What are the ideal % of success for PCs succeeding skills?

[Turanil]

Thinking again about houseruling skills (and using a much shorter skill list), I got to think about the method (how to get them, skill-points or not, what dice to roll, what the difficulty level, etc.) and the percentage chances of success it could give. But now my question is about the % chance at succeeding at a skill check : how much it should be for the game being fun?

In fact, whatever random method you use (1d6 roll in AD&D 1e ; proficiencies checks in 2e ; skills in 3e ; siege checks in C&C ; etc.), in the end it all boils down to a percentage chance of success. For example, rolling 1 or 2 on a d6 wields 33% chances of success; rolling a 18 or better on a d20 wields 15% of success; etc. So, with this in mind, I would like to determine what are the most appropriate chance of succeeding at a task for a PC, for a game being fun (fun is the concern, not realism). Obviously, succeeding at a task only once in ten times isn't fun; but likewise succeeding 90% of the time may become boring (the game loses its fun that stems from uncertainty).

Hence this thread and this poll to get your opinion. I once read that a survey did indicate that players generally prefer that a "trained" character would be able to succeed 70% of the time. I could base myself on this, but I want to hear what you think. That is : what should be the base chance of success at a task for a trained character, and for an untrained character. I mean, for something that requires a roll, thus something deems relatively difficult, not something easy that should not call for a roll.

(Hope the question makes sense? )

 

[Turanil]

For those curious... My own radically compressed skill list is :

-------------------------------------------------------------------------------------------------

Acrobatics: DEX, (Balance, Climb, Tightrope walking, Tumbling, Escape bonds)

Athletics: STR, (Endurance, Jump, Swim, Bend Bars, Break Doors, Move heavy objects, etc.)

Deception: CHA, (Bluff, Disguise, Forgery, Scams, etc.)

Healing: WIS, (First Aid, Treat Injury, Surgery and HP-recovering potential)

Knowledge: INT, (choose one of the five following specialties)
1) Knowledge: Mundane (local + geography + history + heraldry + culture + etiquette, etc.)
2) Knowledge: Arcane (spellcraft, lore about magical things, etc.)
3) Knowledge: Religion (various faiths, planes, etc.)
4) Knowledge: Nature (flora and fauna, survival, tracking, etc.)
5) Knowledge: Military (enemies' usual strategies, armies' organizations, engineering of war-machines and castles, etc.)
6) etc.

Language: INT, (Read and Write own language, + one additional language per Int bonus point.)

Manipulation: DEX, (Open Lock, Sleight of Hand, Pick Pocket, Use Rope, Disable Device, Remove Traps, etc.)

Notice: WIS, (Spot, Listen, Smell, Taste, limited Sense-motive, etc.)

Persuasion: CHA, (Diplomacy, Gather information, Bargaining, Intimidate)

Profession: INT, (Merges together craft, profession and perform. Choose one activity such as Weaponsmithing / Carpenter / Juggler / Musician / Mariner / etc. This skill is the ability to craft / practice / trade / etc. the chosen activity, then to make a living from it, assess if others do it well, etc.)
1) Profession: Merchant (assess mundane items' approximate value and determine if unusual/rare items are junk or high quality, know where and to whom sell the item for best profits, etc.)
2) Profession: Smuggler (streetwise, merchant of illegal things, black market, etc.)
3) Profession: Robber (Innuendo/thief's cant; determine what's best to steal in a hurry; Search, etc.)
4) Profession: Investigator (Search, knowledge of laws, etc.)
5) etc.

Riding: WIS, (Ride + Animal handling of mount species)

Stealth: DEX, (Hide, Move Silently)

-------------------------------------------------------------------------------------------------

There would be two levels of result that can be accomplished with the skills: untrained can only do basic things, trained can do complex things. For example, trained in Notice would help discern that a thief masquerading as a soldier doesn't look like a consummate warrior, and doesn't wear his weapons as a true soldier would. But untrained you would only notice that the guy's chainmail seems a little too large for him (although it doesn't lend in itself to useful conclusions).

 

[Ilium]

I voted 45/15 but actually I'd prefer about 70/10. If a character is meant to be good at something they should succeed at it more often than not, but if they're not trained I still like them to be able to pull it off occasionally (unless it's an obscure skill that can't be used untrained).

I like your differentiation between trained/untrained skills.

 

[Kerrick]

When I'm working on variant skill-based systems (Craft, frex) I figure that someone trained in the skill, with near-max ranks, should succeed about 50% of the time. I never really thought about untrained; someone with 0 ranks shouldn't be able to succeed more than 10% of the time, and someone with cross-class ranks would be somewhere around 20-30%, I guess.

 

[Wystan]

When I'm working on variant skill-based systems (Craft, frex) I figure that someone trained in the skill, with near-max ranks, should succeed about 50% of the time. I never really thought about untrained; someone with 0 ranks shouldn't be able to succeed more than 10% of the time, and someone with cross-class ranks would be somewhere around 20-30%, I guess.

If most Crafters who were Really good -

someone trained in the skill, with near-max ranks

- only succeeded about 50% of the time, we would have a REALLY expensive society as a 50% failure rate would drive most people to stop even trying (and remember you stated that this is the Really good ones....)

I personally think that a Trained Crafter who is maxed in his ranks should succeed at appropriate tasks 90% of the time or more, and less as the training becomes less complete and or the task becomes more complex/difficult.

Example:
Swordsmith creating a Longsword -
1. Max Trained Swordsmith (skill about 10 total) should only fail making a longsword 1-2% of the time.
2. Apprentice Swordsmith (Skill about 5 total) should fail making a longsword 20-30% of the time
3. Joe Untrained Swordsmith (Skill 0-1 total) should fail 90% of the time.

Swordsmith Creating a Quicksilver Longsword -
1. Max Trained Swordsmith (skill about 10 total) should only fail making a Quicksilver longsword 10% of the time.
2. Apprentice Swordsmith (Skill about 5 total) should fail making a Quicksilver longsword 40-50% of the time
3. Joe Untrained Swordsmith (Skill 0-1 total) should fail 99.999% of the time.

 

[avr]

I think skill checks need to succeed most of the time if the players are to be able to make any sort of plans, especially if you may require multiple checks. I voted 75/25.