In issue 6 of "World of Warcraft Insider", an interview was conducted with the lead designer of Blizzard, Tom "Kalgan" Chilton. The insider team interviewed him about the class talent review process and why there are changes and improvements to the various class trees, especially on what happened between patches that made the paladin and priest talent trees change drastically. It is good to see the perspective of an important person in WoW to get an idea on why they are changing and balancing these talent skill trees.

In World of Warcraft, the different play styles must create interesting design challenges for the available talents. How do you approach this fine line between keeping talents valid for PvP or PvE play styles across all the available classes?

Balancing talents for PvP and PvE is obviously an important consideration for us. One of our goals is to ensure that when we create new talents, we design them so they are useful in both PvP and PvE situations. Of course, sometimes we come up with a really cool idea for a talent, but discover on closer examination that it would in practice be more useful for one play style than the other. Thus, there are cases in which we feel we couldn't make a talent perfectly viable for both styles equally without changing the design in some drastic way. We try to minimize that, however, as a general rule. We set out to design a talent tree that would prevent players from becoming over-specialized, because being completely locked into one play style by your talents generally isn't much fun, especially when you're forced into specializing in those talents in order to be as specialized as other players of the same class.

What can you tell us about how the talent changes are evaluated and implemented? How much of the design cycle is usually visible to the players on the public test realms?

Actually, the vast majority of the design cycle happens before players ever see the changes on the public test realms. Generally speaking, our talent reviews begin about two to three weeks before the changes become available for testing by the public.

In the early phase of the review process, the game designers responsible for the character classes get together to discuss our thoughts on the state of the different classes and what opportunities there are to improve the talents for each.. At this point, it's also important that we decide what we won't change. Each character class has some kind of weakness; this is a key part of having character classes that are distinct from each other. It's important that we don't eliminate these balancing weaknesses for the various classes. As a result, we want to make sure that the changes we make will strengthen the intended role of the class and eliminate unnecessary frustrations, rather than create an "uberclass".

Once we have a good feel for what we're trying to achieve with the review, we try to identify most of the talents that we feel aren't yet useful enough and find ways to make them more appealing. During this phase we also experiment a great deal with the placement of the talents in the trees to try to make sure they flow well as a player gains more talent points (and support the key tough decisions we want players to have to make).

Once we've gone through the "mad scientist" phase and evaluated many different options, we isolate the ones we felt held the most promise and begin testing these trees. It is after these changes have gone through a few rounds of internal testing with our Quality Assurance department that players are exposed to the first preview on the website and a run on the public test realms. We receive a lot of important feedback during the public phase, and we evaluate the feedback in the context of the designs we tried in previous weeks during our earlier iterations. Based on this "big picture", we decide which modifications should be made to the design. Once we're confident we've accomplished our goals for this review, the trees are locked and made ready for the next patch.

At which point (or points) during a class talent review do you step back and evaluate how the talent changes for this class will affect other classes that fill out similar roles, like priests and Restoration druids?

Evaluating how changes will affect other classes that fill similar roles is a constant part of the entire design process. We need to keep in mind that character classes don't exist in a sterile Petri dish, but instead interact with many other different classes either in groups or in PvP situations. If we go too long without taking a step back to evaluate the impact on the game as a whole, we run the risk of spending too much time on an idea that will have to be thrown out later down the line. That's why our balancing process goes through several phases. When we're first designing how each talent or spell will work, we decide on our philosophical approach to balancing it.

Some of the considerations are the talent's power relative to other talents, its location in the tree, whether it compounds the effects of other talents or spells (including those of other classes), and how important the effects of the talent are to that character class as a whole. Basically, how does having or not having this talent affect the class' performance in its role? Also, if it's an active talent, we first want to determine the actual gameplay situation in which we think the ability will be useful. Once we have a good idea of when and how players will use the ability, we can balance the numbers to support the design. Most of these fundamental balancing decisions begin at a very early stage of the process, but they are also constant factors during the entire length of the design cycle.

We always compare the feedback we receive with the work we've done for this class prior to the public test phase, basically checking if there is an option we haven't already examined. We also take into consideration the feedback we have received throughout the entire balancing process, including all the past classes. Our overriding goal throughout is to make sure each class has three strong talent trees that play well into that class' key roles.
However, each class has a unique set of design challenges, and they also vary in terms of how close the class' talents are to fulfilling our objectives when we begin the design cycle. In other words, some classes already play pretty much the way they were designed to, while some other classes still need a few tweaks until they're really "there". Thus, it isn't surprising to us that some classes end up having more changes made to their talents than others.

Generally speaking, the way to avoid having one "ideal" build for a class role is to have enough compelling talents that you simply won't have enough talent points to get them all. Within the talent builds that are possible with the available talent points, we try to make sure that there are enough abilities with intangible benefits to make it very subjective as to whether some talents are really "better" than others. The mage talent Presence of Mind is an example of this kind of situational talent, since it doesn't have a drastic impact on a mage's long-term dps, yet it provides the player with tremendous utility for handling difficult situations, or providing a very short-term increase in burst damage.

For the last few patches, the talent review has focused on one class per patch. How do you determine which class is next?

Our process for determining which class will be "next" is actually rather intricate; it's mostly based on what class I'm playing at the time. *laughs* Ok, that's not actually true. The decision as to what class to look at next is based on a combination of our own feelings on the development team as to which class "needs" to be reviewed the most and what we feel would best serve the game and our players as a whole. Also, while the talent review mostly focuses on one class at a time, we still keep an eye on the other classes to check for any changes or improvements that need to be done.

Can you tell us which class will receive the next talent update and maybe also add a little bit about what your plans are?

Well, right now we're in the process of working on the talent review for the mage. Some of the aspects we're looking to improve include easing a mage's downtime between fights a bit and making a Fire/Frost build a more viable combination than it is right now. (Currently most magi feel that having anywhere between 11 to 21 points in the Arcane is a "must").
To accomplish this, it's likely that we'll take a few of the "must have" abilities from Arcane, and make them available to all magi. (Some likely candidates are the talents Evocation and Improved Arcane Explosion.) From there, we'll also look at ways for players to be able to create interesting combos between the Fire and Frost trees. Hopefully, these changes will further polish the general play experience of magi while also opening some new, interesting talent builds.

I think it's a little of both. I feel like our talent trees are getting to the point of being very solid, but keep in mind that the level cap will be raised in The Burning Crusade, and we plan to extend the talent trees in the expansion. Consequently it is inevitable that on one level or another, we'll feel compelled to continue to tune the balance of the various talents from time to time in order to make sure that the balance "keeps up" with the evolution of the game in general.