Certain games, like League of Legends, give such incrementally small effects per skill point that they don’t feel like a reward. In contrast, talent points in WoW either tend to give small but easily describable global bonuses (1% to all stats is incremental but is clearly shown on your character sheet), larger conditional bonuses (30% damage increase to a common power will be noticed), or provide an usable ability. Of course, this wasn’t always true. Today there’s only a few powers in the Druid trees that are difficult to understand or compute (I only vaguely understand bonus healing), and Improved Mark of the Wild now gives the same bonus with 2 invested points as it used to give with 5.

I played Titan Quest recently and it’s an example of a game with a badly designed skill tree. Taking a look at the Rogue tree, you get multiple points per level and have to split them between a “generic” pool that opens up new skills and improving existing skills. So, at a given level up you get 5 points to choose between adding 35 health, increasing the damage of a 12-rank ability by 7, or picking up a new and initially useless passive skill. It turns out ranking up a skill improves more the higher rank it is, which is something I didn’t realize until literally just now. I never really understood what my skill points were doing, which meant I didn’t get any of the primal psychological thrill that results from direct rewards.

Build Variety

Build variety is hard to get right in an online game, because inevitably the hardcore players will try to flatten all variety out of the game as they “discover” the best builds. If you take a look at sites like WoWPopular or disparate internet forums you’ll notice that certain talent builds are considered to be correct. Variety at the high end suffers a bit overall, but WoW does do a good job of encouraging players to diversify outside of their main tree. Once you reach the bottom of your primary tree, it’s a good idea to start working down a second tree towards a synergistic sub goal. In the case of my druid, the Restoration tree has several important skills for Feral druids, so I had a larger set of possible talents to pick from as a worked on my goals.

The real view of character development variety isn’t visible at a static point in time. Instead, WoW needs to be seen as a living, breathing game. For each expansion (and class-by-class between them) Blizzard has dramatically changed the design of the talent trees to fit with the higher level cap as well as solve various Goal and Reward-oriented design issues. This shakes up the playing field and lets everyone explore the full development space. On an individual level, as a player levels up they can respec their talents for an initially low cost. For instance, I eventually realized that I didn’t enjoy Bear Tanking so switched over to be Cat DPS as my primary. Then to effectively double the existing variety they finally added Dual Specialization. Punishing a player for wanting to change their mind on initial decisions or forcing them to specialize on a style of content (PvP, PvE, Solo, etc) when your game supports several is just insulting in today’s market.

Shareable Choices

The final component of WoW’s character development system is actually outside the design of the game itself. If you take a look at a system like Final Fantasy 10’s Sphere Grid it’s got a good mix of Goals, Rewards, and Variety, but it’s missing something critical: There’s absolutely no way to communicate it to someone else. When you’re building a social game, your character development system should facilitate the social element as much as your world design. Based on community support the Talent Tree structure is objectively the best structure for the sharing of character development information yet developed, and has been since Diablo 2.

Several components of WoW’s design are key to this, rather they were intentional or not. The splitting of a WoW class into 3 trees helps as it leads to “Druid 0/58/13” being a useful shorthand for a player’s abilities and inclination. There’s no mechanical reason for any node in a WoW tree to have a specific x/y location, but the spatial nature of the tree makes it easier to remember and discuss. Eventually Blizzard caught on to what the community was doing for them and built the Armory. Basically, if your character development choices cannot be adequately simulated via a single page web application written in javascript, You’re Doing It Wrong.

Half the Game

In my personal opinion, you can blame the success of WoW on two primary factors: the quality of the world, and the drive of the character development system. The importance of Loot eventually overwhelms the importance of talents, but I think many people underestimate just how much talent trees add to WoW. Most players of WoW never make it to level cap or get an epic flying mount, so for all of them the talent system provides that giant carrot on a stick, the one that keeps millions of tired legs fighting against the treadmill of a level grind. Without it, there would be no World of Warcraft.


2 Responses to “The Singular Design of the World of Warcraft Talent Tree”

  1. Ben F said

    Interesting. I definitely agree that WoW’s excellent advancement design is a key reason for it’s success, and the talent system is a near-perfect example. With my warrior, there was no way in hell I was going to quit until I got Titan’s Grip (the ability to wield 2-handed weapons in one hand)!

    What would have kept you going until 80?

  2. JZig said

    Well, you told me there weren’t any cool non-raid story bits after wrathgate 🙂 That would help. Also, I have an entire quest log full of group-required world quests that I have 0 chance to accomplish because everyone is doing dungeon finder.

    I’ll probably pick it up again with cataclysm, becuase they’ll reset the talent system again and stick in some new top level stuff, and they’re adding over advancement stuff. I think having a top level story event for each expansion that DOESN’T require a guild is a good idea.

Sorry, the comment form is closed at this time.

%d bloggers like this: