About Me

My photo
Seattle, Washington, United States
I'm an old time roleplayer who became a soldier who became a veteran who became a developer who became a dba who became a manager who never gave up his dream of a better world. Even if I have to create it myself.

Saturday, December 22, 2012

Class Design: Diablo 3 vs Torchlight 2

So I've played Torchlight 2 and Diablo 3 both, like I'm sure many folks out there have. I don't want to get too much into an argument about which one is better - however, I do want to talk about the differences in classes from a game design perspective.

Diablo 3
In this sequel to the best selling ARPG of all time (Diablo 2), the designers took a very different approach to class design. Some people have liked this quite a bit, and others have thrown venom at it. The truth is, the designers clearly wanted to explore new markets and new business models, and the game reflects that.

First, the idea of a dependency based skill tree was eliminated. In Diablo 2 (for those of you living at Tau Ceti over the last 20 years), characters frequently had to have at least one skill point in a prerequisite skill prior to getting more advanced and powerful skills. Diablo 3 was having none of that. The designers instead went with a progressive level-based skill acquisition - so, for instance, every witch doctor that hits level 16 learns Spirit Walk.

The choices for each character come down to which skill of the ones they know are active at any given point in time. Only 4 active and 2 passive skills can be active at the same time, and so the idea is that different characters will be unique-ish from the combination of active and passive skills their player has chosen.

Allowing further detail is a Rune concept, which gives each of the active skills a choice of a boost, so for instance Spirit Walk could have either Jaunt - which gives a bonus to the duration - or perhaps Healing Journey, which heals the character when the skill is used. Only one Rune can be applied to each skill at any given time.

What skills are active is easily changeable - it can be done at anytime, suffering no penalty beyond cooldown times and perhaps the loss of the Nephalem Valor buff. The latter is only of concern at the maximum level and relates only to endgame loot farming.

I'll also note that Diablo 3 doesn't allow for custom distribution of attribute points. Every level 30 Witch Doctor has the exact same base attributes.

Torchlight 2
TL2 deviates from Diablo 3 quite a bit, and represents a different evolutionary branch from Diablo 2.

First and foremost, it's important to note that the designers at TL2 did one thing the same as D3 - they got rid of the prerequisite concepts for advanced skills. It's interesting to me that, even though virtually everything else is different, that this one decision point was the same. I think that a lot of players found the prerequisite system somewhat non-sensical; in later patches of Diablo 2 the designers even added in a synergy system aimed at this awkward aspect of the game. The synergy system let earlier skills provide bonuses to later skills (like ranks in fire bolt giving bonuses to fireball), so that players wouldn't feel like their earlier choices were a complete waste of skill points.

However, the TL2 designers did keep the concept of skill points, unlike Diablo 3. So, a character might have 1 rank in Shield Bash, and another character might have 10 ranks in the same skill. That second character is much better at smacking things with a shield, but presumably the first character is better at something else.

It's also perfectly possible that a character will have zero skill points in a particular skill - so they might not know how to Shield Bash at all. Each skill has a minimum level requirement, but there is no automatic learning of every skill simply from being at that level. And, there is no reallocation of skill points aside from the "oops" allowance of respeccing the three most recent skill points expenditures.

Additionally, attributes in Torchlight 2 are fully customizable. An Embermage can throw everything into Strength, or an Engineer could throw everything into Focus.

The greatest distinguishing factor between the approaches these game designers took, in my opinion, is the replayability of the game. Diablo 3's identical progression gives no value to creating a second Witch Doctor - why play through dozens of levels again when every skill is already known? The designers at Blizzard clearly wanted to encourage endgame play (and loot farming) rather than the creation of additional characters.

Torchlight 2, on the other hand, has a significant amount of potential in additional character creation. It also allows for a lot of creativity in character progression because of the customizable attributes - I can create a melee Embermage or a spellcasting Berserker if I want, and it is perfectly viable. This type of approach is nearly impossible in D3, where the characters simply are what they are and struggle to leave their well-defined design boxes.

Applicability to Tabletop RPGs
Being able to create exciting character concepts is important to many tabletop players. Noone wants their fighter or rogue to feel like every other fighter or rogue ever made. Different game systems approach this in vastly different ways, from OD&D's super-light fill-in-your-own-blanks approach to GURPS and its complex point-based system.

In my Phoenix RPG, as I consider my own class design (or "portfolio"), I've looked to a lot of sources from both video games and tabletop systems. My design goals are to have strongly defined classes, but with strongly defined customizations that let players create those more unique visions that noone at the table has seen before. As a GM, it gets me excited to see a character that I never dreamed of before because it opens up a ton of story ideas right off the bat.


  1. What I hated in Diablo 2 was getting into the next difficulty level and realising my character was hopelessly underpowered due to poor skill choice (though worked fine in the lower level) and therefore my time investment in that character a total waste.

    1. Yeah, or finding out that your choices, due to a patch or a deep programming bug, were a total waste of time. I had a great Whirlwind barbarian until they nerfed WW over and over into uselessness, and I had to toss the character.

      I figure that kind of thing is what they wanted to get rid of - the "I put all this work into my guy and now he's useless" problem.

    2. @JP - I think that's why D2 added in the synergy bonuses later on (i.e., at least those earlier skills made the new ones more powerful). I do like that in TL2 I still use those lower level skills all the way into high level play.

      @Peter - Totally agree and I think there could be a whole diatribe on blizzard's approach to change management. Most games I've played wouldn't completely overhaul major mechanics every other patch; it really is irritating but Blizzard does it with WoW as well. And in D3 it fuels speculation as to their true intent, when the RMAH sales peak after such changes :)

  2. What other options are there for level design that you've considered.

    I like the examples and your analysis.

    Seems that changing your mind isn't the best way to go in Pheoniz and wouldn''t fit the game well.

    I always find it hard to select skills as I can't see how they will work till I try them (and then it's normally too late. . .)

    1. Well, I wouldn't say I am considering either of these directly for my level design in Phoenix - rather, I like to look at games to see what kind of impact various design decisions had on the playerbase. My current ideas in Phoenix are a bit evolutionary based on earlier drafts, keeping to the core design goals but exploring via "mutation" some alternative approaches. For instance, I like the combination aspects of the most recent draft, so I'm mutating on different ways to get combinations without losing the basic class flavor that I really enjoy.

      With regards to retraining, that is an interesting topic of its own. I've never seen a game system succeed at "explaining" retraining from a story perspective, so they always come across as gamey (like 4e's change 1 thing per level approach which does little other than give min/maxers more possibilities). My thought is, if a player isn't enjoying some choice they made, just let them change it - q.v., Rule 0!


Related Posts Plugin for WordPress, Blogger...