What's the optimal number of skills for an RPG?

It entirely depends on the premise of the game.

What skills exist in a game and how broad or narrow those skills are should always be rooted in what is important to the game, what actually matters to the players. If something is absolutely necessary for everyone to have, it shouldn't be a skill, it should be an assumed part of each character. Meanwhile, if something is completely irrelevant or extremely niche, it should be bundled with a load of other stuff to make it worthwhile. Compressing or expanding skills in this way might seem 'unrealistic', but not doing so just leads to awful design in my experience.

No such thing.

Add them until you are happy

No modern and sci fi shouldn't have more skills. We don't track all the skills peasants in days gone by had that we don't, so why should we track all the skills we have? I just think skills are generally garbage anyway.

Fuck no. Skill bloat is the fucking worst.

Two core problems harry skill design. The first is that skills are too broad in a dissatisfying way, for instance, a weapons skill that covers both guns and swords, meaning there is no distinction between good swordsman and good gunmen.
The second is being overly specific. Suppose a different skillgroup for muzzle-loading and breech-loading rifles. You'd want to suppose that someone skilled at firing one would be at least more competent than others at firing with the other, but a system too granular would mean a rifle expert becomes a doddering fool with a pistol or shotgun. Riddle of Steel had a similar system, but tried to work around it with a list of cross-over skills. For instance, swinging a mace is some respects similar to swinging a sword, so you could use your sword skill at a penalty to swing a mace in lieu of the mace (or mass weapons I think it was) skill

These two core problems will be distorted by the design intention. For instance, a system focused on tactical urban combat would benefit from being more granular about it's weapon skills and could probably collapse social skills into fewer.
Similarly, the appropriateness of skill groupings might change based on design. In heroic fantasy, someone could at tying up ropes could generally be thought of as also good at pickpocketing the same way a good public speaker is probably also decent at lying. So a 'streetwise' skill might encompass both petty theft and ropework.
A swashbuckling game of pirates might want to take ropework out of the broad umbrella of criminal enterprise, and either throw it into a broad sailor bucket, or make such a skill a granular entry of it's own.

so there's a logical constraint (being good with ropes doesn't mean you are better at stalking people) and the contextual constraint (this skill covers the survival tactics of the underclass)

Considering all this, I've found less than 16 skills and things tend to be thrown together that may not go together; one game I was working on somehow found economic skills and stealth in the same group. The exception here would be a rules light/freeform game. You could have very few to none in these cases

More than 20, and you need to start upping player resources to ensure they can enact some measure of consistent talent between similar skills, and even then this can be unstable and does not guarantee a solution. The exemption here would be specialist games, like the old chaosium/palladium games with 50 scalar martial arts styles to keep track of or 60 percentile academic skills.

so about 18-20 for basic needs, I've found.

OD&D didn't have skills and it handled dungeon exploring and problem solving just right, using common sense and stat checks where necessary.

from a design point of view, those attributes are your skills
The old school approach certainly works, but if there wasn't a comprehensive reaction to it's approach to RPG's it would probably just be called the current school approach

wan skil par mekanikall soobsistum, fite mi beaches.

Define mechanical subsystem then mate

>from a design point of view, those attributes are your skills
Your attributes were not your skills. Your skills were whatever ideas you as a player could come up with and how reasonable the DM thought those ideas were.

Attributes were not the primary means of mediating characters' interactions with their environment the way that a modern skill system is thought of, and if you think they were it's just because you're too deep into the modern trad rpg mindset to realize how different things were.

Relying on 'player skill' is dumb.

Yeah, that's just moving the categorization problem into "subsystems" instead of "skills"; same shit, different semantics.

>some antics.jpg

>using common sense
OD&D basically used GM fiat, with interpretations varying wildly between DMs or even the same DM in different years. DMs would interpret "class skills" on pure whim. It was chaos.

For example, would a paladin know how to perform all the rites of their god like a priest (sanctifying, performing marriages, blessing newborns, etc.)? Some DMs would say yes, others would say no, that's a Cleric's job.

from a design point of view, your attributes were your skills. Documented abilities that improve as play progresses. The referential source of mechanical resolutions.
Overlaps completely with boilerplate attributes (or a skill system with no attributes but all the same skills; the difference is semantic)

Ay mekanikal sistum iz any mekanikal element ov tha gaym vat iz disstink frum enee ova elleement, ay gud exhampul iz vee "Sav vusus" sistum fram aydee un dee oorr vee aysee sistum, itch funkshun diffrentlee - wan taks iz kyu frum DEZ und eyetems like arrmer, wiil savs wor deeriveded from ova attreboots - vus wun cud immagin ay sistam war evri attreboot iz linkad too wan mekanik onlee.

>rom a design point of view, your attributes were your skills.

Not really - in pre-3e D&D, your CLASS LEVEL was your skill level if you weren't implementing the optional "proficiency" system that AD&D introduced, and which was ultimately repurposed as the skill system of 3e.

In other systems, other than D&D ones, you generally have attributes as your back up "if you don't have a skill relevent to a situation" thing you roll with, usually with a penalty of some kind, but also often impacting sub-mechanics more than skills would, like how hard to kill your character is or how hard are they to fool or sneak around.

Often attributes act as maximum caps on your skills as well, especially in systems that lack digital "level"ing and classes.

In many ways, skills are one sub-mechanic of many systems, though most modern systems prioritise them for PCs and make them the one mechanic players interact with the most.

This actually.

And not in a facetious, unhelpful way. Consider what skills you actually require for dramatic tension. Do you need to include things like psychology as a skill in your game? Does it add dramatic tension to roll it, or is it just tedious time-wasting? If the answer to the first is no, and the answer to the second is yes, then you don't need the skill. That doesn't mean you can't have a character who is/was a psychologist, and be privy to information drawn from that background, but it should be firmly a background related skill and not something essential to the character's in-game roll-the-dice skillset.

Having an astrogation skill in an SF game is going to be pretty important, especially if PCs are flying the ship. A bad roll could send them off course, use up more fuel, or do several other negative things that not only create tension every time they travel, but leads to taking the story in organic directions.

A bad jump sends them off course, and they end up using a lot more fuel than they would have getting back on course, meaning they need to do a few jobs or scrounge up fuel somehow to get to their next destination. Bam. Instant adventure material.

ohnlee a korter ov it iz. fukkin grate buk tho

as much as needed.