Log in

No account? Create an account
04 November 2013 @ 06:14 pm
Percentile Systems and Too Much Information  
I've written before about how I don't like percentile systems, and while that's not really that true any longer, I think I've realized at least part of where my dislike came from. It has a lot to do with the way the information on the character's skills is presented to the player.

The typical benefit for percentile systems that I've seen when people ask online is that they're transparent and easy to explain and understand. "Shoot Dude 47%" is intuitively obvious in a way that "Run Away •••" or "Gibber Insanely +12"[1] isn't, and it's easy and fast to intepret the results of the dice, too. Rolling two d10s and comparing to the skill is much faster than, to pick a kind of system I use pretty often, rolling from 1 to 15 d10s, looking for 8s and 9s, counting those, looking for 10s and counting those and adding that to the 8s and 9s, then rerolling all the 10s and starting from step 2. And all that is true and reasonable.

The unreasonable bit on my part is why D&D's dice system doesn't bother me even though it works much more similarly to percentile systems than to the dice pool systems that I usually prefer to play with. Each has a randomizer that produces a linear result, each increments up an equal amount when the character improves, and games like Rolemaster even have the same roll-high resolution system that D&D does. This was really driven home to me lately while I was working on converting Only War's system of Aptitudes to Warhammer Fantasy Roleplay and thinking about the odds. Characters in WFRP tend to start with a ~20-50% chance to accomplish actions, depending on their stats and what skills they have. And if you just look at the page, that seems really low. I've seen research--though don't quote me on this because I can't find the article anymore--that the average person feels competent at a task when they succeed at it at least 80% of the time, so obviously seeing the numbers on the page can be demoralizing, and that's demonstrated by doing even some cursory Googling of WFRP+whiff factor.

The thing is, that never bothered me in D&D, even though the underlying math is similar. A starting AD&D fighter with no other bonuses attacking an orc in chainmail has a 30% chance to hit--15+ on a d20. Using ability checks as a substitute for an actual skill system, a guy with average Dexerity has a ~55% chance to jump across a chasm. The complaint can come out through experience--the last AD&D game I ran, I converted the thief skills into roll-high on a d20, and the thief got really frustrated with constantly failing at his thiefitude[2]--but I don't see the same kind of initial impression of incompetence that a WFRP or a BRP character can generate.

Typically, percentile systems I've seen handle this by changing the base for a roll. If a character has climb 45%, then that's not their chance to climb a brick wall under a sunny cloudless sky, it's their chance to climb it in the rain. A sunny cloudless sky adds a 20% bonus, and maybe gale-force winds would subtract 20%. But again turning to Google, you find that in practice a lot of people don't do that, and so complaints about character incompetence mounts. In Dark Heresy, they even had to expand the range of possible bonuses up to 60% in the errata and note that the general difficulty is actually Challenging, so Average tasks should be rolled with a 10% bonus. Obviously, that's not a super-elegant way of dealing with the problem.

I personally fixed this in WFRP by letting people pick the 10s and 1s on percentile dice after they roll, which turns the normal linear distribution of percentile dice into my beloved bell curve and stops the whiff. That does have it's own problems, though--the most obvious being that it's not a "percentile system" anymore because a 50% on the sheet is actually more like a ~75% chance to succeed. There's also a weird stepped progression system: for example, going from 50 to 54 doesn't improve your odds at all, because you could already switch the dice (50->05, 51->15, etc.) so all those numbers are already successes for you. One could make the argument that this is realistic, since people tend to hit certain skill levels, plateau there for a while, and then jump up to a new one instead of improving linearly once they've attained the basics[3], but invoking realism in an RPG discussion isn't just opening a can of worms, it's ripping the roof off the worm factory.

Anyway, realizing that the problem was one of presentation for me was a big part of why I stopped hating percentile systems and stopped trying to hack WFRP into some kind of hybrid monstrosity. And also why I sunk a bunch of money into Runequest 6. I still prefer dice pools, but I'm much more in the camp of "different tools for different game types" now.

[1]: One thing I really do like about Call of Cthulhu is that it uses skills like this to provide character information. For example, you'll have a cultist stat block that says something like: "Skills: Pistol 39%, Fist 60%, Dodge 65%, Laugh Maniacally 50%, Persuade 25%, Credit Rating 20%, Function in Normal Life 05%." Obviously some of those will never be rolled and probably don't exist anywhere else in the game, but it's a great way of telling you information about that NPC.
[2]: And now I wonder if this would have happened if I had left them as percentages. Does 20% chance of Hide in Shadows cause more negative reactions than Hide in Shadows 17+? It does for me even though I know they're the same if I take a moment to think about it, but for other people?
[3]: As can easily be demonstrated with my own Japanese ability. Grr. (>_<)