What do you consider 'medium difficulty' in % ?

edited January 2015 in Game Design Help
I've analyzed various RPGs what "medium difficulty" for a success rate could mean in percent (%) and came up with values ranging from 39,6% to 85%; with most RPGs around 50-65%.

So my subjective take on this would be for a success rate:
90% very easy
80% easy
60% MEDIUM
40% difficult
25% very difficult
(interestingly, my subjective feeling also correlates a bit with 2d6 and 4dF distribution http://anydice.com/program/5085)

What is your personal opinion?
Is there some there some analytical or empirical assessment on this?

Comments

  • I don't find the term "medium" to be at all useful here.

    Generally, it just means "in the middle, between hard and easy". That's not useful, except to tell us that it's easier than hard but harder than easy.

    I'd rather look at what's at stake, and how much a person is willing to risk. Making an attack against a monster in D&D could have a 30-40% chance of success and still be considered "routine": after all, the player and character lose nothing (theoretically speaking, anyway) by making an attack, and will probably have many, many opportunities to try again.

    On the other hand, if I'm the pilot of our plane or spacecraft, and I have to make a Piloting roll to land the craft safely, I'd consider a 90% chance of success absolutely brutal odds. (Would you get in a car with a driver who crashes the vehicle every tenth time he goes out on the road?)
  • There's no easy answer to this question, partly because "medium for whom?" and partly because if the designer is doing their job, the difficulty-vs-ease of character actions is part and parcel with the resolution mechanics. For example, a cinematic game featuring "heroic" characters will have a different "medium difficulty" than a modern-world game featuring "normal" humans and high verisimilitude. It depends on the flavor you're going for.

  • My subjective is based on the 3d6 bell curve, since D&D and first exposure and blah blah. They straight up had the bell plotted out in the book! So it would be more like 7% very easy, 18% easy, 50% medium, 68% difficult, 93% very difficult. But as mentioned above it's all relative to the setting, the assumptions, how much your players hate losing and whether or not someone bribed you before starting.
  • All true. I still don't see the use of the term "medium". Who's it useful for?

    I prefer using a scale where each step has a meaning of some sort, if you must. For instance:

    Easy - 90%
    Challenging - 65%
    Very Difficult - 35%
    Insurmountable - 10%

    Or whatever.

    I find that more useful, as a player and as a GM. "Medium" is weaksauce, as a term.
  • Oh, that's your point! Then I totally agree. In the DayTrippers core rules I opted for:
    1 = no-brainer
    2 = easy
    3 = challenging
    4 = difficult
    5 = hard
    6 = very hard
    7 = unlikely
    8 = ridiculous
    9 = absurd
    10 = insane

    I almost used "sheer folly" in homage to Iron Crown, but hey, I've ripped off Coleman Charlton enough already :-)

  • "Difficult" and "hard" are synonymous.
  • edited January 2015
    Agreed with Paul.

    Without further clarification, "Medium" references two completely separate scales:
    1) The range of all things one might ever try to do.
    2) The range of all things one might roll for to resolve dramatic uncertainty.

    If you picked the latter scale, then I'd have to ask whether we're rolling for:
    A) Momentary advantage, where no one fears small chances of failure.
    B) Big and irrevocable change, where tiny chances of failure are still a big deal.

    If you picked 2A and mandated 5 Difficulty Levels, then I guess I'd go with these:
    How Difficult? - odds of success
    Not very - 75%
    Fairly - 57%
    Medium - 40%
    Very - 27%
    Incredibly - 15%
    That seems like a dramatically satisfying set of possibilities for not-super-consequential die rolling. The odds of failure and success never get high enough to seem certain, nor low enough to seem not worth rolling for.

    With fewer constraints than (e.g.) "rolls for drama and small stakes, 5 Difficulty Levels", I can't even answer the question.
  • "Difficult" and "hard" are synonymous.
    I agree semantically, but not psychologically. :-)

  • edited January 2015
    If the question is about real-life intuition, well, I think "medium difficulty" almost doesn't exist in real life. At any given moment, with most things, we can either do them or we can't.

    Here are the four things I can come up with that I might ever call "medium difficulty" in real life:
    1) A skill that looks like it will take me personally a "medium" amount of time to learn.
    2) A completely new thing I've never tried to do, for which I estimate a "medium" likelihood that I'd be able to do it. An actual attempt will then resolve this guess into a reality of "yes I can" or "no I can't", so the "difficulty" is purely imaginary.
    3) A task that is doable by a "medium" number of people in the context in question.
    4) A "medium" balance of factors hindering and aiding a task.

    Note that (3) and (4) still probably leave any individual's odds of success at that task near 0% or 100%.

    If we're interested in looking only at the tasks people perform at the very edge of their trained ability (I can always throw a frisbee 73 yards unless there's a fluke slip, and I can never throw it 87 yards unless there's a fluke wind gust, but 80 yards is a maybe), then, subjectively, I think the sensible labels would be:

    Very difficult - 80% success
    Incredibly difficult - 40% success
    Nearly impossible - 5% success

    If it was anything less than "very difficult", and you've put the time in, you could just do it, right? It seems to me that that's how we tend to think.
  • edited January 2015
    How about "challenging" instead of "medium"?
    Is there some there some analytical or empirical assessment on this?
    I don't know where who read it or where, but a major forum (ENWorld?) had a poll that asked how often people thought they wanted to succeed and a majority said that they were satisfied if they succeeded 65 % of the times.

    [edit] Here is the only one I could find. I thought it had a lot more voters than 39.
  • 1/3 (almost 67%) is my Platonic ideal.
    I've loved the 4dF idea since I first saw Fudge (even though I no longer like the rest of the game).
    It uses 60% which I am very satisfied with. So easy to see, without any numbers.
    65% is a mod of +7 in D&D which you reach after a few levels if you're proficient.

    Incidentally, this is why dice pool games bug me. Burning Wheel I'm staring angrily at you.

    In the end I realized that dice mechanics is the bikeshed color of roleplaying games. Prep is the foundation. The way you roll dice is like the steering wheel -- it needs to be comfortable, yes, but it's something you can change. But things like D&Ds mapping or AWs MC moves or Wicked Archipelago's oracles -- they can make a game.

    In the nineties, this dice stuff was all we argued about because it was all we understood. We even started naming the systems after which dice it used the most. D20, d6, diceless… many of the games were pretty much the same under the hood.
  • Prep is the foundation. The way you roll dice is like the steering wheel -- it needs to be comfortable, yes, but it's something you can change. But things like D&Ds mapping or AWs MC moves or Wicked Archipelago's oracles -- they can make a game.

    In the nineties, this dice stuff was all we argued about because it was all we understood. We even started naming the systems after which dice it used the most. D20, d6, diceless… many of the games were pretty much the same under the hood.
    Yes! This! +1

    And that's why this thread can get nowhere, in the abstract. We need to know about the actual game(s). We need to know what players are rolling dice for, at the very least.
  • What is "medium difficulty", the replies ask? I read it as "What's a good percent of success to use most often, as a baseline, for an interesting chance, in a game?"
  • In the nineties, this dice stuff was all we argued about because it was all we understood. We even started naming the systems after which dice it used the most. D20, d6, diceless… many of the games were pretty much the same under the hood.
    +1

    Most of them still are, no matter what creative agenda you play with.
  • 60-65%

    Misses usually stick in the player's psyche a bit more than hits, so weighing the odds a bit towards hits makes things feel "even".
  • The best free throw shooter in NCAA Division I basketball is Riley Grabau, who hits 94.2% of his shots. #50 hits 85.5%.

    The best record for shots from the 3-point line is 55.5%.
  • edited January 2015
    Paul_T: thanks for pointing the "what's at stake". Makes a huge difference, indeed.

    Paul_T/Asif/Rickard: scale/ladder is good, I had i.e. Fate scale in mind with "average" difficulty. Probably, "moderately challenging yet doable" would be a good description for my personal flavor of "medium". I also like 2097's definition:
    "What's a good percent of success to use most often, as a baseline, for an interesting chance, in a game?"

    AsIf: to "medium for whom?" - well, for you personally. i.e. you could say "for pulp setting, I'd consider X% medium; for cthuloid setting Y%.".

    Veav: thanks. I'm counting the percentage the other way round, and I get the idea.

    David_Berg: thanks for your precise answer. I had something like "2A" in mind, indeed.

    Rickard: thanks for the link. Also for me, around 60/65% seems to be the sweet spot.

    2097: Yes, 2d6 and 4dF distributions resonate also pretty well with me.

    arscott: yep, same here. I also feel exactly like this: "Misses usually stick in the player's psyche a bit more than hits, so weighing the odds a bit towards hits makes things feel "even". "

    2097/Rafu/Rickard:
    To me the three Cortex Plus systems (Dramatic/Heroic/Action Roleplaying) are a very good example of how dice mechanics _can_ make a difference and support the setting. Of course, for most settings/games, the underlying dice mechanics are probably exchangeable.
    I also agree that it's more important to have a good "steering wheel" and not the "perfect dice scale".
    Still, my main problem with dice pools is that probability distribution is often arbitrary and doesn't scale well for various dice pool mechanics (no. of dice, highest/lowest of, rerolls...). While they are kind of fun, I'm often left with a feeling that I might as well play dice-less freeform, kind of.

    Adam_Dray: We don't play baseball in Europe, sorry I can't relate to that ;-)
    Shooting a basket from a three-point-line would be probably something I'd consider "difficult" for a pro, a two-pointer something of a "medium" difficulty. Sort of.

    Overall, to give typical situation would be i.e. a "challenging yet not life threatening task/encounter in a D&D setting". It's interesting to see that some also perceive a certain "sweet spot" around 60-65% (success) as a base difficulty (and for some this doesn't apply this way).
    As I've said I've gone through some of my favorite and some popular RPGs. Most of them implicitly or explicitly state an average or medium difficulty, I've calculated them as:
    50% 61,73% 64,88% 85% (72,22%) 52,5% 55% 50% 62,5% 39,6%
  • What's that low one? WHFRP?
    Ps Adam did say basketball, which is popular in Europe, not baseball.
  • edited January 2015
    Based on the discussion, here is a revised approach:
    image

    I know this is very subjective and not comprehensive. Some may relate to it, some may not.
    I would consider i.e.:
    Low Danger / Moderate Difficulty: Sneak out inn without waking up host
    Medium Danger / Moderate Difficulty: Win fist fight in a bar
    High Danger / Moderate Difficulty: Infiltrate enemy camp without getting caught
    Lethal Danger / Moderate Difficulty: Sneak out dungeon without waking up dragon


    And yes - though not on par with the NBA - we do play basketball in Europe ;)

    39,6% is not really a "crunchy" system, it's Risus (based on 2.5; thus average between 2d6 [16,7%] and 3d6 [62.5%] for >=10 difficulty)
  • Oooo, I like that task/danger double axis! (For a D&D-like thing, that is).
  • edited January 2015
    Relevant comments first:

    How often a task is attempted - or "how many retries are possible" - is also an important feature.

    The Danger/Task Challenge table is a good idea, and lends a little more depth to the question of "how do I set the probability of success"? I'm not sure I like the actual numbers on it, but the concept is very much heading in the right direction. Good stuff.

    As an irrelevant aside, I find your commentary on dice pools bizarre and surprising. If the dice pool isn't presenting a good probability curve for your game... your game is very poorly designed, in my opinion. I'm sure there are some games like this out there, but... yuck. It's certainly possible to design a very nice dice pool mechanic (and, at least for me, that's the only reason to use one - to get odds you can't easily get with a simple die roll). [One way I like to use this is to make up ways in which the odds of success grow or shrink at different rates than the odds of failure. For example, there might be a way you can increase your chances of success while simultaneously making a disastrous failure more likely.]

    As another irrelevant aside, it's funny to see the term "Average" from FATE mentioned here. They took the adjective scale from Fudge and added the term "Average" to it. I always felt that this was an error, for a variety of reasons (not just aesthetically, but also with how it messes with the size of the steps in the scale). I much prefer the original Fudge scale, without the problematic "Average" term.
  • edited January 2015
    Looks sensible from a meta-perspective. Historically, most RPGs have resisted that sort of "rules, not fiction" transparency, but if you don't mind bucking that trend, cool.

    (In other words, "My odds of success go up because it's more dangerous? What kind of fantasy world is this?")
  • 60/65% does seem reasonable, but I like games with the ability to manipulate probabilities either pre-rolling or post facto, by spending Fate points for example, or with buffs like Bless, etc.
  • I'm ok with pre-rolling but don't like it post rolling. "I dropped it! No, wait, I have it!"
  • Relevant comments first:

    How often a task is attempted - or "how many retries are possible" - is also an important feature.
    True! Makes me think of a table like @BeePeeGee's with this and "danger" as the two axis, and no mention of the "fictional difficulty" of the task. :)
  • edited January 2015
    Yes, "Let it ride" is yet another proof of Luke Crane's genius. Your roll represents all of your tries.

    Edit: Wait, I just remembered that I was slagging BW earlier in this thread. Let's just say it's both genius and not genius at the same time. :p
  • By the way, success rates for typical AW characters range from ~60% (0) to ~84% (+2).
  • When is it ~60%? Maybe I missed something and Crane can be redeemed. I was under the impression that it pretty much flipflopped from 20%- to 80%+
  • Ah, thanks. AW, BW, CW and DW all blend together in my mind sometimes. Yeah, 2d6 maps pretty well to 4dF.
  • edited January 2015
    Reflecting the discussion, I think an important motivation for me is having baselines for GMless play (sort of like Mythic GM Emulator, simplified).

    Paul_T/Rafu: The number of retries would in my view affect the Danger Level; i.e. resolving a conflict in one throw is more dangerous than running a conflict step-by-step.
    2097: and yes, I love "Let it Ride" (I used it often, prior to knowing the BW mechanic)

    David_Berg:
    >"My odds of success go up because it's more dangerous? What kind of fantasy world is this?"
    To me it makes sense from (meta) GM perspective (of course not in-game). When setting up challenges for players in lethal siutations, it certainly makes sense to scale down the failure probability (i.e. a magic item, an NPC ally).
    I know not everybody likes this "fictional difficulty" perspective, it works for me.

    Rafu: thanks :)

    Deliverator: Ok, my goal was to just have a [baseline as a starting point. I also like tilting the odds with Fate Points etc.

    *: on AW: Isn't this the probability?
    2d6 + Stat (average= 5/3) >= 7 (min) or >= 10 (between weak/hard move)
    ==> 72.22% (weak hit) or 27.78% (strong hit)

    Paul_T: As for dice pool, it is certainly possible to design robust mechanics and there are systems that work well. However, I've analyzed and found some quite inconsistent dice pool mechanics, both indie and commercial.
    Personally I feel more often than not, dice pools "feel good", yet give a false sense of control.
    [Most recent example I remember (FU): For bonus, you roll several d6 and pick highest; for penalty you roll several d6 and pick lowest. Sounds very elegant, yet the probability distribution for bonus/penalty is quite different and contrary to intuitive estimation.]

    Paul_T: On Fudge vs Fate ladder: I'm not so happy with the differentiation Fair/Average in Fate, they "feel" the same to me and Fair certainly doesn't feel +2.
    I prefer the Fate interpretation of Medicore as 0, since to me it really means "barely made it" and deosn't feel like -1.
    Then again, the Fudge/Fate ladders(s) are pretty robust and I guess it doesn't matter too much as long as difficulty scale and skill ratings match.
  • edited January 2015
    Reflecting the discussion, I think an important motivation for me is having baselines for GMless play (sort of like Mythic GM Emulator, simplified).
    That makes your goals here much clearer, thanks!

    Paul_T/Rafu: The number of retries would in my view affect the Danger Level; i.e. resolving a conflict in one throw is more dangerous than running a conflict step-by-step.
    Are you sure? Why?

    (This seems like an interesting design question to me. If it isn't to you, ignore it!)

    *: on AW: Isn't this the probability?
    2d6 + Stat (average= 5/3) >= 7 (min) or >= 10 (between weak/hard move)
    ==> 72.22% (weak hit) or 27.78% (strong hit)
    Sure! That's right in the middle of the range I gave.

    More importantly, though, this presents an interesting question:

    How do you estimate "success" in a rules system which has different grades or types of outcome? Is a "strong hit" a "success", or is a "weak hit" sufficient?

    This becomes more complicated in systems which require you to accept a setback in order to achieve a goal, etc, etc.

    Not necessarily an obvious question...

    Paul_T: As for dice pool, it is certainly possible to design robust mechanics and there are systems that work well. However, I've analyzed and found some quite inconsistent dice pool mechanics, both indie and commercial.
    Personally I feel more often than not, dice pools "feel good", yet give a false sense of control.
    [Most recent example I remember (FU): For bonus, you roll several d6 and pick highest; for penalty you roll several d6 and pick lowest. Sounds very elegant, yet the probability distribution for bonus/penalty is quite different and contrary to intuitive estimation.]
    Interesting. Can you explain what you mean about FU? How does the distribution not match intuitive estimation? Are you talking about a player's intuitive estimation (i.e. estimating the odds of a success mathematically - you'd think roll X dice gives you good odds, but math reveals that it doesn't) or a character's-perspective intuitive estimation ("I'm a good skier, so I should be able to make this!" -> Whereas, the way it actually works out, a "good" skier fails most of the time)?

    I agree with you on the Fudge/Fate issue. My main issue with "Average" (aside from it being a meaningless term, for the kinds of things they're using the ladder to rate in Fate) is that it "expands" the middle of the range, making it even harder for low-rating characters to achieve results or compete with higher-skill characters. It is also means higher-skill characters suffer "poor" results less often, substituting "Average" results more often. Finally, it means even more results fall into that ambiguous "Mediocre/Average/Fair" area for everyone, which isn't really descriptively helpful.

    Doesn't seem like the best design choice, to me. I'm guessing they needed an extra step in the ladder to make some of the math work out, so they threw that one in. Would have been more fun to add it at the top end, I think, but that's just me!

    (Of course, it works for the game as written anyway. It's not a huge point, just something that stuck out to me.)
  • edited January 2015
    N.B. for whom it may concern: Although it's tempting to consider the 7-9 range in AW as being the most likely due to our well-drilled acknowledgment of the 7 as the apex of the 2-12 bell curve, that range spans only 3 numbers, which narrows its likelihood to exactly the same as the failure range which spans 5 numbers (note this is without modifiers):
    2-6     41.67%
    7-9 41.67%
    10-12 16.67%
  • That's true. A 7-9 is the most common outcome *only* for a character with a stat of +1 (at 0, it is tied with a miss, and at +2 it is tied with a 10+).

    +1 is the average stat score for an AW character, but if we take an average of the odds for any given roll assuming an even spread of scores from -2 to +3, or the average of the odds for any given roll assuming an even spread of scores from -1 to +2, the cumulative odds for failure and a 7-9 are pretty similar (36% for a miss, compared to 40% for a 7-9).

    However, given this conversation and character success, it's interesting to note how we parse these results *emotionally*: yes, often the odds of a failure are as common or more common than the odds of a 7-9, but in our mind we lump the 7-9 hit in with the 10+, since both are character successes. That makes misses seem rare in comparison. (Even for the character with a stat of 0 - which doesn't get rolled as often as stats of +1, +2, and even +3, in my experience - the odds of a failure, at 42%, are less than that of any success at all, which is in the 58-59% range.)

Sign In or Register to comment.