It’s possible I got this house rule from another game, maybe Lejendary Adventures. It’s used for less-important skills that a PC might pick up if you’re playing 1e or something that lacks a skill system, but isn’t used for tasks like combat or casting. You also wouldn’t use this for trivial things like smoking or narrow tasks like skinning deer.
Or, if you’re writing a game, maybe this is how all your advancement works.
To start with, determine the PC’s skill level on a percentile scale. Maybe equal to their relevant ability score for novel skills or double that for things people in their social class probably have exposure to. That’s their success chance, and they need to roll under that on d%. Add a negative modifier for really difficult conditions, but for easy tasks give double the bonus you think they should get.
If they fail (rolling over the skill percentage), make a note of that. At the end of the game session, award +1% to every skill if it had at least one failure during the session, and if 2+ failures, then there’s an X in 10 chance to get an extra 1%.
This means you can’t improve a skill if you don’t use it, you learn fast at the start but slower if you’re really good at it, you learn more by taking big risks instead of grinding easy tasks. You’re incentivized to do a wide variety of tasks throughout the session.
It can work really well for learning languages.
If you use this as your main advancement mechanic, consider throwing in additional skill point awards which players can choose to apply to any skill that’s 75% or under. Maybe a bonus 1% for finally defeating the bandit lord, 1% for rescuing the brainwashed prince he kidnapped, 1% for returning the entire treasure the bandits stole. A given session might offer 0-2 of these achievement-based bonuses.
I also think if you’re using this as your core advancement mechanic, it could work to help limit the power of casters by saying the caster has to improve each spell independently, just like any warrior would need to learn all their weapons independently. But, to reflect that there is some cross-knowledge where knowing Shortsword makes you a little better at picking up Longsword, you’d need an adjustment in there. Maybe half your chance to hit comes from the specific Fireball skill and half from your Evocation group skill.
Another cool percentile rule I found from Basic Roleplaying (I think) is that on any percentile roll, doubles are a critical, but it’s a critical success if it’s a success, critical failure if it’s a failure. So this naturally produces critical outcomes 10 in 100 which is equivalent to 2 in 20 for natural 1s and natural 20s on a d20 roll. But it also spreads those out in a way that makes high-skill characters more likely to get critical successes and low-skill characters more likely to get critical failures. And that just feels right.
Anyway, I think this is a cute idea, but I’m also just very happy with the comprehensive secondary skill system and using d20 under ability score checks for miscellaneous tasks.
1d30 