Thoughts on Talent Trees

[ladubois 2]

So I really like the concept and feel of talent trees like were once found in WoW and SWTOR, but I do recognize that they have flaws. The two most prominent being the difficulty they create for maintaining class balance, and the road block they create for creating class/spec identity.

 

The first is fairly obvious - more "moving parts" increases the likelihood that something is going to break. Also, I've uet to see any game that manages to have talent trees that don't end up having "road maps" for a decisively superior build than all other options (which is really just a specific manifestation of the aforementioned breakage). Choices should allow for increased variety in playstyle, not a test to see if someone's looked up a guide or not.

 

As for the second, this is basically what motivated Blizzard's scrapping of talent trees in WoW (which, much as I like talent trees, I do feel was for the better in this case). The WoW/SWTOR-style talent trees basically illustrate a dichotomy where if you allow players to pick talents from any spec, your core class abilities have to support all builds - meaning that, for example, all druid are still going to get Regrowth, Swipe, and Starfall, even though no one can effectively use all of those with the same build. This causes a similar problem to the previously mentioned one where the player has way more choices than they need that they then have to sort through. And while in this example, it's pretty easy to figure out which you should put on your hotbar, given how different the three specs are, classes without different roles for each spec can be much trickier. Additionally, abilities that are central but unique to a given spec require spending precious talent points to get to, basically like 3e D&D's infamous feat taxes. Even setting all that clutter and tedium aside, I like specializations each having their own clear identity.

 

Anyway, I think I've come up with a solution for how to design my talent trees that allows them to cut down on clutter, maintain distinct identities, and also allow for multiple distinct playstyles even for characters with the same spec.

 

To illustrate this, let me introduce you to one of my classes, the Animist. The animist is a lightly-armoured caster class that controls the elemental forces of the world. The specializations open to the Animist are Fire (a DPS spec that focuses on DoTs and burst damage), Earth (tank spec that uses armor buffs and self-heals), and Water (healing spec that focuses on steady group heals, and can essentially average out the party's hp to make its healing more efficient).

 

Now, here's what I'm thinking for the talent trees: Your specialization is a discrete choice that grants you access to spec-specific abilities while leveling up, as well as that spec's talent tree. Each spec's tree being made of the typical four columns, with the middle two making up what is essentially a main "branch", and then two side branches which each essentially overlap with the other two specs. These overlapping beanches are identical to each spec that has it. An Earth Animist would, for example, have access to an Earth branch that focuses on improving their damage mitigation abilities, a Magma branch (overlapping with Fire) that improves their AoE damage spells, and a Life beanch (overlapping with Water) that improves their self-heals. There will be sufficient talent points and talents for a fully leveled Animist to take all the talents in their main branch, and a few in one or both (undecided - will require some playtesting) of their side branches, or all of one side branch, and about half of their main branch. The end result is that, while having roughly the same strategy, a Magma-leaning Earth Animist and a Life-leaning one would be quite distinct, with the former having greater DPS, but would be more taxing on a healer than the latter would be. However, thanks to the variety of class types and specs, neither of these builds is definitively better than the other - a Water Animist would probably love to have a Life-leaning Earth Animist tank, sonce their self-heals reduce the need for targeted healing, and allow them to focus on their strength with group heals, while a primarily single-target healer may rather have a Magma-leaning Earth Animist whose AoEs will keep the attention of stray mobs away from squishier allies, so they can just be sure they're keeping the tank alive.

[lonequeso 1,921]

Personally, I love skill trees. They can be hard to balance, yes, but the possibilities of customization is something I value highly in an RPG. Being able to spend skill and stat pints and having many choices within equipment is the holy trinity for me. A good skill tree should have branches so the player at least a couple paths to get to certain skills. Usually, the later skill int he tree are superior in one way or another, but not always. Some early skills are essential. For example, one of the first skills for a mage in Dragon Age 3 is a barrier that comes in super handy throughout the entire game. One of the warrior's first skills deals more damage, the lower your health is while another shoots a grappling chain at an enemy do you can pull those pesky archers and mages to you. 

 

The Dragon Age games feature my favorite skill trees especially 2 and 3. Each class has several trees. Most are readily available form the outset, but there are specialization classes that are unlocked, too. The latter are a bit superior to the others, but not so much that you'd want to rely sole on them. Not every tree should necessarily be used by a character. My sword and shield wielding warrior doesn't need to learn any archery skills. The reason I really like the trees in these games is because there is a good mix of active and passive skills, and all active skills can be upgraded once. I've spent a lot of time pondering which tree to spend a point in and whether to unlock an active skill or a passive on or upgrade a current skill. The cool thing is it's pretty easy to reset skill points so if you don't like the way you set your character up, you can rebuild them. That comes especially handy after unlocking specializations. 

 

As you pointed out, some combinations of skills are always better than others both for individual characters and to set up combos. One character puts an enemy to sleep and another hots them with a skill that deals bonus damage to sleeping enemies. Stuff like that. yeah, someone can look up a guide and see what the best combinations are, but personally, I like tinkering around and seeing what works well and what doesn't. There are so many various strategies you can use, specially if stats and armor are customization as well so your play style still effects how you build your characters. For example there's a class that specializes in laying traps. I'm sure there's some really great awesome strategy you can imply with them, but I don't really like traps so I don't use that tree. Don't worry about people suing guides to get the best builds. You're never going to be able to create a balance so perfect that it makes guides useless. Just focus on making the classes balanced both within themselves and with the other classes. 

 

The way you make a game challenging is how tough you make your enemies. You got all these tress with all these awesome strategies for the player.

How are your enemies going to ruin them? Giving enemies a diverse enough skill pool to effective combat all the player's skills is paramount to making battles worthwhile and challenging. Good AI should be able to recognize certain scenarios. The player just launched an AOE skill? Is your enemy smart enough to try and get out of the area of effect? You poison an enemy. Is the healer smart enough to see this and remove it? It can get really complicated really fast especially if you want your enemies to use complimentary skills, but very much necessary. Remember my previous example with the Sleeping enemy? Why can't the enemies do the same to the player? Bottom line is the more robust and diverse the player's skill pool is, the more the enemies' must also be. That's how you create balanced and challenging fights that force to player to formulate strategies instead of spamming the same skills over and over. 

 

Another important aspect is to make it so enemies can't simply be felled by two or three hits. That's one thing I really don't like about the Pokemon games. It's really easy one one or two hit an opponent. So much so that the faster Pokemon usually has a severe advantage. The longer it takes to defeat an enemy the more opportunities they have to adapt to the player's strategy or receive support from their allies. It keeps battles fluid so it's harder for the player to rely exclusively on a handful of skills and be successful.

 

So there's my thoughts on the subject. As you can see, strategy is very important to me. 

[ladubois 2]

Well you touch on a few other design elements I've worked through. First off, I should point out that the vast majority of my talents will be passive effects. For the style of play I'm going for, characters' core abilities need to be more-or-less guaranteed, and I don't see much use to cluttering things up with a dozen variations of "deal X damage of type Y" and the like. I'm planning for each class will have about a dozen abilities or so. Half will be common across all sepcs of a class, and only about half (independent from that first half) will be the abilities you regularly use. In my experience, that's usually about all anyone but the absolute most hardcore players uses anyway, regardless of what type of game is being played or how many abilities are available.

 

Aside from a capstone, virtually all talents will simply modify your existing abilities or grant passive bonuses. Additionally, because of my distaste for "feat taxes", the power curve of individual talents will be slight to nonexistent. Your very first talents will be just as useful as your highest level ones. They'll probably tend toward the more "boring but practical", so that higher-level talents will be more exciting to get, but a humble 1% increase to all fire damage still benefits you at level 50, just as much as it did when you first got it, even if that new talent that grants your Fiery Wrath spell a 20% chance to deal extra damage is more fun to have. This approach also helps make mixed builds just as viable as focused ones, since you aren't giving up more powerful abilities to go halfway up two branches as opposed to all the way up one. Remember that since you always have your lower-level talents, simply getting another makes your character more powerful overall, even if that new talent isn't, on its own, more powerful than all your previous ones. Also, I'm probably going to give the talent trees some spider-webbing, granting multiple paths to the same talent. This will probably take the form of each talent taken granting access to all adjacent talents. Sort of like the system in FF12.

 

Second, concerning the length of battles. Consider the alternative where every time you run into a random encounter you have a lengthy fifteen-turn battle to fight. Simply getting to Veridian City from Pallet could easily take an hour or two. I've got a couple of ideas for how to solve this particular predicament, but it'll take some testing to figure out which solution and how much of it works best. The main ideas that I have in mind are wandering monsters so that players looking to avoid encounters can do so, and exponential character growth that allows minimal grinding to let players smash through lower-level trash mobs with ease while still being properly challenged by at-level ones.

 

Anyway, I'm using my phone to post this so that's all I say for now.

[lonequeso 1,921]

Lots of random encounters annoy me, too. I prefer quality over quantity though the simplest enemies probably shouldn't take fifteen rounds in a turn based game. Personally I don't like them ending in one or two rounds either because it becomes boring really fast. Oh look another Caterpie. Lemme use Ember for the 20th time. At least in Pokemon, you level up pretty fast so those quick battles can get you a lot of XP quickly. Still, I prefer more challenging battles than mindless skill spamming  Most of our previous examples were regarding action-Rpgs. Turn based is an entirely different animal so the things that work for skill trees in action RPG may not translate as well when it's turn based. Especially, if it's pure turn based (no ATB).

That being said, the number of skills you have feels about right. You usually don't need a super long list of learned skills. It can quickly become cumbersome or just unnecessary if you're not careful. I'm not a fan of systems that have skills that basically become obsolete once you hit a certain level. You don't want pigeon hole players with few active skills either. If you get creative with damage formulas yo can make a really diverse set while keeping a nice mix of passive and active skills. 

[ladubois 2]

Yes, I realized a bit too late that where I left that point probably made it seem that I was planning to just have quick battles. Quite the opposite, or else there'd be no point in having a talent tree. This game will be focused on two things: exploration, and dungeons. The world will be large and expansive, and filled with relatively weak enemies. Challenging enough at the early levels when you won't have many abilities to strategize around, yet, but trivial at higher levels. Aside from some pocket change and nominal XP, their primary purpose will be to simply fill out the world and keep it from feeling empty. The real challenge lies in the dungeons, of which there will be many, and which will easily take an hour or more to get through, even if they only have 6 or 7 battles (they will vary in length - this is just the minimum). But this is all tangential to the topic.

 

P.S. It's going to be CTB - the potential for tactical variety of ATB, but with the ability to give players a moment to think and actually capitalize on that potential of traditional turn-based. I honestly don't know why more (professional) games don't use this system. FF10 and the DS Digimon games are the only ones I know of.

Edited March 25, 2017 by ladubois

[Rezanta 373]

Well, let me start with saying I like the amount of homework you put into the skill trees.

 

I, to be sadly honest, dislike skill trees. In Crystal Saga, for example, it's basically you pick one side or another with the trees, then automatically, it makes you forcibly put your stats to that tree's bidding. That's really the only major negative I have. The rest is small, personal feelings.

 

Some skill trees are really good in what they are aimed to do, but need a lot of testing to give players the freedom they would like. I hope this helps.

[Kayzee 4,032]

I know it's been a week or so, but I wanted to give my opinion on talent trees. I can understand the idea behind them, but to be honest, 99% of the time they just annoy the heck out of me. Why? Because too many games do this thing where in order to invest in good skills/talents you have to put a bunch of points in other skills/talents that are functionally worthless or not conductive to the player's play style. Look, maybe I don't want to invest my precious points in that useless light spell that I don't need just to get a fireball okay? Yeah I get it, you need practice in simple magic in order to do any of the good stuff, but seriously I rather just invest in a stat or skill level to make the requirements and then buy it directly or something. It's so much easier just to have a list and let you freely invest ion what you want when you want. Not to mention that most skill trees are nonsensical and force you into particular roles.

[lonequeso 1,921]

That is one drawback, but good skill trees keep that to a minimum. The really good ones have a lot of passive bonuses bridging those gaps so there are less active skills you may not want/need. They also allow at least two paths to most skills so you have a couple options. The very best trees are in games where every active skill is good in there own ways. Skills are well tested and balanced so none are useless. Then it just comes down to your play style regarding what you choose. 

[Kayzee 4,032]

I just don't find that to be true in my experience, but whatever. Mind you I don't have any problem with the basic idea of unlocking skills or classes or such in a tree-like way, I just don't like forced investment in skills I don't need. And there is a difference.