You could add some more depth by lets say having a specialization-tree with unlockables that require some level in the general skill.

I like this idea! Knowing that as the game designer you must plan out every step of every branch of that tree, having a special move/buff at certain levels of advancement can encourage specialization. Deep knowledge of a specific weapon might let you know that every 5th shot of continuous automatic fire from that make/model has a high % chance to cause a jam. This could translate into a bonus to reduce or eliminate that % chance because you know how to maintain the weapon to keep the jam from happening.

A bonus for the pistol line could be explained that by studying the use of the pistol so much, the character reflexively knows the proper forms and uses for that class of weapon. This leads to unlocking trick shots and advanced tactics available to them in gameplay at various levels.

I'm thinking that you are looking more for a system like (but not exactly like) Star Wars Galaxies where there was an overall combat class, but could specialize down certain weapon-type paths to get specific skills when using that class of weapon.

I'm just thinking out loud here...

Thanks for the great replies! Definitely reconsidering critical failure elimination.

Quick background on how I'm thinking of doing skills to frame this better: All skill levels are open ended, as are specializations, raised either by paying for training or through use. Skills have levels, called Ranks, which are pitted against challenges that have a corresponding Challenge Rating. If Rank equals Challenge Rating, you get a 50/50 chance at success. Each Rank above confers a percentage bonus (not determined yet, may be +10% per rank up to a cap of 95% chance of success).

Example: You have Gunnery at 5 and a weapon that has a CR of 5 to use. This gives you a base chance of 50/50 at hitting a target (excepting other factors like range, evasion modifier of target, etc).

I should mention also that it's for a party-based game possibly with a form of permadeath. I intend to discourage reloading by zapping a slowly building resource that powers character abilities if you reload anything other than the active game (which saves upon exit).

Can the player predict how and when skills are going to be useful?

For Light Arms specialization to be a real choice, there must be other useful specializations (e.g. Sniping) and their corresponding weapons (e.g. a large caliber rifle). Are you including useless skills? Does the player know that they should be avoided?

Third, there must be a way to win the game with any choice of specializations, forcing you to avoid, among other things, unavoidable fights that are much easier with one skill choice.

Honestly, I'm not really sure if the player's going to be able to predict how to use specific specializations other than by experimenting with different challenges. I'm trying to procedurally generate a lot of the game and may end up deriving the nodes of challenge from a mix of what the player has invested in skills (along with incentives to skill up) and somehow telegraphing a warning on the things the procedural mission will decide they need. But this is very good advice to keep in mind.

Shadowrun (PnP) skill web system...

Very helpful example. I was considering faster advancement for the specialization, but I'm not sure of what to do if you completely overpower a single specialization other than create situations where the player could flat out fail (which makes it a bit of an investment hazard). What did Shadowrun do to keep specializations from becoming overly powerful?

Are specializations a one-off cost?

No as mentioned above. The idea you propose for one off advantages would work well with the character special ability system, though, where you have to spend character points (earned only through missions and a few other sandbox activities) to purchase abilities.

-The player gets a bonus related to the specialization but that isn't an attack modifier, i.e. specializing in Pistols lets the player notice more easily if an enemy has a pistol, or confers a defense or dodge bonus against pistols

I like this idea. It's unexpected, which is a nice touch.

I could probably come up with some more diverse ideas with more examples of skills and multiple specializations within them.

Not a complete list, but a sample

2) Reloading(how fast do you reload a gun, very important)
3) maybe gun-maintenance(keep your gun from breaking down due to be worn out, or make it usable even in the rain, again depends on your game)
4) doublehanded(using two guns at the same time, on lower levels with big penalties, level up to cut the penalties)
5) and maybe a skill to increase the damage from guns(dunno whether it would clash with your level of realism)

Thanks, I'll take these!

If your system has a concept of critical successes, a specialist might also have an increased (but never guaranteed) chance of critical success.

What about general skills having no chance for critical success but still including critical failure? Too harsh? Greater and greater specialization would unlock more abilities and a greater chance for criticals, which as you say would not be guaranteed but would go nicely with special character ability traits that could be domain specific, like criticals with weapons, criticals with computers, etc.

You could add some more depth by lets say having a specialization-tree with unlockables that require some level in the general skill.

I don't have trees but you gave me the idea of maybe cross linking these open-ended skills with character abilities, which have to be purchased through character points you earn during play. Maybe abilities require specializations, something like "Lightning Draw" which gives a plus to initiative in gun battles when holding a pistol requiring a specialization of X in Light Arms.

For instance shooting at unaware target might require a 3 success to hit, but if you have the sniper specialization then 5 success means a head shot that kills non boss enemies.

I like this. I could see generalizing this to something like "called shot" requiring a specialization of X against a given enemy or "fast hack" giving a bonus to only certain types of security systems.

I also like the package system you're proposing but I don't think I can work it in with open ended skills.

Shadowrun (PnP) skill web system...

Very helpful example. I was considering faster advancement for the specialization, but I'm not sure of what to do if you completely overpower a single specialization other than create situations where the player could flat out fail (which makes it a bit of an investment hazard). What did Shadowrun do to keep specializations from becoming overly powerful?

The game system itself did not handle that aspect. It was up to the Game Master to keep things interesting. Like if someone was rolling with 15 levels in their Musashi Samurai Sword specialization, they would be neigh-unstoppable within arms-reach. So the GM (usually me) would add in a sniper from a few hundred feet away to give that character something to deal with creatively. Teamwork would ensue as someone else would have to take out the sniper that was harassing the melee guy. In their element, the specialist would be overpowered. The trick was to moderate how often that element happened. You didn't want to completely eliminate that element, though, because that's where that specific player had the most fun.

For video game mechanics, a simple solution would be to cap specializations at a certain level (10?). The rationale is "you can only get so good at something." Once you reach perfection, further practice only helps maintain the perfection. You don't actually get any better at it. Maybe make that last level "get rusty" over time if not used. Say you plan for 10 specialization levels per weapon type. The 10th level grants a super move for that class of weapon. If that skill isn't used for 5 missions, the level drops to 9, and you lose that super move. You can earn that 10 again later, but at a reduced cost since you knew it once already.

Shadowrun didn't limit how good anyone got at any particular skill, but hyper-specializing meant there were points not being spent in other skills. And there were several dozen, if not hundreds, of skills. That specialist might be awesome at hacking a computer, but couldn't dodge a spitball to save his life. The best teams wound up being a group of five or six specialists that all knew their roles well. Nobody could run alone and succeed at the tough jobs. Kind of like a sci-fi Ocean's Eleven.

So it often happened that a Firearms skill would be at 4, Pistols skill would be at 6, and Baretta 9mm skill would be at 8.



In that case, to raise the skill levels, you would have to spend:

8 points to raise Firearms from 4 to 5 (2 x current level)

9 points to raise Pistols from 6 to 7 (1.5 x current level)

or 8 points to raise the Baretta 9mm from 8 to 9. (1 x current level)

(Each line is paid for and advanced independently after character creation.)

How the first level was handled (it should cost zero)?

An interesting system BTW...

So it often happened that a Firearms skill would be at 4, Pistols skill would be at 6, and Baretta 9mm skill would be at 8.



In that case, to raise the skill levels, you would have to spend:

8 points to raise Firearms from 4 to 5 (2 x current level)

9 points to raise Pistols from 6 to 7 (1.5 x current level)

or 8 points to raise the Baretta 9mm from 8 to 9. (1 x current level)

(Each line is paid for and advanced independently after character creation.)

Do these all add up to one "final skill value" (general + concentration + specialization) that is used as a skill check?

How the first level was handled (it should cost zero)?

An interesting system BTW...

I am operating from memory here as it's been a long while since I've played Shadowrun. Likewise, it's Shadowrun's ruleset I'm relaying. I didn't come up with it.

The skill levels denoted how many dice were rolled in the skill test. The separate lines do not add up when using the skill. The most appropriate line was used. For example: if the character shot the Baretta 9mm, the 8 dice were rolled. For a Desert Eagle, Pistols rolled 6 dice. For a shotgun, Firearms rolled 4 dice.

On character creation, you had so many points to put in skills. All the points were given to general skills. After all the points were used, concentrations could be selected if desired. The appropriate general skill would be lowered by 1 to gain a concentration under that generalization for 1 more than the original level. Then a specialization could be selected under that concentration with the same process, but by lowering both the general skill and the concentration by another level.

To get the Firearms example above, the Firearms skill started at level 6. Concentrating in Pistols, the Firearms went to 5 creating Pistols at 7. Specializing in the Baretta 9mm, Firearms is reduced again to 4, Pistols is reduced to 6 (from 7) and Baretta 9mm is created at 8 (7+1).

If you had a favorite Baretta 9mm that you always carried around like a family heirloom (and enough ammo for a prolonged encounter), that's not a bad deal. But if you rely on scavenging weapons and ammo from defeated foes as the adventure progresses, it might behoove the player to remain a generalist so that their effectiveness isn't diminished once they run out of ammo.