Wednesday, June 1, 2011

Dev Watercooler--Cataclysm Talent Tree Post-Mortem

You may have noticed we changed class talent trees for Cataclysm. We changed not just the trees themselves, as you might expect for an expansion, but the entire structure of the trees and the way you choose talents. Now that the Cataclysm model has been in play for several months, the team has been discussing what we like and don’t like about it, and I thought that might be of interest to some of you. As always with this series, this is design rumination, not a list of upcoming changes.
What Worked Well

    The talent trees are simpler now, but without losing a lot of depth. Most of what we cut were passive talents that everyone took anyway, or really lame talents that did nothing.
    Choosing your specialization at level 10 and not having to delay playing your character the way you want feels great. You can play your shaman as Elemental or Enhancement with the tools and bonuses to make either work.
    Specs within a class feel different. This was a big challenge for the DPS warriors, warlocks, hunters, and rogues in particular. Nowadays those specs have different rotations, different strengths and utility, and a different flavor overall.
    Mastery integrates into the trees well. We can delay the complexity until higher level, and we’re at the point now where it’s a competitive stat for many specs (though to be fair, not all yet).
    There are some legitimate hard choices for many of the specs. Usually these come in two varieties: which talent you want before you can advance to the next tier of the tree, or where you want to spend those remaining talents after you’ve hit the bottom of the tree.
    At the risk of catching flak for this statement, I feel that the game is as balanced as it’s ever been. When you look back at the vanilla or Burning Crusade days there were many specs that were just jokes and the difference between the highest performing and worst performing specs was on the order of 30-50% or more. Nowadays, players worry about 5-10% differences. Those are differences we still want to fix, absolutely, but we’ve come a long way. The talent trees have helped us do that.

What Didn’t Work

    I’ll admit there are still a few clunker talents -- those that are undertuned or just not interesting enough. There aren’t many though, and they’re relatively easy to replace.
    On the other hand, the talent trees still have traps for the unwary. For example, a Fury build that skips over Raging Blow is making a serious mistake. That may seem obvious to current players but it’s the kind of thing someone returning to the game after a hiatus might not understand immediately. (After all, you didn’t robotically take Ghostly Strike just because it was a gold medal ability.) While there is something to be said for safe choices, it would also be nice if the talents we expected players to have were talents they always had.
    Some players miss true hybrid builds. (Hybrid in this context means spending near evenly in two trees -- I’m not talking about the more common use of “hybrid” as a tank or healing class.) To be fair, these builds were either not very competitive or were just cherry picking a few powerful talents in order to create something that was likely overpowered, especially in PvP. In other words, the reality of the hybrid build never lived up to the myth. But it’s fair to say that it’s impossible now to have a hybrid build, and we understand some players want them back.
     I said above that there are tough choices within the trees of many specs, but there aren’t very many of them within each spec. Often it can come down to where to spend those last 1-3 talent points. While that was our goal, it would be even more exciting if there were more of those hard decisions. Hard decisions can be painful when you’re faced with excluding an ability or mechanic that’s fun to have. But overall we think hard, exclusive decisions are a good thing. They encourage experimentation and discussion and give players a chance to try out different things, all of which can help keep them engaged.
    Even worse, one potential place to spend those points is in the first two tiers of the other trees of your class, yet those talents are extremely design-constrained. First, they have to be attractive to the main spec using that tree, so chances are you’re not going to find much interest in the healing tree if you’re a damage dealer (unless you want to improve your limited healing). Second, those top-tier talents can’t affect higher level abilities since the talents are available at level 10. Finally, because those talents are available early, they should really be relatively simple to understand for new or returning players. You don’t want to put complex procs with lots of exceptions and internal cooldowns that high in the tree. All of those reasons mean that it’s rare that there’s a true game-changing talent available in those first two tiers. This would be totally broken, but imagine you could spend those last 10 points anywhere in another tree. Much more exciting, huh?
    This is a personal pet peeve, but I don’t like the talents that have a 33/66/100% chance to do what you want them to do. That’s just an awkward way of making a valuable talent cost more than one point. The new Cataclysm talent tree design didn’t cause this problem, but it didn’t fix it either.

The Future
This is the part where I’d really love to share our ideas for how we could address these problems, but some discussions are still a little too rough even for the dev watercooler. When we’re a little farther along, we’ll be able to share more. In the meantime, this is a great topic for further discussion. Players like to evaluate the talents in their particular class, but it’s also useful to evaluate the talent tree system as a whole. It’s an iconic design for World of Warcraft for sure, but that doesn’t mean it can’t be improved.

Greg “Ghostcrawler” Street is the lead systems designer for World of Warcraft. He once spent a summer capturing live radioactive alligators. True story.


Post a Comment