Friday, July 15, 2011

Difficulty with Difficulties

Ok, so yesterday we established that we’d be looking at a d6 tally system with a baseline difficulty of 4. We want to expand on that so as to better handle a die pool of up to ~5d6. Given that, there are two different vectors of approach here. The first is the purely mechanical, while the second is conceptual, speaking to the role of numbers. They weave together, so let's five in from one direction and see what comes of it.

First, we'll start with numbers. Now, one great thing about d6 systems is that the numbers are pretty well known, and there are decades of games trying to come up with interesting fixes to smooth out the progression of the average roll, since that troublesome 0.5 makes life complicated. These decades are part of why, as I noted yesterday, 4 is such a magical number. And the good news is, there are a few other magical numbers.

The first candidate is 7. 7 is a great number for two reasons. First, it's the likeliest outcome of a 2d6 roll, and second, it's the first roll that's outside of the possible scope of a 1d6 roll. That latter is handy because it provides a fantastic model for something that may not be hugely difficult, but which requires training to be able to accomplish. For example, even if you know how to use a computer (1d6) you don't necessarily know how to write even a simple program. That requires specific knowledge and training - you're not just going to "get lucky" if you keep trying. That's exactly the kind of scenario where a difficulty of 7 is a handy tool.

After that, the next magic number depends on how you look at things. 13 has many of the same benefits of a 7, except that you need at least 3d6 to hope to hit it. The difference is that it's a bit less likely to succeed - 13 on a 3d6 is harder to hit than a 7 on 2d6. Now, this might suggest a compromise middle step of 10, since that's a midpoint on 3d6, and it's +3 from 7 and -3 from 13, and given that 7 is 4+3, and 4 is 1+3 (1 being guaranteed success), there's some numerical elegance in making the progression 4,7,10,13 (especially because it can be expressed as "base difficulty of 4, with quality of success increasing by one 'step' for every extra 3", something similar to what's done in a few other dice pool systems).

The problem is there's some conceptual roughness to it. I mean, yes, I could easily say:
4 - Mundane
7 - Difficult
10 - Complex
13 - Boggling
or something equally pithy and it would still be better than a lot of games (which set their baseline too high) but really it would be utter bullshit. Those terms are crazily subjective, and while I don't object to the GM interpreting situations, they provide the GM no practical guidelines for how those things are actually set, which would be irresponsible of me.

But if we drop the 10 we get something that's not quite as intuitive a progression, but is one that gives us a real, concrete basis for the progression: the numbers are such that if you are not at least operating at a certain skill level, you can't hit them. That means 4,7,13 (and 19, if we really feel it's necessary).

Now, what those things will mean are going to vary depending a lot upon the specifics of the skills, and when this ends up in a system, some of that is going to have to be offloaded, but the basic progression is pretty simple.
4 - Normal. The difficulty for day to tay tasks that might be difficult, but require only familiarity with the action being accomplished. For example: Disinfect a cut, perform the Heimlich maneuver.
7 - Expert. Difficulty for a task which cannot be accomplished without at least some proper training and experience. For example: Perform more advanced first aid (proper splints & bandages) or give CPR.
13 - Master. Difficulty for a task which requires intense, specialized training. Example: Perform Surgery, prescribe drugs.
19 - Past-master. Perform a hyper-specialized task. For Example: some sort of specialized medicine, like brain surgery or the like.

All of which is to say, the difficulty levels are built on a clear understanding of "Do Not Try This At Home". If this is something that anyone with a little familiarity could do with luck or hard work, then difficulty is 4.

Now, I should note that I view success as trickling down. If you hit a 7, you also implicitly hit a 4, which may suggest certain bonus or not, depending on the task. If you are doing neurosurgery (19) and roll a 17, then the failure is in the specialized part of the activity. In contrast, if you rolled a 12 (lower than needed for less complex surgery) then the problem was with the surgery as a whole. You get a lot of meaning trickling down through the tiers.

Also, doing this implicitly folds in duration of activity. Climbing a mountain is a task anyone could conceivably accomplish(4). Climbing a mountain in an afternoon probably requires training and experience(7). If you roll to climb the mountain and roll a 6, you still successfully climb the mountain (you beat a 4) but it's going to take you longer than an afternoon (since you didn't hit the 7).

Now, this is still just a starting point. We still haven't added in fiddly bits, and we haven't subjected it to the real test - a conflict system - but this seems like a solid start.


  1. So one issue I see looming here, that I see in a lot of dice systems, is that increased skill also means increased variance.

    First, let's use stddevs, because they're more useful. So, your average result is going to increase linearly with extra dice:

    1. 3.5
    2. 7
    3. 10.5
    4. 14
    5. 17.5

    But your stddev will also increase (these numbers are approximate, gained from simulation of 10,000 rolls):

    1. 1.7
    2. 2.4
    3. 3.0
    4. 3.4
    5. 3.8

    So as you increase in skill, yes, your average result will increase, but so will the variance of your roll. This is probably more of an issue for opposed rolls than for these static difficulties, though, and you've already ruled opposed rolls out. But I just wanted to bring this up as something to think about.

  2. Have you considered a progression of 4 -> 7 -> 12 -> 18? While 7 makes sense for a "must be skilled at all" roll, one of the great parts in fiction, gaming, and real life are those successes that are made against all odds (rolling the critical hit that is the only way to defeat an enemy, making that last-minute life-saving move you never thought you had in you, etc.), and those numbers will make that a possibility without cheapening those moments.

  3. 10 is usually in there for the same reason that 4 is, and that's because it's roughly the mid-point of a roll (3d6 in this case). But I think it's interesting to think of this in terms of "can't possibly get this number."

  4. I know that this is completely beside the point, but I'd swap Heimlich and bandaging around in your example of difficulties, I've trained both and bandaging can be done with a bit of common sense and dexterity while the Heimlich really needs specialised training.

    (As an aside, Dr. Heimlich is an... 'interesting' character which explains why the 'Heimlich' is called 'abdominal thrusts' by medical professionals)

  5. I'd probably talk in terms of success rather than difficulty, so that:

    4 - Ordinary Success
    7 - Expert Success
    13 - Master Success
    19 - Past Master Success

    And then require that certain tasks to have a certain level of success to succeed fully. A lesser result is a partial success.

    This changes the emphasis from overcoming a specific difficulty to achieving the best result that you can.


  6. One element that always bugs me, is the that there are some failures that just shouldn't happen.

    If we follow the surgeon example, rolling 3d6 he still has a chance of failing to disinfect a cut, but weirdly, the neurosurgeon rolling 4d6 does not.

    And that kind of scenario always bugs me

  7. Unsurprisingly, Rob, this has some conceptual links to a post you made a long time ago about tiers of competence in fiction. Glad to see that notion is still bouncing around in your head, as its one of my favorite posts of yours.

  8. @Kit: Actually I don't think that the variation is going to be that much of an issue, since what we are actually doing is using the random dice roll to simulate random (and generally hidden) variables innate to the skill test.

    The skilled neurosurgeon is still going to approach the job with the same 5d6 of ability, regardless of whether they actually roll a 5 or a 30 on the skill test.

    So what does each roll actually mean. Well in the case of a 30 it was obviously a textbook case. For example the AVM, stroke, or clot was located at the surface of the brain and easily removed or cauterized. A trivial case. In the case of a 5 the problem was located deep in the patient's brain, close to critical autonomic structures, and couldn't be safely accessed without risking the patient. In other words it was beyond the ability of any expert neurosurgeon (rolling 5d6) to perform. It does not mean that the patient dies as a direct result of rolling a "5" on the surgery. It just means the neurosurgeon can't help with this problem. It was inoperable (which is probably why the roll should take in the entire operation, from diagnosis to operation to recovery).

    We are inured with a success or failure regime relating directly to the character's abilities. However rolling more dice doesn't mean that the character performs better - it simply means that the character has greater expertise at overcoming likely adverse results that would impede a person of lesser skill. The fact they can't cope with the situation as determined by the dice roll doesn't mean that they generate the failure - it just means that actual situation is beyond their ability.

    Similarly a low mountaineering roll means that difficulties were encountered that the mountaineer couldn't overcome at that skill level. It doesn't mean the mountaineer botches. [Although it may be good to add a "botch" rule to the system indicating something has gone wrong if all 1's are rolled. That's where hidden circumstance, such as the ice flow (hidden under the snow) delaminating, gets you.]

    So your neurosurgeon (5d6) can overcome greater difficulties than your surgical registrar (3d6) or surgeon (4d6). It doesn't mean that your surgical registrar failed to disinfect the wound site, only that the surgery was beyond them (although since they couldn't make the 19 difficulty they would obviously face a board review and malpractice claims even for attempting it, as would a surgeon who wasn't a member of the College of Neurosurgeons, unless it was a critical emergency and there was no other option).

    On the other hand it would only require an expert success for a surgical registrar or doctor to stabilise a patient and notice the problem and refer them to a neurosurgeon, which is 7+ on 3d6.

    And on this point, it also probably means that a character's skill base should become narrow and more specialised with height. That is, a character might be a neurosurgeon (5d6), but when presented with an orthopedic case might only have the ability of a surgeon (4d6), or as a doctor (3d6) when dealing with a viral outbreak, or even as a med student (2d6) when doing something they haven't had to do since med school.

  9. One pet peeve I have that I think your system here addresses is that I loathe systems that say, "Okay, X constitutes a success" and then go on to say, "Actually, X is a minimal success, kind of lousy, really, and in fact, almost not a success. In fact, sometimes you may need more than that for it to be a success." Gee, thanks.

    But, and maybe it's all in the wording, what you have here is "X is a success. A solid success at its level. If that's higher than a lower level, you've succeeded at that level. And if you miss X, you may well have done Y, which is a useful thing. You just haven't done X."

    Somehow, that sounds so much more positive.

  10. I love this idea. I was looking at Supers! the other day, and it uses a similar number of d6s.

    In it 3d6 is the maximum that a normal human can have in a stat. So, using your guidelines, we could have the difficulties as:

    4 - Normal human
    7 - Skilled human
    13 - Peak human
    19 - Superhuman
    25 - High Superhuman

  11. Pavane: Yes, exactly. But what you say applies if and only if the rest of the system makes it true. Particularly, the results have to stand. The facts about the situation that are determined by the roll have to play a role in the next roll (and that sentence works better written than spoken).

  12. Narrowing focus is absolutely something I'm angling towards. :)


Note: Only a member of this blog may post a comment.