I am working in the context of an action RPG similar to Diablo II which will have a tree of skill which a character can learn. Action RPGs frequently appeal to the min-max crowd of people who want to get the absolute best character build. My problem is that skill points spent at low levels tend to do almost nothing by the end of the game. To a person entirely focused on getting the best character possible, this encourages them to spend the absolute minimum points on their skills at low levels, saving up for higher levels. I don't like that. Players should not want to punish themselves at the early game to get a slight benefit hours and hours later on. I don't find that fun. So here's my idea: Since I also like open-ended character development without restrictive classes, I thought that having one giant skill tree for all characters would be best. The skill would start out with two basic skills, and each layer down, it would double the number of skills (each skill would have two new ones that require it to learn, until the bottom layer). Each skill's damage or effectiveness would be directly related to the skill just before it. For example, one section of the tree could be like this: Magic Missile -> Fireball -> Meteor Storm Magic Missile's damage could, say, increase by three for every point that you spend, starting at five. Fireball's damage would be entirely based upon Magic Missile's damage: its damage could equal 5 times the skill points in Fireball plus the damage of two magic missiles. Meteor Storm, furthermore, would depend upon Fireball for its damage: its damage would equal one half a Fireball's damage every X seconds (where X is calculated from the skill points in Meteor Storm). This would, in effect, encourage spreading out of skill points. Early skills are valuable because they make the next tier of skills more powerful, which makes the next ones more powerful, etc. etc. To max-min, the player would put roughly equal amounts of points into every skill. However, trying to actually make this sounds very difficult to me. The balancing complications of this system seem almost impossible to overcome. Furthermore, I'd like to be able to get the game "working" with just a subsection of the content. In a normal RPG, that would mean fewer classes. But with one large tree, I'd need to implement it ALL or else some skills will be less worthwhile than others, due to this system (skills which don't have two dependent skills will not benefit as many skills, and thus be less worthwhile). So, is this worth it? Should I just not bother and go Diablo II's route by giving small bonuses to later skills based on early skill points (like, say, +1% damage in Meteor Storm for each point in Magic Missile). Any other methods? Should I just ignore the max-miners and assume they'll have fun breaking the system no matter what I do? :)

I rather liked the Witcher's method, where you got lots of points you could spend on lower level improvements, some points to spend on mid-level ones, and the high level improvements were much rarer. Possibly something like that?

Alternatively, more advanced skills could be no better overall, just more specialised, or some other form of equal-but-different method like that.

respec?

why not let the players remove old skillpoints and use them in new skills?

immersion you say?
why not let the players go to a hypnosist or letting the player wait a while for its character to "forget" a skill

Quote:

Original post by Ezbez Magic Missile's damage could, say, increase by three for every point that you spend, starting at five. Fireball's damage would be entirely based upon Magic Missile's damage: its damage could equal 5 times the skill points in Fireball plus the damage of two magic missiles. Meteor Storm, furthermore, would depend upon Fireball for its damage: its damage would equal one half a Fireball's damage every X seconds (where X is calculated from the skill points in Meteor Storm).

But this would still encourage spending of points in certain skills. We could examine the effects of spending skills. Let x,y and z be the amount of skill points spent on magic missile, fireball and meteor storm, and let M, F and S be their damages. Then:
M = 5+3x
F = 2M + 5y = 10+6x+5y
S = F/2s = (5+3x+2.5y)/s
Thus we see that spending a skill point on a magic missile will always have a greater effect than spending one on a fireball, even on fireballs themselves. I'd consider it impossible to balance such a system successfully. If the lower-level skills makes a greater contribution, then people would only make higher level skills accessible and then focus on the lowest level. If all contribution was equal, then people would still focus on the lower levels since they contribute to more levels. If the higher-level skills were somewhat better, then people would still save skill points for them. Thus, no matter what, people wanting to max-out their character will focus skill points on specific skills.

Quote:

So, is this worth it? Should I just not bother and go Diablo II's route by giving small bonuses to later skills based on early skill points (like, say, +1% damage in Meteor Storm for each point in Magic Missile).

Aren't you already going down Diablo II's route just giving larger bonuses? I think larger bonuses could make choices less fatal for the more casual players, but people ready to analyze the system will find a way to maximize their stats.

Quote:

Any other methods? Should I just ignore the max-miners and assume they'll have fun breaking the system no matter what I do? :)

Don't ignore them, but don't try to work against them either. Let them "max-mine", but let it be fun. Find a way for them to be able to play (restating perhaps?)

Alternatively you could simply force players to use points in different parts of the skill tree. For instance if you have an 8-level skill tree and you gain a level, then you could simply get 1 point for use in every level. As players advance you could change the distribution to let them get more high-level skills. Or you could use unlockable skill levels which are unlocked every 5 levels, but do it all the way to the last level, not just for the first 30%. This would of course require that skill points are spent in the skill levels that were available when you gained them.

Go for breadth not depth.

If you have a broad skill tree, one where the is not many dependencies (if any), then they don't waste skill points in the prerequisites.

In a deep tree (like diablo), there are many dependencies, so to get the good skills, you have to navigate the tree spending points in the dependencies to get the skill you want.

The problem with a broad tree is that it can overload the player with too many choices. The solution is to not have too many skills and to group them into groups of around 5 skills. These groups are really just a cosmetic thing, but it helps the player understand what the choices are and not overwhelm them.

I think there are a few simple options you're leaving out:

1) Don't let the player save skill points. You have to use them all when you level up.

2) Use talent trees, like WoW, where you have to progress deep into a tree to get the best stuff anyway, so you're not "wasting" points by taking the early tree slots.


I can't remember exactly how Diablo 2 did it, so maybe I'm not grasping the problem.

I found WoW did a really good job with this. The lower-tier talents have more broad, general effects, and the higher-tier talents are more powerful, but only affect one or two abilities.

If you are worried about low-leveled (or early-gained) spells becoming obsolete, you could simply just make it so that your Spells' power depends on the character's magic proficiency, so that both lower-leveled and higher-leveled spells will increase while you level up your magic proficiency; as well as that, make it so that every spell has a purpose.

For example, in your provided example of Magic Missile, Fireball, and Meteor Storm, each could be designed more so that each would be optimally used in certain conditions over others. Magic Missiles for example, would work best in close-combat situations as the other spells' splash damage might affect you or your allies. It could also probably home into its target, making it a sniper weapon. Fireball would then work best against fast-moving large masses of weaker-HP enemies. While Meteor Storm would work best against slow-moving masses of larger-HP enemies.

The key in the wow / diablo with their talent / skill trees is the way it affects killing things. In diablo, you have click-2-attack, and whatever skills you have points in.(save passive skills)

In wow, you gain skills as you level up, and talents augment those skills, without being skills themselves.(for the most part, each tree has new skills in a few places)

Meaning, in diablo, when you get the new spell, you use it, because its either that or the old spell. In wow, you still have your set of skills, you favor certain ones because they have bonuses, but you still use other skills, each skill still has its place.

If you still want to stick with diablo, with each thing in the tree being a new skill (mainly), you just need to make sure that new skills dont kill off an old skill entirely, try to leave some purpose intact.

[edit:] Tangireon beat me to it :(

I have to agree with an earlier post. Give people free or low cost respecs.
Allow people to REALLY mess with specs without suffering for it. Let a person get low level skills and then as he progresses be able to switch it up without having to spend 2139283018 gold or monies to do it.