Someone posted this article a while back that I think is spot on:

http://en.wikipedia.org/wiki/Reinforcement

It basically goes over positive and negative reinforcement, what types/combos work best, and the idea of extinction (separate article, but you can link to it).

To sum it up in game terms, if a player is receiving items he/she finds valuable while playing (a new level, a new spell, a new item, gold, whatever) then the player will continue to play. However, when the value of the reward diminishes too greatly (ie, it takes too long to level, it takes too long to collect enough gold for the next item, it takes too long to complete quest, the value of the item drops starts to fall, or the player suddenly doesn't value any of the items obtained anymore), then the player will stop playing.

Unfortunately, items in-game don't just 'have' value, the player has to be made to believe the in-game items are valuable. So, providing a solution to this problem is where I believe all the magic happens.

The decision isn't as binary as we tend to make it. CLasses, with skill trees are a hybrid of the two, and given enough depth in the skill tree you actually end up with a variety of "class-like" derivatives. An arms warrior and a protection warrior in WoW are different creatures entirely. A drone robo and a crate robo in Global Agenda are different creatures as well.

So have both. I like the idea of a skill system for diversity but the reality is you end up with cookie cutter specs regardless. A class system forces some level of diversity, and then it's up to you to create a compelling skill system that reinforces it.

I think if you want to know which is better in a system, you have to be aware of the goals you're trying to accomplish with that system. What are the variables you're trying to maximize here? Stickiness? Approachability? Flow?

A lot of design is knowing what's actually the most important goal, and which things are fluff that detract from that goal.

Quote:

Original post by Cygnus_X Someone posted this article a while back that I think is spot on: http://en.wikipedia.org/wiki/Reinforcement

Great link, very handy to have on hand :)

Quote:

Unfortunately, items in-game don't just 'have' value, the player has to be made to believe the in-game items are valuable. So, providing a solution to this problem is where I believe all the magic happens.

I dunno man, I think you're overestimating the difficulty here. You put a number on almost anything and people will want to get more of it. I mean, I remember back when they used to have a page showing something like the "top 10 posters" (in terms of post count), and people would game the forums just to have the highest # of posts. Or somewhere like reddit where people are constantly karma whoring. Honestly you can turn almost any activity into a game just by adding some form of "points", even if they aren't intentionally supposed to be points.

Quote:

Original post by Cygnus_X Unfortunately, items in-game don't just 'have' value, the player has to be made to believe the in-game items are valuable. So, providing a solution to this problem is where I believe all the magic happens.

The value of a chess queen is worth a bishop, a pawn and a knight. It is so not by some declaration from on high, but because the rules of the game have granted the queen power approximating the tactical advantage of nine pawns.

In a game universe with some rules framework such as how combat is resolved, the items have intrinsic value because of the rules in which they participate. In some games armor is everything (Dwarf Fortress), but in other games it makes little difference (Age of Conan).

It's not by trickery or mass hypnosis that value is assigned; the value is there to be discovered as the game is played. Recently the top tier units in Starcraft 2 received a nerf because their performance was excessive -- they had too much value compared to the way the developers wanted the game to play out.