Skills have been my buggaboo for nearly as long as I’ve been writing about RPGs. Every time I look at a skill system I see some new problems that I want to fix. I don’t know whether I’m working towards some platonic ideal, or if I’m just fickle. Either way I can’t help myself. There’s always something that can be done better.
For the last few years, most of the referees I play with and learn from have been using some variation on Skills: The Middle Road. If you’re unfamiliar, the breakdown is pretty simple. When you check a skill, a roll of 5 or greater indicates success. For untrained characters, each skill check is rolled with a d6. As a character advances in a given skill (either through class features or training), the die they’re allowed to roll increases from 1d6, to 1d8, to 1d10, and ultimately to 1d12. The elegance of the system has always appealed to me. As recently as last year I was planning to integrate it into my games, and a variation of the system was originally going to be described in my LotFP house rules. But then I sat down and did something I almost never do: math.
I love math, but I’m shockingly bad at it. It was never my strong suit to begin with, and being homeschooled from the 3rd grade up didn’t help. As such, obvious mathematical realities sometimes escape me. Using AnyDice, I tried to work out the functional difference between The Middle Road, and the way skills work in Rules-As-Written LotFP. (Skills start at a 1-in-6 chance of success, improving to 2-in-6, then 3-in-6, etc).
While The Middle Road gives players a higher success rate at lower levels, RAW LotFP quickly outpaces it. The very highest level of ability in Middle Road only gives a 7-in-12 chance of success, or 66.67%. In RAW LotFP, a character with a 6-in-6 skill rolls 2d6, and the check fails only in the event of double sixes. That’s a 35 in 36 chance of success, or 97.22%.
Now when I first figured this out, it convinced me that I wanted to stick with RAW LotFP. At the time, my reasoning was that players ought to be able to truly master a skill if they choose to devote themselves to it. After all, choosing to master a skill completely leaves other skills underdeveloped. There’s a natural balance there.
More recently, my thinking has changed. A system which allows total mastery of a skill is a system which allows certain challenges to become completely trivialized. If the players can always unlock a locked door, why even place locked doors at all? As a means of justifying the player’s investment? Even players seem unenthusiastic about rolling dice when they feel certain of success. And in the rare events that they fail one of these nearly-certain rolls, they seem more upset by it than usual. As if the dice just told them they’d failed a “walking” check.
When failure doesn’t seem like a possibility, it’s a lot more frustrating when it happens. It might be better to allow players to reach a level of mastery that no longer called for a roll at all. One where locked doors simply no longer existed for a master locksmith. It’s not the sort of system I want, but I think it would be better than a 2.78% chance of failure.
It could be argued that if players have reached such a high level of mastery, they ought to be adventuring in areas with locks that require checks made at -1 or -2. This is a possibility described in the LotFP rules, and is entirely legitimate. But I don’t like it.
Part of the reason oldschool skill systems appeal to me is the lack of any need for the referee to determine the difficulty of a task. Getting rid of Difficulty Checks was one of the best things about quitting Pathfinder. DCs are the fuckin’ worst. They add a ton of boring, yet necessary preparatory work to the referee’s job. They encourage the referee to reduce player agency by saying “Well, this door allows the players to skip half the dungeon. So it’ll have a REALLY high DC.” Furthermore, when a DC has to be improvised, it’s difficult to choose a good DC off the top of your head. And finally, if the referee is placing ever-increasing DCs in the game, at the same pace that the player’s ability to surpass those DCs is growing, then why have any advancement at all?
If I place a lock that requires a check at -2 in my adventure because of the fact that my players are really good at opening locks, then why did I allow them to waste skill points that I was just going to invalidate through future dungeon design?
The Middle Road has none of these problems. Furthermore, it has two great benefits.
First, it enables the referee to include weirder skills in the game. Consider the problem of the Law skill that I mentioned when I drafted the Lawyer class a few months ago. I think the law skill is super neat and fun, but it’s also powerful. If a player were able to master it, they’d be unstoppable. But so long as they still have a decent chance to fail any given check, they won’t be able to push their luck too far.
Second, and more importantly, it preserves the low level experience. It’s almost universally agreed, at least within the OSR, that low level play is the best play. The most challenging, the most engaging, the most fun. People prefer low level play so much that many campaigns (much to my frustration) are terminated just as the PCs start to reach mid levels.
The reason low level play is so treasured is that the game fundamentally changes as it goes on. As PCs reach higher levels, more and more challenges become trivial for them. They don’t need to worry about torches once the magic user learns Light, they don’t need to worry about rations once the cleric learns Create Food and Water, and they don’t need to worry about locked doors once the thief has an almost certain chance of opening them easily.
Keeping skill checks uncertain, even at the highest levels of mastery, maintains that low level play that we all enjoy.
And as an aside, Skills: The Middle Road could easily be expanded using a d14 and d16. Not all Zocchi dice are reliably random, but those two are. I’d love to see them come into more common usage, since they so nicely fill in a lot of the gap between the d12 and the d20. And even with a d16, a player would only have a 75% chance of hitting 5 or higher. Which is still a respectable failure rate.
I’ve been thinking about skills quite a bit recently as well. It occurred to me that not all skill checks serve the same purpose, even when they are the same skill. Lockpicking, for example, can serve at least two purposes in the game.
One, as a resource expenditure (of Time). In that case, the better you are at picking locks, the less time it should take.
Two, as a point of interest/interaction within the game. In that case, we’re talking about a large, intricate, or somehow adversarial locking mechanism that has to be overcome in order to accomplish some goal within the game. A simple die roll is probably not going to be very satisfying.
How do you think similar skills with different purposes could be objectively defined (or, can they) and differentiated?
I think it’s a matter of how much focus you want to put on a given system. I’m open to, and even interested in, the idea of turning skill checks into more complex mechanisms that require the players to participate in a kind of “mini game,” but it’s not something I’m focusing on in my own games right now. For the moment I’m enjoying the simplicity of quickly resolving success/fail with a single roll, and using time as the expended resource.
That being said, my buddy Courtney Campbell over at the Hack & Slash blog actually did run a more complicated lockpicking system, based on the game “Mastermind,” I think. The idea was that there are 4 different lockpicking “techniques,” and you had to figure out the right technique for each lock. Some locks might have multiple layers to them. So if you ran into a “4” lock, then you’d have to figure out 4 techniques in a row in order to get past that lock.
He explains the BURP Lockpicking system in this post: http://hackslashmaster.blogspot.com/2014/07/on-locks-and-keys-redux.html
I played in that system for a year or two. It was really fun!
Awesome, I’ll have to check that out. Thanks!
There might be some contradictions between the desired simplicity and the features sought for a skill system…
I believe my players and I don’t have any problem with variable difficulty checks. I certainly agree it can be tricky to assign difficulty check scores, often in just a few seconds… but this might be the only drawback. (Disclaimer: we don’t use Pathfinder either, our rules are all homebrew). And I think a DM adjusting DCs explicitly to counter the skills of his player is just a bad DM.
A master climber doesn’t need a check to climb a wooden palisade, because he’s simply over the minimum even if he rolls a 1. He might even be quite safe climbing regular stone walls. I agree that players don’t want a 2.78% chance of failure on what should be trivial tasks.
But what about a smooth natural rock wall? What if there’s an overhanging part? What if he’s being harassed by flies, or hippogriffs? If its raining? If he’s holding his pet goat in one hand?
The variable DCs are exactly what makes skill masters shine. It’s not the trivial stuff; it’s when they pull off stunts that are absolutely heroic. And that’s when the players know the skill points have been well spent.
Could you expound on that first statement?
“There might be some contradictions between the desired simplicity and the features sought for a skill system…”
What’s the conflict between the simplicity / features I’m looking for?
Summarily, I believe the desired features are:
– When skilled enough, mundane/trivial tasks should be an auto-win (and not 2.78%, the roll is just annoying and failure even more so)
– When skilled enough, performing extraordinary and heroic tasks now becomes possible, with a serious possibility of failure
You mention the problem that locked doors may simply no longer exist for a master locksmith. Well, there’s a very wide range of locks out there, some of them even enchanted, or perhaps divine in nature!
To me, the most direct way to resolve this is with a difficulty check: a random roll, plus a constant bonus from the skill level, against a given number. The only real drawback is the spontaneous improvisation of DCs, I agree on that.
By the way, I found amusing the mention that a DM might be tempted to raise the difficulty of some checks, to match their players’ skills…
I believe I unconsciously tend to do the opposite: I may be generating situations where my players can put their best skills to use, giving a chance for each one to shine for a moment (example: a subtle but useful hint that can be obtained by a successful Bowmaking check on a very peculiar bow). It may help to ensure everyone has a great time.
Great blog by the way! I especially appreciate the quick lists of ideas (like these 100 curses), to be inspired by or to borrow directly.
Ah, see here is our misunderstanding. I wouldn’t agree with either of the desired features you listed.
I never want any skill based task to become an auto-win. A locked door should remain an obstacle no matter how skilled you become at picking locks.
Nor do I want completing heroic or extraordinary tasks to become possible with skills. If it’s possible, the success should be possible at any level.
You’re viewing skill based tasks with greater gradation than I am. With some being more or less difficult than others. When viewed in that way, DCs are necessary. However, I view skills as having only 3 possible difficulties: Yes, No, and Maybe.
If a player tells me they want to open a door barred from this side, I will say “Yes.” This task is trivial, and no check is required for it.
If a player tells me they want to open a door that is barred from the other side, I will say “No.” They have no access to the locking mechanism, and so the task is impossible.
If a player tells me they want to open a locked door, I will say “Maybe. Roll your lockpicking skill to find out.” It doesn’t matter if the door has a rusty old padlock, or a brand new internal lock built by a master locksmith. All locks are functionally the same.
I’m glad you’ve been enjoying the blog!
Thank you for the reply, though I may still be confused by the misunderstanding. The original post says:
“Even players seem unenthusiastic about rolling dice when they feel certain of success. And in the rare events that they fail one of these nearly-certain rolls, they seem more upset by it than usual.”
And I agree! Rolling dices when you have a 2% chance of failure is bothersome: the outcome is never exciting (and failure somewhat frustrating).
But then you say:
“A locked door should remain an obstacle no matter how skilled you become at picking locks.”
So… to clarify this, the greatest lock picker in the realm, who famously managed to enter the king’s vault, could still be defeated by a crummy lock made the local blacksmith of a small backwater town. And then, someone else totally unskilled at lockpicking could open it up with ease.
If I’m not mistaken, with that kind of system, it’s not the level of the skills that matter, it’s how many people can throw a dice.
Let’s do a little math to check. Let’s say an unskilled character has 20% chance of success, and the best lock picker has 98% of success. If I were to hire 20 unskilled commoners to try to pick the lock, the total probability of success would be 98.8% (1.0-(0.8^20)). Bottom line, it’s much better to hire 20 commoners than waste time seeking (or becoming) the best lockpicker in the realm…
Personally, I think such a system could be improved, though I also realize it’s a matter of style for each DM.
That assumes that multiple checks are allowed without any penalty.
To again use lockpicking as an example: the first check is to see if you can open the lock quickly and quietly. If you fail, then the lock has stiffened, and if you want to try again it will take 10 minutes, and require an encounter roll. If you fail a second time, then the lock is really fucked up, and all subsequent attempts will take half an hour of time, or 3 encounter rolls.
So it’s always best to have the person with the highest probability of success making the check.
Well, you assume the time is a constraint; they could have brought the locked chest in town, asking 20 random commoners to try to unlock it. (and if you would say the lock is getting damaged, we can suppose they are trying to identify a strange tome/tablet)
I realize you want a system where players always have a chance to succeed (and fail) at anything. Personally, I think that may devalue the skills themselves.
In real-life, luck doesn’t replace a lack of skills. A non-programmer will never be able to write a small piece of software in a few hours, no matter how lucky he is.
That also applies to physical tasks. My friends and I are rock climbers, our levels are between 5.10c and 5.11d. Our probability of failure climbing a 5.9 route is exactly 0%, it just never happens. For someone with no climbing experience whatsoever, the probability of failure on a 5.9 is close to 100%… unless he has exceptional strength to compensate the absence of technique (there comes the strength and constitution modifiers).
Once again, it’s also a matter of DM gaming style. I know my personal preference is for skills that make a big difference, way beyond what pure luck can achieve.
There are a number of points here to address, so I’m gonna go through them in a quote / response format.
“Well, you assume the time is a constraint; they could have brought the locked chest in town, asking 20 random commoners to try to unlock it. (and if you would say the lock is getting damaged, we can suppose they are trying to identify a strange tome/tablet)”
1. Since we’re talking about how I run skills in games, I’m not actually assuming anything. I’m describing my methods.
2. If the players go through the effort of hauling a chest back to town, then I have no problem allowing them to open it for free if it is possible to open via lockpicking.
3. Technically it would still be faster and more efficient to have the person with the highest skill do the work; but it wouldn’t matter. If they haul it back to town I’d be happy to handwave the opening of the chest. (Again, assuming it’s possible to open via lockpicking.)
4. I don’t use knowledge skills in my games, so the only skill that might be used to identify a strange tome or tablet would be the Language skill. The Language skill may only be attempted once, and if it fails, then the language is forever added to the character’s “not known” list. Village commoners wouldn’t even get a check, though perhaps there is a sage in town who might get a check.
“In real-life, luck doesn’t replace a lack of skills. A non-programmer will never be able to write a small piece of software in a few hours, no matter how lucky he is.”
If the truth of things “in real life” is important to you, then you and I are coming from very different schools of thought. I hold that realism is the death of good games.
I look for rules that create a fun game, not rules that emulate reality.
Fair enough, different rules for different gaming styles.
Like I said, my personal preference is for skill mastery to open up new possibilities way beyond what luck can achieve. (You want to try picking that lock sealed by divine magic?… All right, roll it!) That can be fun too.
Thank you for the discussion!
Always happy to chat about games. I enjoyed this.