I think this is why KaVir linked and alluded to the other countless discussions over this very subject. It's always the same "well, would you consider this classless?" questions recycled ad infinitum.

IMO, it's easier when you agree that things leaning towards the classless side of the spectrum are "classless", including those grey areas, and things leaning towards the most strong prototypical linear character development are "class based." It's not really interesting or useful to have endless debate about if something is slightly more class based or not if we all know it's generally an open ended character development system.

I do, but I also understand the limitations, having designed and implemented several systems using both approaches in the past. I'm sorry if that's not the answer you wanted to hear, but there's no reason why you can't give it a go anyway. I did find the non-exclusionary approach okay, just harder to balance and more restrictive, with less variation between characters.


I already did, in post #4.




I think you've missed the point; the whole concept of races is exclusionary.



I agree, however there needs to be a clear definition of the system before you can really discuss it.

I would agree, and I'm really not interested in classifications anyway. What I'm really interested in is quite a simple concept to explain.

Typical "class-based" systems are usually balanced by asking a simple question: given classes A, B, and C, will a given player be just as powerful if they choose A as they will be if they choose B? And so on. Because I know KaVir would've chimed in on this one :wink:, assume for the moment that a class has access to all of its possible skills.

With "classless" systems, you can use the same method, but instead of having a small, finite set of classes, you now have a -huge- set of possible "skill combinations" (read: classes). Now the question becomes: is A+B+C+D+E just as powerful as H+I+J+K+L? Well, it's highly impractical to look at every possible combination and compare them to all of the other combinations as you would do in a class-based system, since if you have even a moderate amount of skills in your game, the number of possible combinations becomes too large.

So, what I'd like to know is how are systems like this balanced? Do you select similar skills (say, weapon types) and balance them against each other, and balance in a divide-and-conquer sort of way, or is there some other, better method that you've found to doing this?

I think the answer might be 'trial and error', but :shrug: what would I know?

This type of balance is not achieved easily. It often boils down to something far more complicated than picking "+5% damage or +10% damage". Yes, min maxing to some extent will let you figure out how to best do one thing with very specific rules. Where it gets really difficult to balance is when you have multiple disciplines in single builds. Like giving access to healing and damage. It's hard to balance +5% healing vs +5% damage. It's hard to balance things like +5% damage or a new skill useful in very niche situations. The best way to balance in these situations is a lot of testing. Having dealt with this before I can tell you for certain that what looks good on paper is going to need to be tweaked. I guarantee it.

There's always going to be some degree of trial and error, but as I've mentioned before, "in my experience it's much easier to balance the combinations if you first balance the basic building blocks".

In a certain sense isn't balancing an unrestricted system a dead end by the very nature of the system? Other than quantifying certain important values (max damage-per-second, max heal-per-second, max soak, etcetera), which is trivial in a programmatic sense, what exactly are you trying to balance in the first place?

Usefulness. Speaking personally, my goal is to make every ability a useful addition to at least one viable and competitive character build.

It's almost like you'd have to identify what players do in your mud (PK, running dungeons, figuring out puzzles), write test cases for each one, then let the game run all build combinations through each test case, gathering statistics as you go.

I'm scared of Deimos' avatar, it hides in my closet and speaks to me when I am alone. I think it knows
I am afraid…

Wouldn't a balanced system render the players' choices meaningless? "-)

Bravo! You need to go work for Blizzard and explain that to them. :)

Everyone seems to forget the actual multiplayer part of an MMO. You aren't SUPPOSED to be able to solo everything in the game, nor are you supposed to be able to solve every puzzle, win every battle, or see all the content BY YOURSELF. Even on a PvP game, balancing everything for one on one combat turns it back into a solo game again. Pointless.

Balance around group dynamics. Design your game so groups that lack a particular archetype will have a much harder time doing SOME content. It's OK for a fire mage to be able to solo all the snow monsters, but they should get their butts handed to them if they try to go up against the fire elementals without some backup.

Now, that' s not to say there's no balancing to be done, but it's at the skill level. A skill that a fighter acquires at a certain level of experience should be roughly the same power as a spell a cleric gets, although their application may be entirely different. The fighter skill may do massive knockback and stun an opponant for long enough to land other shots. The equivalent healing spell might cure all poisons/diseases, return a large chunk of health, and place a heal-over-time to help mitigate the next few blows. Are they balanced for 1 vs. 1 PvP? Of course not. Are they balanced for group fights, or for adventuring with a group? Maybe.

A classless system, where everyone gets to (try to) fill every role means you have to somehow make your abilities scale so that the specialist is rewarded more than the generalist. If nobody has to choose, everyone will take everything and be Generic Hero 372.

The problem with optimising for group play is that the number of players found on a typical MUD at any one moment isn't that high; it's quite likely that there simply won't be anybody logged into the MUD of a remotely similar level when you go on; so if it's necessary to be part of a group to function, you may never get anywhere. I think you need to assume that a lot, perhaps even the majority, of a character's time will be spent operating solo, and design accordingly.

I agree that it's good try and have skills that are very different in nature but of equal utility (indeed I said as much back in post #18); but you have to be very careful about how you do this, and I'm not sure that your warrior/cleric example is a good one. If a warrior's skill has high combat potential and a cleric's skill doesn't, then what that tends to lead to in practice is a MUD where everyone is a warrior. Why? Because warriors' skills benefit them just as much when soloing, but clerics' skills are only of benefit when they are part of a group.

So if you're going to do this you need to think of ways in which skills can be of use when the player is exploring solo; and if you want non-combat skills to be as viable as combat ones, the whole MUD has to be designed in such a way that situations where the only way to achieve a goal is by fighting are no more common than situations where the only way to achieve a goal is by notfighting. Very few MUDs are designed that way: combat is usually central to the whole process of acquiring anything and advancing the character. So long as that's the case, solo combat potential actually is the only useful measure for how useful a skill is.

And that's going to be even more true if there is any significant PvP element.

@quixadhal:
Inter-player balance and intra-player balance are both very important. Ignoring either one will cause you problems. Also, making the "specialist" more powerful than the "generalist" is a very poor design choice. It leads to everyone becoming specialists, and that's boring. I think you are slightly confused about typical classless systems. They allow a player to be anything, not everything. There is always a strategy for limiting the number of things a player can choose to be good at. If there weren't, what you've effectively created is a class-based system with only 1 giant class. And that would suck :tongue:

I agree with your point, although I would take it a step further: Building up a playerbase is hard work, but a good place to start is by appealing to the type of player who's happy playing on their own - because these are the people who won't quit just because nobody else is online. Get a few of those hanging around, and other players are more likely to stay. But if you want to appeal to those first few players, you really need to give them a fun solo experience.


Agreed. This is the exact approach I took.

Regarding class vs classless, the only form of classless online play I know of is first person shooter instagib. Introduce weapons and you'll have people who specialize in rocket launchers or flag cannons, the moment you select a certain weapon you become in essence someone of a particular class.

It's possible to have balanced classes with different playing styles that appeal to various players. Explorers will appreciate other skills than player killers. Then there are the standard cooperative gameplay classes, tanks, repairmen, and cannons.

I doubt a truly classless game would be very interesting. So I think the trick is to find and fine tune all desired classes, and determine how easy it is to switch between classes. I think what's most important is to give each ability group its own advantage, though this typically only benefits the smarter players, which isn't necessarily a bad thing as it might inspire many to achieve greatness beyond their reach, which like the carrot and the horse is a good way to keep people hooked.

I disagree. I think as long as you structure your game in such a way that explorer skills are just as "useful" as player-killer skills, then you'll be just fine. For example, maybe without some kind of explorer transportation skill (horseback riding, dragon riding, etc.), a player killer is fairly immobile, and can't get to the victims he intends to kill fast enough. Or maybe an explorer can't get into a cave to explore it because it's guarded by gryphons that he would have to kill or otherwise dispatch using non-explorer-skills.

If there's only one task, then people will be choosing skills only useful for that task. Even if other skills are "useful" for other tasks, if the players aren't interested in those tasks, nobody will bother.

If I'm understanding you correctly, I don't see how that's undesirable.

The point however is that this would still result in a class based MUD, just that there are too many different classes to bother giving them a name, though one could try.

The point I argue is that the fluidity of changing classes is what determines the classless nature of a game, as adding classes is close to impossible to avoid. In Unreal Tournament changing classes is changing weapons, and takes one second. On DikuMUD you have to logout, log in a different classed character, and head to the location where your other character quit, and depending on the complexity of the game you might have to perform other tasks to fully complete the class change, not to mention it's not all that much of an immersive experience.

Then there is the issue that fast class changes might not be desirable as it could be equivalent to giving players all available skills, so everyone is the same, which has a big impact on game play. So for a decent classless MUD you might need to add a form of transition, like having to ride a horse for 10 minutes to become proficient at riding, and for an hour to become an expert, with your familiarity with the skill waning as you don't ride a horse for a while and increase other skills instead. Alternatively you can make the interface complex and interactive enough to require some form of player skill, but that'll easily turn your game into a botting competition.