Designing Roleplaying Systems

You're barking up the right trees.

Designing the back end roleplaying system for a game is an interesting task.

Firstly, you need a means of defining what makes a character. These form the base stats - the inherent ability of the character for being what they are.

How many of these you have is up to you, but should encompass anything that might inherently have an effect in game.

Next, you need a means of defining limits or secondary stats based on those stats (and therefore some kind of sensible stat range). For example, in the WSRules set (part of the licensable Primogen suite) we have stats that typically range from 1 to 10, with an 'average' human rating 5 in each.
A human has a hit-table of locations, where END (endurance) determines hitpoints at each location, by simple multipliers. The carry weight is a factor of strength (for power) and endurance (for stamina).

Now, you have the choice of how to model skills. Many games are class / profession based, where the 'job' of a character is directly responsible for their abilities. Some are 'truly' skill-based - look at Runequest, and GURPS.

Next, item stats like damage and so on. These should be balanced around the primary and secondary stats. Always bear in mind 'average human' for rough balances - for example an average human might get stabbed with a dagger twice (perhaps once if hit critically) before dying, but go down in one slash from a sword.

The system we use incorporates 'primary' stats into the skill equation (and effects them in terms of levelling up) which creates a nice dynamic in terms of 'related' skills and 'being a natural', where a character's primary stats can really lend it some extra weight in certain skills.

Finally, game flow. For a tabletop RPG you don't want too much dice rolling, or looking up of tables. For a CRPG tables aren't a problem (but dice rolling can be! Random generators are often slow!) so you can really go to town compared to a tabletop game. Think about whether you're going to use a turn based system or real time - do you need to model 'initiative' or 'action speed' or 'action points'?

There's a brief overview on the Bloodspear website at http://www.bloodspear.co.uk/

But in terms of the design 'path':

1) PRIMARY STATS - inherent abilities
2) SECONDARY STATS - based on primary, no amount of practice changes (eg HP)
3) SKILLS / PROGRESSIVE ABILITIES - class based, 'pure skill' or 'stat-based skill'?
4) ITEMS - PARTICULARLY COMBAT ITEMS - balance with 'average human' in mind.
5) GAME FLOW - combat, initiative etc.

Hope this helps.

Having a thing to plot and manipulate graphs easily [like this applet] is useful. Making a quick mockup of a character generator and/or battle system is useful.

It's hard sometimes to see how things would work conceptually or on paper, and since the figures are likely to be tweaked, having something easy to edit and recalculate is good.

A good place to research for you may be MUDs, a lot of highly original rulesets have been developed for them and you can get a basic MUD to test out your ideas.

I recomend trying to keep it simple.

I took a stab at this years ago for a pen and paper RPG. Character stats were centered around bell curves. Wrote a program to generate large numbers of characters and everything. Basicly tried to find ways to give modern day average people stats and what would stats be for exceptional people. Then went on to try and figure out how many calories a wide range of tasks would be and how many calories the person ate would determine the things they could do in a day. And of course, a system for magic was needed for possible future expansion of the system and so I started that figuring that a magical unit could be figured out somehow by how much energy (in calories of course) was required to generate a certain amount of light.

Lots of scribbling on paper, charts, graphs, and notes. I did try to run a game using this system, but it didn't last long enough to get beyond the story telling between all the players. The whole thing got out of hand real fast and the actual stat system never got play tested.

Was lots of fun trying to figure it all out though.

While simple stat-number systems are everywhere, I start to love more sophisticated stat-number calculations. Intead of burning all your cpu threads to the empty shell with beautiful landscapes, leave some for the complexity of the char stat-numbers.

The weakness of char development in current mmos is that, char development path is quite predictable, so all you need to do is max it out. And you can do nothing about your char, so kill some mobs to level up is the only way you can do about the development of your toon.

So I say again, thou I am playing a less popular game, the mmo model behind it is far more advanced than those in the main stream mmos.

Take my toon for an example, it started with the stats supposed be an excellent crafter but a lame tank. During lvl1 - lvl30, it's not that obvious. From lvl 35-41, I painfully feel its lameness so I have to try my best (within the means provided by the game) to *improve* the lame char, to an awesome tank (at lvl48 now).

And not just that, my awesomeness is abit different from the awesomeness of those toons supposed to be awesome tanks (yep, very difficult to explain with my limited liguisitc ability, because everything in the game seems to be co-related).

The balance is that the supposed to be awesome tanks but lame crafters will have to paid their efforts to tune their toons to an awesome crafter.

How many level-based game out there allow you to fine tune your chars? vitually 0, simply because all they adapt is a simple char system with simple stat-numbers. Simplicity sometimes is good, however, a market with only simplicity but without complexity is definitely not healthy.

I'm interested in this topic as well, for we are beginning to approach the point where we need to figure out this sort of information in our battle system.



Our game is pretty unique in that both characters and enemies can target enemies in multiple places. For example, characters have targets of: head, torso, arms, and legs, each with their own set of stats for damage calculation. For example, getting in a head shot ([wink]) causes the most damage, but it has a higher probability of missing. Torso is easy to hit, but low damage. Arms and legs are both hard to hit, and additionally drop a character's strength or speed. (By the way, I call this system MAPS, for multiple attack point system). This allows the player more strategic options in battling their opponents and will hopefully make the battles more fun than that of typical RPGs (which I usually find trite, simplistic, and boring). However although it sounds like a good idea (at least I think it is), it is meaningless if we can't properly balance the algorithms and weights correctly.


Like a previous post mentioned, we also utilize Gaussian random variables in many parts of our system. The idea here is that when I use character A to attack enemy X, I don't want to cause 56 damage every damn time. It's predictable, and IMO predicability is boring. Instead we give a mean of 56 to a Gaussian distribution and using a defined variance, select a number that will usually be around the vicinity of 56, but sometimes may be very weak or very strong. This is also used in other places like actually selecting the strength of the enemy party in battles.


Anyway, like I said we haven't really gotten this far into things yet (focus has been on the engine rather than the game in the past). What I was planning to do for implementing this system was less pencil and paper calculations and more real-time playing and balancing. You could scribble away on paper all day for a week and still not know if you've come up with a reasonable and balanced system until you play it. That's my opinion on the subject anyway. [smile]

I've always wanted to experiment with an RPG system where the building of certain stats will effect the building of others, potentially even lowering them. If you have a character with uber-high strength, it will be more difficult to raise your agility/speed, simulating the general idea that worlds strongest man may not even be able to scratch his own ass from being so big. Seems like if designed right it could be an interesting dynamic stat system as opposed to ever increasing numbers that in the end don't really mean much.

Designing a "from-the-ground-up game system" for an RPG :

- Start with an empty character class.
- List the actions you want the characters to be capable of.
- For each action, add the necessary data to the character class.

Once you know what you want the game to play like, designing statistical functions to match is relatively simple.

The other design path is to create the character class from the start, and define your actions from there.

Quote:

Original post by Roots Like a previous post mentioned, we also utilize Gaussian random variables in many parts of our system. The idea here is that when I use character A to attack enemy X, I don't want to cause 56 damage every damn time. It's predictable, and IMO predicability is boring. Instead we give a mean of 56 to a Gaussian distribution and using a defined variance, select a number that will usually be around the vicinity of 56, but sometimes may be very weak or very strong.

Roll 6 sided dice, on 6 roll again and add all results.