Saturday, 12 July 2014

Skill atrophy, part 2

So last time I was talking in general terms about how skill development in real life doesn't really match the constant escalation of most games. Let's see what we could do about that.

Introducing skill atrophy to games

If we wanted to have a system where people's skills shifted more naturally to match their activities, what could we do? What would we be looking for?

  1. Skills increase with use
  2. Skills fade with neglect
  3. Neglected skills are easier to revive than new skills are to acquire
  4. Skills follow a curve - it's easy to master the basics, but every improvement is a little harder than the last

I feel like Call of Cthulhu's system for skill increases is one of my favourites. Traveller is also decent, but is based on only training one skill at a time. In each case, using skills offers the opportunity of increasing them. In case anyone doesn't know, Cthulhu lets you increase a skill if you 1) pass a meaningful roll on that skill, and 2) fail a second roll - this makes it easy to pass rolls with higher skill, but hard to increase that skill. Traveller lets you spend time to gradually increase skills, but each increase takes a number of weeks equal to your total skill points; this means refining one skill isn't specifically harder than improving a weak one, but getting more skilled in general is increasingly difficult.

One consideration is uptime and downtime. If a game assumes a lot of downtime when skills might be practiced, or has a broad range of life skills, then a system with player-allocated practice is best. If the game assumes most events take place onstage, or has skills tied firmly into uptime actions, then a system where practice is determined by in-game use of skills makes sense.

A fairly simple system would give you N practice points to use each Duration. These allow you to maintain existing skills, improve them, or gain a new skill. Unless you allocate practice points equal to your current skill, your skill will decrease.

A slight variation would be chance-based, with a roll to determine changes in skill based on the practice points allocated.


Let's assume we have a system vaguely along WoD or Traveller lines, with skills going from 0 to some smallish number like 5. Let's also assume, for convenience, that along Traveller lines, Skill 0 is a genuine benefit in allowing you to roll without penalty, compared to lacking the skill entirely and rolling on halved attributes.

For practical purposes, let's assume that over the timeframes of the game, a person can't realistically completely change their skillset between maintenance sessions, which would take years. They have five points of Focus which indicate areas they're concentrating on. These might be chosen by the player (in a downtime-heavy game), agreed with the GM based on the events of a game (in a more onscreen game) or emerge through a system like ticks for skill checks.

Focus can be used to try and increase a skill, or to acquire a skill in the first place. Let's assume a Call of Cthulhu model here, so a failed roll on your existing skill is needed to gain the new skill point. However, training a skill will leave you less time for maintaining others. Each skill point gained through Focus must be transferred from an existing skill at level 1 or higher. Adding a skill point to a skill you don't have gives you skill 1.

In theory, this model allows you to game the system. It's mechanically beneficial to keep opening new skills at skill 0, so you can roll without penalty. However, this will depend on the frequency of maintenance sessions. Bear in mind that spending long periods with most of your skills at 0 will leave your character bad at most things. It's possible, of course, to have one skill point that you keep swapping around to new skills; whether this is a problem would depend on the system. If skill points are relatively scarce, skill changes occasional, and the boost from each point substantial, then this is a less tempting option than keeping the point in a skill you really want.


This is a more brutal system. The idea here is that keeping the edge on a well-honed ability is hard, while basic competence is easy.

Assume there are 10 skills in the game and it's a percentile system. You have 10 Focus to allocate to these skills during each maintenance session, by whatever means. You may allocate 0-2 Focus to any skill.

Allocating 1 Focus means you maintain the skill at its current level, using it regularly. Allocating 2 Focus means you are using it intensively and it will tend to increase; you can roll a skill check, and if you fail, your skill increases by 1d10.

For any skill you don't allocate Focus, you must roll a skill check. If you pass, your skill decreases by 1d10.

In this system, you can quickly gain low-level skills and it's relatively hard to lose them. The basics are easy to learn and hard to forget, while also providing a relatively large increase over zero skill. In contrast, high-level skills will degrade quickly if they aren't constantly practiced, and are hard to attain in the first place. Bear in mind, though, that a high skill losing a few points remains high (a relatively small decrease) whereas a low skill gaining or losing points sees a proportionally large change.


This system might work in a game with obvious "spheres" of skill pertaining to different kinds of activity - like combat versus politics versus technology.

Assume there are 5 spheres, and skills go from 1-5. Rank the spheres in order from 1-5, with 5 being the highest. This ranking indicates where most of your energy has been expended. High-ranked spheres will tend towards a score of 5, while low-ranked spheres will tend towards 1.

The exact mechanism doesn't really matter. Suggestions:

  • Shift skills one step towards the rank they were assigned. Simple, blunt; indiscriminate, with no difference between rank 2 and rank 5 when you're looking at a rank 1 skill.
  • Roll for each skill, with a modifier based on the difference between current skill level and the rank assigned, to see if it shifts one step. Finer-grained and with chance of change proportionate to (essentially) the ratio of effort to current skill; somewhat unstable as no overall balance of skill.
  • Roll a dicepool for each skill, with maybe two dice plus one die per rank. Each 4+ represents a point of skill retained or (if it exceeds the current skill) gained, but if successes don't equal current skill, the skill is reduced accordingly. Offers unpredictability and potential for rapid change, with high skills very difficult to attain or maintain.

The first system is simple, and might work for a system where the fine details of skills aren't all that important, as well as one with pretty frequent maintenance sessions - you're not worried about the magnitude of change, only that consistent focus on or neglect of particular skill sets will tend to affect them.

The second might work for a system that isn't too worried about exact mechanical balance at any given point, but that does want to convey a sense of characters changing to reflect their actions. It's slightly unforgiving (or potentially, generous) but you can predict reasonably well what might happen. I'd expect a relatively long gap between maintenance sessions. I'm thinking this might be okay for something like WoD games? The decay of old street-smarts as you adapt to managing contacts and underlings is inevitable; so, too, the strain of remembering and recovering those skills when some betrayal leaves you isolated and without political influence.

The third one is pretty chaotic, so might suit a game that has little interest in mechanical balance. Of course, some tweaks could be introduced to tame it a bit. Because skills can change quite a bit, it's most suitable to infrequent maintenance, as much as every few years - as such I'm thinking it might work for a long-term strategic game like (I think?) Ars Magica and so on, or games with time-skips.

Little-known alternatives

The next two ideas feel like a bit of a departure from the sort of mechanics you typically get in skills, so I'm shifting them off to one side.


In this model, you can move some arbitrary number of skill points around between games.

That's it.

This doesn't intrinsically offer any kind of advancement, although you could easily pair it with an existing advancement system, or blend the two in some way. You could, for example, rule that the number of new skill points gained with advancement is halved, but that the remaining points can be shifted between skills.

This would potentially work well in games that have skills covering most situations, so that it made sense for points to shift around as the characters age and change their activities. In a game where skills are restricted to certain spheres, it may be less appropriate - if all your skills relate to laboratory work, then becoming manager of the lab and taking on all the administration and meetings doesn't logically explain why your dissection got worse and your distillation improved.

It could also work in a game that's not very worried about the actual mechanics, except as a prop to frame gameplay around. Something like Inspectres crosses my mind, and FATE could also be an example.


For Call of Cthulhu and other BRP games, there's a simple adaptation you can make to introduce atrophy. This is a variation on the Erosion method.

Simply add a second means of ticking - either an additional box, or using the tick-then-cross method. The first tick indicates that the skill has seen some significant use in play, and I think it's important to say here that "significant use" shouldn't necessarily indicate a roll. The second tick, however, should come from a success on a significant roll, as in the original rules.

A chemist character who spends significant time and energy on doing chemistry should get the first tick, even if their activity doesn't call for a roll, or is unrelated to the official plot. What we're trying to do here is represent what the character gets up to. If they take two days off from investigation to attend a conference, or analysing samples for the CPS to pay the bills, those are fine and respectable uses of Chemistry.

When skill maintenance comes up, a single tick indicates no change - you have kept in practice. A double tick allows for a roll to see whether your skills increased by failing a roll (rolling over your current skill). If the skill is unticked, it's liable to erode towards its default value - make a roll, and if you pass this roll (i.e. get under your current skill) you lose d10 points.

To be honest, many Call of Cthulhu games have such a short timeframe that I wouldn't tend to use this - while there may be significant downtime between sessions, ticks represent uptime activities often over a few hours. It doesn't really make sense to me for skills to change based on a few hours of frenetic investigation followed by months of ordinary life. On the other hand, it might work better for a longer-running campaign where the same kinds of activities often occur repeatedly over long periods.

Any thoughts, opinions, questions, disagreements etc, post below.


  1. Interestingly I do have a *sort* of skill-erosion system built into the Regency Requiem game (which is progressing slowly due to scheduling conflicts, but is now reaching the 1820s) in which players can voluntarily lower their skills in return for - in essence - xp bonuses.

    I think the basic problem with skill-erosion type systems is that they introduce a lot of bookkeeping (I've just noticed that the word "bookkeeping" contains three double-letters in a row, how cook is that) in order to introduce a feature that can, ultimately, make the game less fun (at least in the narrow sense that it's fun to be good at stuff and not fun to be bad at stuff).

    Again, this might be a difference in taste (I think you prefer things with more of a world-simulation feel while I'm more tolerant of metagame features) but I think if you wanted skill atrophy in a game you'd want it to have some kind of payoff, like you can *choose* to atrophy some of your skills in order to get bonuses to others, rather than having skills atrophy by default, otherwise you'd have the feeling of running to stand still.

    1. Oh, cool. How's that working out?

      It definitely does introduce more bookkeeping, true. Whether it's anti-fun is going to depend a lot on what it is you want to do with the game and its mechanics, which is always something that needs considering.

      I think I do tend towards a more simulatey feel than you (not least because I tend to struggle to understand abstract systems) but again it's going to vary by game. There are plenty of games where it wouldn't add much - if you're trying to run an actual Bildungsroman then there's not a lot of point worrying about how little Hector's chicken-farming skills atrophy when he takes up his father's sword and travels the world to seek the whatever that will let him return in triumph.

      Partly this is just an intellectual exercise, because that's how I roll.

      Partly it's to address a problem you've mentioned before: it often doesn't make mechanical sense to do things in certain ways. In this case, in a lot of games the optimal way to deal with all problems is for the PCs to personally handle them, or for the most senior vampires to handle them, or whatever. Regardless how important they've become, they're often so much better than their underlings that it would be essentially wrong to let anyone else tackle the problem. Why send an army to fight the orcs when you could take them on single-handed without all those (human) deaths? Or better, intimidate them into surrender, because which of them will take you on? Why let an underling do a job that matters to you when you could do it better AND claim the credit personally AND make sure nobody's double-crossing you? If that stuff bothers you, then having people's skills atrophy is a reasonable way to make the mechanics match up with both expected reality and the way you want the game to go. It makes sense to send your warrior underlings because they're better at fighting than you.

    2. Basically, if the result of introducing atrophy to the game would be "running to stand still" then I’d generally assume it’s a bad idea, because most games aren’t looking for that feeling. If you’re playing a game full of struggle and failure – let’s say, oh, a Magnificent Seven-style game about heroes (or criminals) brought out of retirement to face one last fairly lengthy adventure – then struggling to keep your skills sharp might make sense as part of a deliberate atmosphere. Call of Cthulhu already does something like this with injuries and SAN loss, except not really the way we play it.

      My intention was to offer a way for people’s skillsets to change in response to what they’re doing, rather than permanently accumulating. I think that’s very different from “running to stand skill”, partly because I’d expect it to tie in reasonably well with a change in the game’s focus, and so degradation of little-used skills isn’t a problem because they’re, well, little-used. If atrophy would detract from the game or not fit the genre, then I wouldn’t recommend using it.

      I mean, if you look at it another way, skill atrophy can be seen as a tradeoff for having your skills increase *at all*. It only seems odd because we’re trained to expect continuous increase across the whole range of abilities, even when in many cases it doesn’t particularly fit the genre. Atrophy as I’m suggesting it is essentially a way to let people respec to fit a change in game tone, it just looks like a penalty because the reward part is already built into most games.

      To take a broad example: you could probably make your game actually simpler by cutting out XP awards altogether and only cashing-in of existing skills and abilities. People get better at the stuff they care about now, they get worse at stuff they care less about, but they care less about that stuff. Minimal bookkeeping, deals with the mechanical issues I mentioned, you still get to advance characters and get cool new stuff. Except, in your case it wouldn't be a great fit because vampires getting more powerful with age is a kind of a thing, so I wouldn't do that.

      In a pulpy adventure game, though, I would totally be up for a simple system where you just swap skill points around to represent different points in your career.

    3. If you’re playing a game full of struggle and failure – let’s say, oh, a Magnificent Seven-style game about heroes (or criminals) brought out of retirement to face one last fairly lengthy adventure – then struggling to keep your skills sharp might make sense as part of a deliberate atmosphere.

      That's very true, and I have actually always wanted to run an anti-bildungsroman, where you start off at the top of your game and gradually spiral downwards.

      I think the issue I have with - say - a Cthulhuesque tick system is that while I agree it isn't particularly punishing for you to lose points in skills you aren't using anyway (and it actually could make for an interesting feature of a very grim style of game), it's administratively punishing to have to go through all of the skills you didn't use and check if they degrade (particularly in Cthulhu, which has a *hellishly* long skill list).

      I agree that ultimately the steady increase of skill levels you get in most RPGs is just as implausible (arguably a lot more implausible) and that we mostly accept it for legacy reasons (up to the point of crowbarring it into games that really do not need it, like Hellcats and Hockeysticks), but I think there are some fairly sensible - for want of a better word - domain reasons that skill degeneration doesn't really get a look in in most games. For a start, the list of skills you don't use will always be (in a sense) infinitely long. Certainly it is likely to be a lot longer than the list of skills you *do* use. In a game like, for example, second edition D&D there basically isn't anything on your character sheet that you don't use basically all the time (especially if you don't use nonweapon proficiencies). In a game like Call of Cthulhu, by contrast, the character sheet is full of things your character will hardly ever use, but a big part of that is the focus of the game. If I'm an accountant and I take two days out of my regular accountancy job to investigate a mystery, and that mystery does not involve my doing any accountancy, I'm not going to be a meaningfully worse accountant when I come back to work.

      This said, I think you probably *could* work skill atrophy into a system in a way that wasn't an administrative nightmare. I think it would make most sense in a game that had a time-based experience system (like my Vampire game, or default Ars Magica). It would seem to run very much against the grain of something with uptime-based XP, because it seems very peculiar for actively doing things to make you worse at other things.

      Anyway, having dinner now. Hope you're enjoying Parts Foreign.

    4. As I mentioned at the end, I don't actually think most CoC games would suit atrophy anyway: in-game skill use mostly corresponds to only a fraction of your activity, and actually the tick mechanic itself doesn't make much sense most of the time. Skills you don't use decreasing would make things even worse because it'd just emphasise how much the game isn't about any of the skills that supposedly define your characters, but about spotting, dodging and shotgunning. In your example, the two days you take out of your accounting job should not make you a meaningfully better private eye, either.

      The length of the skill list would be another issue, certainly; at least with ticks you're perhaps getting something for your trouble.

      I can see a CoC game working something out here if you ran very slow-paced games and did spend time worrying about everyday activities. CoC as a setting lends itself fairly well to skill atrophy: I can easily see a mini-campaign set at 20-year intervals where characters should change substantially. But you'd probably want to handle that with player decisions rather than mechanics. The main reason I put it down as an example was simply that I thought of how to do it.

      Fundamentally, I'd say if you want characters to change focus over time, the best way to do this is mostly just going to be having players move points around. The other options offer something different if you specifically think it's more "lifelike", or want to introduce tension or a sense of decay over time. I don't personally think I'm likely to use any.

      In general, yep, I agree.