Talent Trees and The Graph

Talent tress and character progression through leveling is something that jumps to the front of the queue from time while I level in WoW or other games.

So after reading a response post by Ghostcrawler quoted below I thought I would bring this subject up.

For context the post he is responding to in the thread: Least favorite 3 talent trees by design?

Does a Talent tree have a majority of passive percentile increase talents? 3% to this, 5% to that, 10% over here, 15% over there and 70% chance to prevent here etc.

If Yes, it’s a boring tree.

If No, it’s a fun tree.

I love talents that DO “stuff”. It doesn’t necessarily have to be an actual new spell or clicky, but it should have some actual gameplay changing effect. Speccing XYZ should mean I alter how I would play compared to if I didn’t spec XYZ.

It’s when you dump dozens of points into a tree just for raw throughput, that’s when things get dull. When the only difference between speccing and not speccing the tree is a mere increase or decrease in the numbers.


We agree with this to a degree. We have put a lot of passive increases in for a few reasons, including:

1) It’s the main way for hybrids to distinguish themselves (otherwise Elemental heals as well as Resto).
2) They are popular in a wide variety of situations (other than “Oh that’s a PvP talent.”)
3) They are no-brainers, which to be honest, a game as complex as WoW could stand to have a little of.
4) They are very easy ways to tweak healing, damage or tanking when a particular spec is underperforming or overpowered.

But while all those are laudable goals, in the end we’re not sure it’s worth it. We wonder if the talent trees would all be more interesting with fewer passives and more interesting talents. Consider that even the situational talents wouldn’t feel so expensive if they weren’t competing against a 5% damage talent. Cherry picking talents in another tree wouldn’t be so scary either since you wouldn’t be sacrificing pure dps etc. to do it.

I’d actually look at the warrior Prot tree again, especially comparing it to the BC version. There are still several passive talents, but we combined many of them and got rid of really lame ones like Defiance (+ threat). We have ambitious plans to get more talent trees in that vein.

I agree with the philosophy that GC is discussing here, that the talent trees will not really be all active or passive talents.  There are just some cases that to differentiate out a spec you need to have passive buffs for feel.  The flavor is in the active talents, those that actually change game play.

What has me concerned when looking forward at the next expansion what ever it may be, is the depth of the current trees.

One of the reasons Dual specs are essentially required in WoW, is due to the depth of the trees.  Characters have to specialize into the end of a tree to really be worth a grain of salt.

Classic WoW you had the [31/20 | 30/21] specs, then with tBC we got the [41/20 | 40/21] now with WotLK we see specs like the following  [51/20 | 51/17/3].  There is one exception in WotLK which is a warlock spec that is [41/30].  In the next expansion do we really need to have another 2 tiers added to each of our talent trees?  When planning for the next expansion we have the perfect opportunity to do something different from what was done in the past but still similar enough so it does not need to be reexplained how it works.

What can be done in the next expansion to continue a progression path while leveling that does not involve making the talent trees deeper.  There are 2 obvious choices.

Tree Options

Broader Trees

Broader trees are almost a no-brainer, players complain about bloated trees but bloated does not imply broad nor does a broad tree imply bloated.  So what does broad imply without bloat, not as simple as the term but essentially that decisions matter and there is no real right way.   Due to the tiered nature of the trees this will be a bit more difficult than just adding talents but may be a good idea.

New Trees

Rather than making existing trees deeper or broader we add new trees to all characters.  This is essentially similar to the AAs in EQ.  I say similar because the AAs represent a generic set of skills that characters can learn to improve.  I have not seriously played EQ so can give details to the system.

For WoW I would think the new trees could come in 3 different flavors or even have several implemented.  The first type would be the Racial Tree, this tree would be available from level 1-10.  I would think the racial tree would be a simplified tree like those created for Hunter pets, except every character has one.  We would also gain talent points that are spendable only on this tree from 1-10, maybe even tie an extra 1 or 2 to the nursery to get it going.  This would not add that much undo complexity to a new player joining wow and it would be nothing for alts or returning players to pick up.

Another option for a new tree would be a generic tree that has nifty things available to all players.  This tree would most likely only be used by levelers and non-raiders, it really depends on what is put here.  Due to balancing issues I don’t think a generic tree not tied to race/faction/class would work.

Another style of new tree that could be added would be cross spec trees.  These extra trees would require X# of talents in tree A and Y# of talents in tree B to “unlock”, it could also be tied to specific talents as the perquisite.  For a warrior example have a small tree that requires Titan’s grip and Shield Mastery.  This would be the Tanking with a 2-handed weapon and shield tree.  A better example would be from the mage class and an actual viable elemental build.  Have it require Pyroblast and Icy veins and then open up a new progression path for walking the line between Fire and Ice.  Something with the warlock Demon/Afflic or Demon/Destro could change the type of demon you metamorp into.  The combination and capabilities when a path is created by combining existing themes in new ways are nearly endless.   Hell then Lissanna’s Restokin build which with another 10 talent points allows for both Moonking and Tree forms is at least available for leveling without some other help.  But some of this potential I fear is lost due to dual specs.   The only thing would be that the combo new talent tree would be as viable as specializing in only one flavor.

Now after all of this and my title being and the Graph what does a graph have to do with any of this.  Well that is where the next section and set of options comes into play.

New Structures

The Graph

A graph is just a set of nodes connected by edges.  A graph does not even need to be fully connected so in a sense we are already using a graph structure for our talent specs.  In game it just happens to be on 3 different tabs to filter out the other information that is not relevant to the specialization at hand.  Now my favorite example of a progression graph has to be the grid from Final Fantasy X.  I used and abused that system so wrongly when I use to play that game.  I had a save file from the game that I had over 150 hours on.  I got sick of the mini games to get the ultimate weapons and ended up farming the alternate uber boss for enough dark matter to make analogs of them for Titus, Lulu and Kimahri. I eventually got Titus’s real ultimate weapon, all I have to say is I hate those damn birds.

Anyway back to the graphs.  The unique thing that having a graph would allow is for a much better blending of lower tier talents.  We could even go farther by making some talents count for more than one tree as you get today.  So rather than 9-12 first tier talents we could reduce it to 6-9. For example in the druids you would make the following.

Furor = Feral & Balance points.
Mark of Wild = Resto & Feral
Nature’s Focus = Resto & Balance
Genesis = Resto & Balance
Thick Hide = Feral & Balance
Intensity = Feral & Resto

I know some of these are not tier one talents because there has been so much QQ about resto druids in PvP but if there was a paradigm shift in leveling progression some of this would need to be re-evaluated.  I think a graph structure would work well but there would need to be a lot more thought and discussion to make it fit.  Essentially I think looking forward at another 10 levels and trees 10 points deeper, we are going to see a lot more balancing issues.

Posted in World of Warcraft

One comment on “Talent Trees and The Graph
  1. Chris says:

    Something like this was talked about during Wrath discussions, probably on the beta forums. Also adding in something of a pvp tree, even a general one. I can’t link to the forums at work, but I recall them.

    The real issue comes down to balance. Adding more trees, or combining trees in any way creates more balance problems. The trees are doing fairly good right now, class power is different at different levels due to talents available, but in the end they turn out to be roughly the same (or as close as Blizz will make them). Blizz really does try to make the trees different, that is why not every caster class has a +10% damage increase in t2, reduced mana cost in t4 etc etc. The balance is much easier when you can place the more powerful talents out of reach.

    Take the warlock specs you mentioned (40/31 and 41/30, those specs exist because blizzard did not make high end destruction appealing to many people. Chaos Bolt did not do enough damage to justify taking Fire and Brimstone (because this talent sucks and was counter intuitive to the spec). Best talents in those specs are conflag and the Felguard(FG), now if you could get them both how the hell can you balance that, without spending a ton of time on it? Hell, Blizz has actually had to nerf conflag once already (and as all nerfs this comes after buffing it to make it actually useful). I also believe that GC has mentioned the talents are easier to balance when it is known that everyone wants/needs the 51 point talents. The partial specs were never intended.

    I don’t play many melee classes aside from the obligatory DK, so I can’t comment on them as much though I have a feeling it would make Rogues OP in pvp again… somehow…

    As for druids personally I just think they should have quests for the forms. Give everyone moonkin and tree form and just make the talent trees boost them to “useable” levels. Like tree of life form has a 1% reduced mana cost of all spells. Moonkin does 1% more damage for all damaging spells. On my druid I spend most time in caster form anyway:s, maybe I druid wrong.