(For those that are curious like me, I think Thomas is referring to this video – https://youtu.be/yCNxr1O-4qk?si=DMa_bV1yM43Qy7va&t=369 – skill tree & player class discussion starts at 6:10 and ends around 12:30).
Thomas, first-off, kudos to you & your team for the dream qualities of this game’s design & execution. I’m extremely excited for April 18th!
From the interview, it sounds like a Skill Tree / Class system was rejected for (2) main reasons:
- Increasing player autonomy (i.e., wanting to dial-down how much the designers dictate the boundaries of a given playstyle)
- Reducing homogenous gameplay experiences (i.e., if I go up the Rogue skill tree, my gameplay could look very similar to another player’s Rogue since we’re climbing the same tree)
I feel like the current implementation–a Soulslike Attribute System in which leveling-up allows players to make small, incremental increases to some numbers–doesn’t quite solve the (2) mentioned problems and instead adds a few more to the pile.
If player autonomy is a concern, then there are few points of tension:
- Attribute requirements dictate which weapons I can equip, so my access to certain playstyles is still being prescribed by which numbers I choose to incrementally raise by 1.
- As a player, it feels less autonomous to be a classless entity when the weapon types from loot are random, weapon stats are random, and I can’t predict / plan for what attribute spread I’ll need to have in order to try later weapons. Maybe there’s a really fun Scythe later in the game that’s DEX/INT but I’ve already committed to a STR/FTH spread; or, let’s say I want to test out the Great Sword I get from defeating Warrick (just to see if I even like the moveset / playstyle), but I can’t until I’ve spent 12 points in STR (and not spent those points elsewhere). Both of these scenarios result in less player freedom. This can, at best, be relieved by a low-cost respec system, but then that’s drawing attention to the overall tedium of the system.
For the issue of homogenous gameplay experiences, the root problem seems to be not the tree but the placement of, and balance between, its branches.
- Looking at Ori 1, the (3) Ability Trees are linear and branchless. As I go up the Utility Tree, I’m never prompted with a choice to commit to Branch A at the cost of Branch B/C/D, so of course my gameplay experience is going to look similar to other players; the system did not supply stakes or alternatives.
- The current system seems more likely to resolve into homogenous gameplay experiences since the only thing that separates Dexterity Player A and Dexterity Player B is the difference in their attributes (and the gear that those attributes permit them to equip).
Outside of those (2) issues, I think it’s worth mentioning the difference in the sense of progression & satisfaction between a Skill Tree system and an incremental, numbers-go-up system. With a well-grown skill tree, it feels exciting to earn that next level and have the opportunity to unlock a new skill, to commit to a choice with active gameplay consequences that reinforce the fantasy of playing as an X (e.g., do I choose the lower cost & shorter distance Blink, or the longer distance Blink that empowers the next ranged attack?). I can’t say the same for increasing a few numbers by 1.
All that said though, I’m not a game designer, and I imagine the implementation of a Skill Tree system (especially at this stage of development) would be exceedingly resource-heavy given the design complexity, UX demands, and all the tuning required across multiple other systems. However, if it ever becomes a possibility down the road, Skill Trees would be another one of those dream qualities that would massively improve the sense of player expression and progression.