Thursday, July 1, 2010

Game Mechanics

Okay. Shattered Worlds will be a skill-based system — because I farkin’ hate the rigidity of character classes. Also there will be no nonsense about ‘alignments’. Characters will do as they see fit, according to their nature, whether they be players or the lowliest spear-carrying non-player. Good old D’n’D achieved its position by being the first modern RPG, not the best.

In a skill-based system, actions taken by characters are assessed for success or failure by a dice roll against their ability in a relevant skill. You want to shoot an arrow accurately? You need archery skill. Want to pick a lock? You’d better have some training there, yeah?

When it comes to using the skills, the GM assigns the task facing the character a level of difficulty. The character is then required to make a dice roll which matches the level of difficulty assigned to the task.

This process — this dice-rolling mechanism — is central to the game. You need to have it pretty thoroughly nailed down before you can go doing much of anything else by way of game design. In general, there’s two ways of handling it. The first is pretty simple: you go with a sort of linear scale based around something like a d20 or a d100 roll, and base the success roll against the character’s skill.

This is the road taken by a lot of well-known and popular RPGs. It’s the DnD method, for instance — insofar as DnD actually uses a skill system, that is. Fantasy Games Unlimited used a d20-based system to good effect: your character would develop levels in whichever skill, and when you had to test it, you might score a few modifiers on the roll depending on circumstances and how the Game Master felt, but ultimately, you had a Basic Chance of Success which was somewhere from 1 to 20.

For a long time, I worked with a system based on d100, but with a few twists. Players never knew exactly what their skill level was. I used to keep track of their skills myself. All they knew was a vague verbal description - anywhere from ‘abysmal’ to ‘master’, with ‘average’ falling somewhere in the middle.

The reason I chose such an obfuscatory tactic was, very simply, that the linear-scale systems are quickly and easily ‘cracked’ by savvy gamers. You know your character’s abilities down to the last decimal place; you throw the dice, and you know instantly the outcome.

I wanted more room to move. ‘Fudge factor’, if you will. I wanted to be able to keep characters alive if it suited the story, and I wanted to be able to do it without blatant, obvious intervention. I called it the ‘Hidden Hand’ system, and it worked pretty well, on the whole. But there was a lot of book-keeping, wasn’t there? A typical character might have fifteen different skills at various levels of expertise. Six players in a group: that’s ninety separate numbers to keep track of, session after session, game after game. Sure, you can cut corners with computers and good record-keeping, but... ouch, y’know?

D20 systems never really caught my interest at all, by the way. With only twenty possible outcomes... well, if you incorporate the idea of ‘critical’ successes and failures in your game (which is always fun) then you’re looking at a game process wherein one in every ten attempts to do something is either a critical success or a critical failure. So unless you start designing complex ‘critical’ tables (I’m lookin’ at YOU, RuneQuest) or throw in a secondary ‘intensity’ roll, pretty much any gaming session is going to be filled with a lot of seriously wild outcomes.

I’ve driven a car for nearly thirty years. Can't say yet I’ve had a ‘critical failure’. Possibly one occasion I’d call a ‘critical success’. So — one in ten? That’s a fucked-up game mechanism right there, if you're trying to simulate something that 'feels real'. (Or at least has that sense of verisimilitude you get in all the best stories.)

The alternative to the linear-scale systems is a sort of grab-a-handful method. In these systems, the necessary score for success gets higher as the tasks get more difficult. The twist is that as your character gets more skilled, you get to throw more dice. The old ‘Ghostbusters’ game used to work that way, and there are quite a few modern games that work in similar fashion.

I’ve never done much work with it before, but this is the mechanism I’m using for Shattered Worlds. The core die is 1d10, and levels of difficulty are based on the number 6. So: if you’re faced with a Level 1 task, you need to roll a 6 or better to succeed. Since you get to roll 1d10 for every level you have in the appropriate skill, it’s pretty easy to carry out a Level 1 task by the time you have 2 skill levels. In fact, if you do the math, rolling 2d10 (as you would with 2 skill levels) means you have a 90% chance of rolling a total of 6 or more. By the time you’re rolling 3d10 (for 3 skill levels) the probability of success jumps to just under 99%

The really interesting thing about this system is that the nature of the bell-shape curve you get from all that dice-rolling brings about some funky probability shifts as things get more difficult. Take the situation above: level 1 problem, versus level 2 skill — 90% success; level 1 problem, level 3 skill — 98.9% success. Now: let’s consider a level 5 problem, shall we?

Since our success number changes by 6 per level, you need to throw 5 x 6 = 30 to beat a level 5 task. If your skill level is 5, you have a 50% chance of success — exactly the same as a character of skill level 1 encountering a level 1 problem. However, if you have a level 6 skill, you’re throwing 6d10. So... what are the odds of success now?

Since I’ve had a good stiff rum drink while I’ve been typing this, I can’t tell you precisely. I can, however, tell you that it’s distinctly less than 90%. And if your skill is level 7, your chance of success is considerably less than 98.9%

In fact, the more challenging the problem gets, the more the probabilistic ‘playing field’ levels out. So: while a character with skill level 4 should be pretty bloody confident about being able to throw a 6 or better to beat a level 1 task, a character with skill level 8 doesn’t have anything like the same degree of confidence about beating a level 5 task.

I like this. I like it a great deal.

I like it because it solves a major problem for me in terms of magic. As a storyteller, I don’t like supremely powerful magical types because it’s just too easy for them to fuck up your plotlines. How do you pose interesting problems for characters who can solve ‘most any difficulties by dropping an asteroid on them? But: if you grade magic spells by power, and raise the task level appropriate for powerful spells — well, even quite powerful magicians stand a reasonable chance of failure, and failed magic is never, ever a good thing. Thus, by virtue of the central game mechanism, you create wizard types who are reluctant to hurl around seriously overpowering stuff because it’s dangerous, dammit! They could get hurt!

I like it also because it means while there’s a big difference between a highly trained practitioner and a noob, the differences get less important the higher up you get. Two characters battling, one with a skill level of 8 and one of 7 — they’re almost even. On the other hand, between a 2 and a 1 there’s quite a big difference, probabilistically speaking. That’s useful. It gives new characters incentive to stay away from really dangerous types until they’ve got some mojo of their own. It means you can always provide a challenge even for highly skilled characters just by creating a few skilled enemies. And it means that the highly skilled sorts will sneer at base-level problems — but will never have reason to be overconfident of their ability to handle things that are close to their own level.

There remains only the question of critical successes and failures. Personally, I like ‘em. I like the added fillip they hand the storyline. I enjoy the element of improvisation it forces on me as game master, and I really enjoy watching the players when they roll a critical: they tend to get awfully excited, which is fun.

Therefore: to the central mechanism of rolling numbers of d10 equal to your skill level, I’m adding a familiar element: 1d20. I’m calling this the ‘Kaboom! die’. A roll of 1 is bad. A roll of 20 is good.

Obviously, this still looks like a ‘ 1 in 10 is critical’ mechanism. But remember: it’s decoupled from the basic success/failure system. It’s an enhancer, not a dictator. If you don’t like the odds, you could change it to a d100 roll for the Kaboom effects, for example. However, I’m keeping it relatively simple. Roll a success plus a 20 on the Kaboom die: that’s a true critical success, with all the shiny special effects. Roll a failure plus a 20 on the Kaboom die: lucky you! You can gather up all your d10 and roll again, adding the new total to the old. In this way, characters of low skill have a chance of succeeding at really difficult tasks. Roll a success and a 1 on the Kaboom die: minimal success - no special effects or extras of any sort. Roll a failure and a 1 on the Kaboom die: true critical failure, with all the disastrous side effects and vile trappings.

So — with the ‘critical’ outcome decoupled from success or failure, we no longer have that 1-in-10 nonsense. And with the bell-curve system provided by throwing numbers of dice, rather than throwing against a linear structure, we get a subtle bias that evens things out towards the upper levels, making the game more interesting and challenging for high-level types. Of course, it does mean that highly skilled characters can pretty much wade through the lowly types... but really, that’s the kind of effect we’re after, isn’t it?

That’s my take on things. Anybody with a better head for probability: if you can kick holes in what I’m doing here, I’d be grateful.