One issue I find troublesome in RPGs is skill caps because they don't seem to be constructed with actual limits of human capability in mind. In general they fall short and there isn't enough difference between low and max skill levels.

I remember the text from an old Rainbow Six game manual talking about how Rainbow operatives have skills that represent the limit of human capability. How would you try and represent something like that in a game?

Are there official stats for how often, say, elite infantrymen hit certain targets in training under certain conditions? Would it make sense to try and model those stats at one end of the spectrum and then work down from there?

Skill caps tend to make me furious when they're like in VTM where your max skill is totally pansy.

Well, if you could find those stats, that would be great and allow you to model skills on that. If you find, for example, that an army 'grunt' can hit a human sized target at 50m with a particular weapon 80% of the time, and that an elite special operatives can hit the same target at the same distance with the same weapon 99.995% of the time, then you have some idea of what kind of range of skill you need to represent. Of course also having a % for 'joe average, first time picking up a gun' would be very handy as well.

Yes I've always struggled with the limit of human capability, so far I've yet to see a RPG replicate that correctly so far, so I will keenly watch this thread for ideas, as I don't really know where to start.

It's not as simple as just falling short; the limits often fall quite long as well—just look at SR3 firearms combat where a trained (Skill 4) individual with a heavy pistol can hit someone walking (+0) at 20 meters (Medium range) in minimal light (+6) slightly over 20% of the time. Hell, put the shooter behind cover (+2) and we're still talking over 10% hit rate on snap shots. Note when I say "snap shots" I don't just mean no time spent taking aim, I also mean emphasizing defense over offense (no Combat Pool use). With Combat Pool our shooter can pull a Clarice Starling and take the shot in total darkness (+8 instead of +6) for nearly 14% hit rate.

~J

You could always impose a limit on dice thrown for any particular skill, but the skill level remains with no 'hard cap' - to show that skill will overcome any obstacle. If someone has a Firearms skill of, say, 25, but the dice cap is 20 - then he's still throwing substantial dice if he's trying to hit someone from Extreme range or in pitch darkness (or both). Allowing characters to further train in those skills will continue to minimize penalties until only the luck of the dice themselves will give the character pause.

The exponential falloff of the TN-based system is sufficient to avoid the need for die caps (though the details may need tweaking); the issue is more finding a good probability distribution and consistently applying it.

Just one more argument for my "dice with continuous distributions" proposal.

~J

Of course you want to put the skill cap one or two above what gives you the nice "best" averages. One cos being the best of the best is fun if you have to work for it and two if its an average then there are going to be some above that any way.

I've never really understood the reasoning behind skill caps. Dice pool caps I can understand, though I may or may not agree with them in principle. But beyond chargen the way skills are raised is primarily through karma expenditure, and that becomes a self-limiting procedure long before it would outweigh the dice pool effects of modifiers that already exist for many of the scariest skills.

In SR3, absolutely. In SR4, much less so; thresholds don't lay large numbers of dice low the way high TNs do, and you get Immunity to Normal Modifiers pretty quickly.

(My take on the reasoning behind skill caps is that they're a clumsy patch for a broken base mechanic, but that all got discussed to death in '06.)

~J

I suppose the devs were trying to enforce some level of generalization. Sure, it would be expensive to raise automatics to 10, but you'd still see plenty of people doing that instead of getting a half dozen other skills to 2. I don't know why the devs felt they had to bind people's hands like that, but it's just my guess as to why the skill cap was imposed.

I see your point.

Let's look at the actual costs. Say the PC starts play with Automatics at 6. (Hey, if the player wants to go to 10, he might very well do this.) So, to get to 10, it costs (7+8+9+10) X 2 = 68 karma. And that will buy roughly 1.3 extra hits on average. But let's say it also allows a commensurate number of extra modifiers, so it nets 2.6 extra hits.

I don't know. You really have to want those 2 or 3 extra hits to pay 68 karma. For the same Karma you could build a whole Skill Group from 0 to 4 and be 1 Karma point shy of getting that Automatics Skill to 7, just because. Or, you could raise you AGI from 5 to 7, giving you 1. 3 extra hits on Automatics anyway (with the extra modifiers allowed) and be better at every AGI related Skill (and therefore be more rounded).

It almost sounds as if I am arguing against opening up Skill caps. I'm not, I just think the fear of someone sinking everything into one Skill is overblown.

In any case, we have Skill caps, so there.

Well, I suppose it really depends on just how abstract you want the mechanics in your RPG (OP is asking about all RPGs, not just SR versions).

Ok, so lets say shooting range or olympics or something else that is recorded is what we use, do they represent those as percentages and use a couple of 0-9 dice to represent 0-100%? How much do you think the fortunate of fate/circumstance plays into how well the best of the best do? Do you allow for that Mr. Murphy ('s Law) can even smoke the best of the best that humanity has to offer? Also, do base your best of the best on some set that someone is born with, take the Aryan nation super soldier BS approach or that anyone can realize their full potential (despite the odds stacked against them deliberately or unintentionally, they only just got a President with a darker skin tone after all) USA BS approach?

My best guess would be to make it percentile in a range, where scoring a hit in the range is a successful outcome for the character utilizing the skill. The bigger their natural ability + hard earned skill experience applicable knowledge = a bigger range representing a better chance for success. Have the Mr Murphy circumstantial factors decrease the range, maybe even throw in the "dying is fun" 0-1% roll for you are hooped no matter how awesome you are a human at this skill attempt, everyone meets their fate eventually.

Though, you could just be asking why some RPGs have their high skill/ability attempts have low chances of success, thus making it seem silly to have invested in that character skill/ability. That makes me think of all the crappy stat rolls I would end up with for character creation in D&D. No matter how I tried to tweak the situation to my favour, my character might apparently be better than the average human but it didn't seem to make my monk's quivering palm or stunning strike actually stun anyone ... and even if those things hit they lasted for dick all or made their save vs dying or whatever. I always felt like my monk was useless inside and outside of combat and never really did get to parkour up in the dungeons to foil traps or flank monsters from unexpected angles. Best of the best, my ass, unless I roll nearly all 18s for starting stats, I was getting jack shit from that character and even then since I was playing a monk in the first place because I don't actually care much for the loot the other players are fighting over...

Skill caps have two purposes.

The first, and most obvious, is game balance. Unlimited modifiers can get broken in a hurry. I recall a D&D game where I destroyed a heavy war galleon. In three hits. With a sword. Okay, that was more ridiculous damage modifiers than skill modifiers, but it illustrates the point.

The second is a little more subtle. Many skills have a point where you simply cannot do 'better' anymore. Okay, you hit a bullseye. Even if you 200 more ranks in the skill, you still just... hit a bullseye. There is no more better possible result.

The difference is, if the cap is, say, 20, the guy with a skill of 28 will be able to hit that bullseye at 1800 kilometers in a driving rain with gale force crosswinds while being shot at. He'll still have a skill of 20 whereas a guy with a skill of just 20 will have been reduced to 2.



-karma

Sure, if you choose a mechanic that doesn't scale, but then you've just moved the error up one level.

~J

I wonder if part of the issue of skill caps is related to time. Many games, SR included, go for the ease of acquisition rather than trying to realistically chart out skill development combined with time required. In SR Runners gain karma quicker the more runs they end up doing (well...its more an out of game issue, the more you play the more karma you tend to earn, unless you've got a really stingy GM). Then you can turn around and plop that into as many skills and attributes as you have karma (and gm consent) between 'adventures'. As opposed to real life, where some skills could arguably take months, if not years, to raise 1 or 2 points, in SR you can do it in at most, a few weeks (assuming the old GM standby of "Ok you guys rest up for a couple weeks").

Sometimes I wonder if it might be helped (though horribly cumbersome) to approach skills with a 'points invested in skill' plus 'points invested in penalty mitigation'. KarmaInferno's firearm example works for this. I would argue that in terms of raw marksmanship, you do run into an inherent cap. In a managed environment, say shooting range shooting paper, you'd be hard pressed to tell the difference between a guy with a 4 5 or 6 in a firearms skill. Unlike SR, where even possessing professional rating in a skill, its possible to miss the hanging paper target entirely, in rl, you'd maybe see a difference though slight, in the grouping of shots on the paper target.

But again, there's a difference between the guy that can hit a dime sized target on the range, and the guy that can hit a dime sized target while upside-down, at night, in a tornado. But you wouldn't really be able to tell the difference between the two in the range. The second guy can usually make the chaotic shot because they've invested experience in such 'real world' applications of their skill. There's a difference between the dead-eye hunter that can peg a stationary deer, vs a deer on the run, vs say the guy who can blow your head off from a mile away while you're riding in a car.

SR3 has training time rules. In my experience they're universally ignored for what I believe to be two reasons: first, they're optional rules that appear as a "gotcha" in SRComp, and second (amplified by the first) they're just not fun—the opposition doesn't train and learn skills, so really it's just delaying gratification for no reason.

As for the paper target, I don't think that holds up, especially with combat pool in the mix.

~J

SR4 has training time rules as well, and they are similarly ignored most of the time.