Torchmodders

Modding => Classes & Skills => Topic started by: gytfunke on September 05, 2014, 07:46:47 pm


Title: The Case for a Single-Level Skill Model (or, The Short Forest of Skills)
Post by: gytfunke on September 05, 2014, 07:46:47 pm
Okay, so I mentioned this before in that other thread where we all aired our ideas and hopes for where we would take TL2 modding.

I think we should switch the way we do things when making classes and abilities for our TL2 mods, especially going forward with total conversions.

Currently, each active and passive skill has 15 levels.  I feel there are a few serious negatives to this approach.

Cons

-Longer development time
-More complex to balance
-Skills that open at level 1 and skills that open at level 35 have to scale at different rates, making balancing even more of a headache

Pros

+Incremental progression, which is satisfying
+Vanilla TL2 compatibility


What I propose is a '1 skill, 1 level' model.  You invest 1 point and you get the skill.

Pros
+Less dev time
+Easier to balance than the 15-point model
+This all adds up to more time spent developing OTHER stuff

Cons
-Vanilla Incompatibility
-What can I spend my points on every level!?!?!?


So, why do I care so much about this?  I think the bottom line is that there are a very finite number of meaningfully different skills available in a combat-based game.  New class creation doesn't add a lot of meaningful content to the game.  It's still going to play mostly the same way because you can't escape the fact that the 4 vanilla classes have almost every conceivable skill type possible.  You can change fireball to iceball, but when you put them in the game they play the same way.  Skill creation is very limited.

Since it adds so little value to our mods, new skills should be deprioritized in future mod work.  Removing the redundant levels each player skill has would reduce the amount of time we have to spend creating and balancing them.

Unfortunately, this would decrease the pleasure players get with each levelup.  I don't have an alternative progression path in mind that would replace skill investment the way it works now.  I do, however, have another proposal.

Reduce the level range to 1-20 for future standalone mods.  I'll discuss my reasons why in a new post.
Title: Re: [Discussion] The Case for a Single-Level Skill Model
Post by: lolesch on October 10, 2015, 03:15:38 pm
Hey Gytfunke, do you have an example for something like that? How do the skills "level up"? using a graph? For Affixes with Duration we still need the level model, no?
My passive skilltree could use some level1 skills but i still want them to be increasable...
Title: Re: [Discussion] The Case for a Single-Level Skill Model
Post by: Phanjam on October 11, 2015, 05:50:59 am
I think gyt was outlining a base philosophy for mod design / mod making, and I think (gyt, pls correct me if I'm off here) gyt's Lure of Ember and Classless Character mods were the brain-children of this philosophy.
Title: Re: [Discussion] The Case for a Single-Level Skill Model
Post by: gytfunke on October 11, 2015, 11:01:06 am
I think gyt was outlining a base philosophy for mod design / mod making, and I think (gyt, pls correct me if I'm off here) gyt's Lure of Ember and Classless Character mods were the brain-children of this philosophy.

Phanjam knows me well.

@lolesch The skills don't level up.  The idea was that character progression would be more lateral (the character's skills would grow in variety and options) rather than vertical (the skills grow in power).  The example is in the Lure of Ember mod.  There's a link to that abandoned mod somewhere round here.
Title: Re: [Discussion] The Case for a Single-Level Skill Model
Post by: lolesch on October 11, 2015, 11:13:26 am
well isnt that a good point to start again?
Lets say we can choose our "normal" level 1 skills  (maybe 21 like in vanilla for variety) and the rest we can spent into "passive" skills modifying our characters. Gain bonuses, choose elements, transfer Mana int health, gain auras, all the funky stuff...

Hm.. well that would mean creating 131 skills  ??? 21 class based and 110 class open passives which could stack or remove eachother again 8) ooh thats gonna be fun to develope!
Title: Re: [Discussion] The Case for a Single-Level Skill Model
Post by: Chthon on October 11, 2015, 12:39:25 pm
I strongly suggest taking a look at the GW1 skill and stat system for inspiration.

-------------
[edit: Additional thought. Since several folks seem to be interested in a "skills are items" system like PoE, I did a bit of experimenting.

The bad news is that the "spell" scroll/tome system is hopelessly hardcoded. The 4-spell limit is unavoidable and no part of the system can really be repurposed for anything else. So, you've pretty much got to do everything some other way.

The good news is that there is another way. You can do skill learning and unlearning via consumable items that uses the "set skill level" effect, so long as the PC unit has that skill. If you shadowed the skill levels with custom stats, you could do things like completely hide skills the player doesn't know and unlearning from a menu. For double the custom stat overhead you could even do temporary unlearning -- which would make de facto class changes possible.

Title: Re: [Discussion] The Case for a Single-Level Skill Model
Post by: lolesch on October 11, 2015, 12:54:35 pm
Thank you for the input! It already made me rethink about the single level system  ;)

We could add affixex with statwhatchers to controll lategame scale based on your statpoints spent. So a physical skill will benefit from weaponDmg so you want to max out Strength. but you might get a bonus on that skill if you invest into focus too
Title: Re: [Discussion] The Case for a Single-Level Skill Model
Post by: lolesch on October 14, 2015, 03:47:36 pm
edit: Additional thought. Since several folks seem to be interested in a "skills are items" system like PoE, I did a bit of experimenting.

The bad news is that the "spell" scroll/tome system is hopelessly hardcoded. The 4-spell limit is unavoidable and no part of the system can really be repurposed for anything else. So, you've pretty much got to do everything some other way.
Wait what? the 4 slots are preset in the "containers" and can easily be changed! Or is there something else you cannot edit?
though I dont like the Idea of random skill drops ;D

Quote
The good news is that there is another way. You can do skill learning and unlearning via consumable items that uses the "set skill level" effect, so long as the PC unit has that skill. If you shadowed the skill levels with custom stats, you could do things like completely hide skills the player doesn't know and unlearning from a menu. For double the custom stat overhead you could even do temporary unlearning -- which would make de facto class changes possible.

Actually I'm thinking about this: Single-Level Skill => Event => Affixes with different level value and stat whatcher.
So in longer text: the skills are linked to the player.dat with no level req. Unlock the skills is stat based and by adding a stat a higher affix inside the skill gets enabled. these stats can be applied in any way, by reaching a certain attribute value or by clicking a button is a skilltree...
Also: did anyone tried the "Skill_Enable/disable" effect?
Title: Re: [Discussion] The Case for a Single-Level Skill Model
Post by: Phanjam on November 01, 2015, 04:50:22 am
@lolesch The skills don't level up.  The idea was that character progression would be more lateral (the character's skills would grow in variety and options) rather than vertical (the skills grow in power).  The example is in the Lure of Ember mod.  There's a link to that abandoned mod somewhere round here.

Yup that link is over here (http://torchmodders.com/forums/lure-of-ember-tc-public-board/lure-of-ember-mod-files/msg1734/#msg1734) ;)

@lolesch sorry but all I can contribute right now is to cheer you on. This is one of those mod concepts which I think is important at a core level; it would be so great if it moved forward some more!