Addendum: Why "Roll under" Ability checks really are the best of checks

My recent meditations on roll under ability checks and rant against the D20 unified mechanic has generated a bit of commentary and further clarified my own thinking on the matter. This post is an addendum to my Using Ability Checks in B/X article, seeking to further explicate why the "Roll Under" ability check truly is the best of ability checks.

Earlier today, during my delvings into the blogosphere, I came across this box from Quarrel & Fable, a Fighting Fantasy spinoff:
First thing that struck me was how similar it was to my proposed resolution for Ability Checks. And secondly, it combined those thoughts with my memories of the old Fighting Fantasy [FF] gamebooks and set my mind spinning into that cross section and how much I always liked the elegant simplicity of the FF mechanic.
The best part about doing a post involving Fighting Fantasy is the chance to showcase some of the brilliant art in the gamebooks

Now, I've given reasons already in previous posts on why I favour the old ability check. Things like:

  • It gives mechanical significance to the ability score in and of itself, as opposed to just being a number used to derive other numbers - This is aesthetically pleasing, especially since the ability scores is the most prominent part of the character sheet
  • It's probably the most intuitively obvious mechanic in D&D - Give new players a B/X character sheet and sooner or later they will start to wonder if they couldn't be rolling against those scores somehow.
  • For B/X, it can do much the same as the X-in-6 roll where ability scores come into play.
  • It avoids DCs and fidgety modifiers
  • It's a test against one's ability rather than a test of the world's difficulty - And as such is a momentary mechanic that pops in and out rather than a passive constantly available one.
But let me tell you why it really is the best of checks when it comes to resolving tasks: 

It all boils down to the Fighting Fantasy gamebooks and how they made task resolution real easy and exciting.

Those who have played it will may have an inkling of where I am going with this already, but I will try and unpack the case for everyone.

The fighting fantasy series were single-player "create your own adventure" gamebooks. It had three stats - Skill, Stamina, Luck. You roll 1d6+6 for Skill and Luck and 2d6+12 for Stamina. Damage goes off your Stamina and Skill and Luck you use to resolve outcomes by rolling 2d6 to match or roll lower than your stat. Everytime you used luck, you lost a luck point (and thus will eventually run out of luck if you overuse it). Luck was mostly used for things that were hard to consciously impact and Skill for everything else, including combat. Very simply game engine really.

Your starting rolls meant your FF hero on average had a Skill score in the range of 7-12. Let's try and translate this to a D&D ability score, since FF uses a 2d6 bell curve vs D&Ds d20:
  • 7   = 12
  • 8   = 14
  • 9   = 17
  • 10 = 18
  • 11 = 19
  • 12 = 20.
Overall, this makes FF heroes quite capable and there will be more very capable FF adventurers than D&D ones rolling 3d6 straight. Fair enough perhaps, given that the FF hero is a solitary adventurer.

I can't speak for others but I generally re-rolled 1s and 6s for Skill and Luck, since I found 7 too punishing and 12 too boring. A skill of 9 seemed like a good baseline for having a decent chance of success to complete the book, whilst still being quite tough and needing to make optimal choices all the way. It meant succeeding on 83% of rolls. Which, at a glance, seems like an altogether high and snugly comfortable score to roll against. There were also never any modifiers to this roll that I can recall. Testing your Skill was Testing your Skill.

Here is the thing: Whenever I was asked to roll for Testing your Skill with a score of 9, I was still shitting metaphorical bricks. There was always a pregnant sense of drama in the air.

It turns out, failing 17% of the time is quite unsettling when you know that the consequence of failing a check can be quite punishing, or at any rate have significant consequences for where you go from here.

And this is where Fighting Fantasy succeeds brilliantly with such a simple mechanic. You weren't tasked with setting a difficulty for rolling when making checks. The baseline for when and how to roll was baked in - If it wasn't a proper challenge with consequences, you did not roll. Anything less than that was simply a decision. The act of making the roll already tells you how challenging this is.

Now of course, there are some variables to consider. The D&D hero will fail more often due to generally lower Ability scores compared to the FF hero's Skill. But he has companions to shoulder the burden with him. And where a failed roll in FF has quite fixed negative consequences ("If you fail, turn to 312"), the more open-ended world of D&D has some scope for creatively working around a failed check.

But the point remains - If you bake the right baseline into the check, you don't need to fiddle with modifiers to establish difficulty, etc. because that is implicit in the act of the roll in the first place. The very act of having to make a check should be more than enough for most players to suck in a bit of air in the dramatic pause that follows such a declaration. It's not quite a Saving Throw, but it lies just next to the Saving Throw's dramatic space. In other words, don't roll for shit that isn't fun and cool.

Make a Challenge Roll against Wisdom to not shit your pants.
Wait, make that a Saving Throw vs Paralysis.

It does requires a sharp conceptualisation of the mechanic, and the baseline of that mechanic, to really unfold. Here is my take, inspired in no small part by the Quarrel & Fable box above, that I call a "challenge roll" to emphasise the baseline:


As a general rule, how you describe your character's interaction with the world will determine the outcome of your actions. 
In some cases, where an action is especially challenging, for example due to high risk, time pressure, sheer difficulty or pure foolhardiness, the DM may ask you to make a CHALLENGE ROLL against one of your ability scores to determine the outcome.

A CHALLENGE ROLL is never trivial. Whilst a character with a high ability score may still succeed quite often in these challenges, a single failure can have severe consequences or turn an adventure in a dramatically different direction.

To succeed at a CHALLENGE ROLL, roll 1d20. If the result is equal to or less than the ability scored being challenged, you succeed.

If circumstances are especially favourable, you may be granted advantage on the roll, where you roll two d20s and only need one of them to be equal or less than your ability score for the CHALLENGE ROLL to succeed. 

If multiple advantages might apply, it is probably not challenging enough to roll for.

If the attempted action is deemed especially difficult, the DM may apply disadvantage to the roll, where you roll two d20s and both dice have to be equal or less than your ability score for the CHALLENGE ROLL to succeed. 

A cost roll works in much the same way as a CHALLENGE ROLL, where you roll 1d20 against the relevant ability score. 
But where a CHALLENGE ROLL determines if you succeed or not, a CHALLENGE COST ROLL assumes success, but a roll is still made to determine if there is a cost to the attempt. For example, injuring yourself on a difficult climb, losing valuable time or breaking the climbing tools. 
The DM will use the margin of failure/success to determine the outcome. Higher margins of success may also result in unexpectedly positive outcomes.
Why is this wording important? Let's compare this with how the d20 mechanic is first introduced back in the day, in the 3e PHB, emphasis mine:
A character can try to do anything you can imagine, just as long as it fits the scene the DM describes. Depending on the situation, your character might want to listen at a door, search an area, bargain with
a shopkeeper, talk to an ally, jump across a pit, move, use an item, or attack an opponent.
Characters accomplish tasks by making skill checks, ability checks, or attack rolls, using the core mechanic.
So here you have  the baseline, baked in from the start - roll for fucking anything. The 3e DMG elaborates with more bullshit:
The whole game can be boiled down to the characters trying to accomplish various tasks, the DM determining how difficult those tasks are to accomplish, and the dice determining success or failure.
While combat and spellcasting have their own rules for how difficult tasks are, skill checks and ability checks handle just about everything else.
So not just roll for anything, but actually roll for everything. Fuck that. And fuck it twice on game nights.

"You actually planning on rolling dice for that, bub?"

By 5e, designers had wised up a bit:
The DM calls for an ability check when a character or monster attempts an action (other than an attack) that has a chance of failure. When the outcome is uncertain, the dice determine the results.
A bit better than being explicitly told to roll for everything, but still a bit shit really. The 5e DMG elaborates:
When a player wants to do something, it's often appropriate to let the attempt succeed without a roll or a reference to the character's ability scores. For example, a character doesn't normally need to make a Dexterity check to walk across an empty room or a Charisma check to order a mug of ale. Only call for a roll if there is a meaningful consequence for failure.
When deciding whether to use a roll, ask yourself two questions:
Is a task so easy and so free of conflict and stress that there should be no chance of failure?
Is a task so inappropriate or impossible- such as hitting the moon with an arrow-that it can't work?
If the answer to both of these questions is no, some kind of roll is appropriate.
Oof. For a moment, it looked like going somewhere when it says "meaningful consequence." But apparently, this is just shoddy writing for "chance of failure" since that is how it is elaborated. 
The notion of meaningful consequence has no follow through elsewhere in the text either, where we are given DCs for "Easy" and "Very Easy" tasks (though, despite giving the DC, it does say that most times very easy tasks require no roll).

This is starting to sound a bit like my previous rant against DCs, but I am actually going somewhere slightly different with this. This is not about widget-making ruining roleplaying, but about baseline. 
That is to say, when is making a roll going to be a fun and cool thing to do? What is the threshold for making a roll and what does that mean for the kind of roll you make?
We interrupt this post to show you what a regular tavern is like in FF

Using the Challenge Roll proposed above sets a baseline - The difficulty is already set by the act of rolling the die. Now it is a matter of meeting the challenge with a test of your ability.
The mechanic doesn't enter the picture until a situation arises where an action is challenging and has a potential cost. And then it disappears again. It's basically like pushing a Drama Button. You know shit can go down when making a Challenge Roll.

It's the same reason why the B/X rule of thumb for trying stuff is basically 2in6. If it's not dramatic somehow, don't bother.

There is another aspect which is a mostly aesthetic, and disputable, point, but I also think there is something something somehow more dramatic about testing your own ability in a dangerous situation, than testing a situation's difficulty with your ability. It's a difference between watching Indiana Jones walking an old tightrope and wondering "can he do it?" vs "how tricky is this tightrope?"

It's basically about having a game that allows for dice rolls for players to test their abilities in challenging and dangerous circumstances vs a game that allows for rolling dice for resolving whatever you want.

In the former game, you don't need a lot of modifiers. It's baked into the baseline of the roll. The dice are there for dicey situations. In the latter, you do, since it needs to resolve a much wider range of situations than the former. The dice are there for any given situation.

The roll-under ability check gives you a mechanic with just the right level of granularity for a dramatic baseline, and as such has little of the ambiguity of other mechanics in setting difficulty and applying modifiers. It uses the most prominent part of your character sheet to good effect; gets out of the way when not needed; is intuitively obvious and mathematically sound in terms of success-to-failure ratios in dicey situations. It is, all things told, simply the best of checks.


  1. Big agree. I actually have my own drafts of these same few blog posts. Moving away from roll under put the math onto the player instead of the system. It's like having to do your own taxes even though the IRS/CRA already knows the answer.

    You can even include variable difficulty or costs into it. Like Whitehack has roll under for the attack roll and AC can be 0 to 6. So you have to roll under your attack score but over your opponent's AC. Such a number could also be used to indicate success at a cost.

    I'm also annoyed by requiring the poly dice but not really using them for anything except weapon damage. I've been toying around with the idea of each attribute having a different die assigned to it and having to roll 3 or higher. Mathematically it functions similar to roll under but it seems even more fun because of all the fun shapes!

    Don't even get me started on what the stats are called and what they mean.

    1. Agree with the post and with Lawson, too.
      @Lawson, check out the systems for Savage Worlds and for Mazes, both games rating stats in polydice and resolving rolls in different ways. In my view, Savage Worlds piles on way too much mechanical stuff, and it could have been to my taste with a target number of 4. Anyway, check these two out.

  2. Totally agree, and this is basically the approach I use. For me, B/X (or BECMI - there isn't really a meaningful distinction) - provides precisely the right level of crunch.

  3. I like your point especially about testing the character not the environment. I wonder what your thoughts might be on the 'blackjack' approach to opposed rolls that can be used for combat for example. As in: roll under, roll high. Highest roll wins giving the person with the higher skill a greater range of winning rolls.

    1. I have my doubts about opposed rolls in general, tbh.
      I suspect they are better resolved as situation rolls, ie a roll for the situation that can take modifiers from both 'sides' rather than than dual rolls.
      Bit like the D&D attack roll really, which takes modifiers from attacker, defender and conditions.

      Blackjack sounds like a fun way, provided it is a single-die roll (would be utterly screwed in GURPS where the 3d6 bell curve always hews towards the middle).

      Atm, I am leaning towards something a bit like the reaction roll table in TSR-D&D, with the type of dice being subject to the size of the modifiers employed. But basically, two dice rolled to have a bit of a bell curve towards the middle, gradated outcomes and small modifiers remain significant.

      I don't know, my thoughts on opposed rolls are still digesting. But I am in general leaning a lot more towards situation rolls for a lot of stuff that tends to be resolved as group efforts or contests these days.

  4. TL;DR: exactly what Into the Odd does, for exactly the same reasons :-)

  5. I have not found roll-under to be attracative, but you make some compelling points. A character having a 75% chance on every Str roll they make (15 stat) really turns me off... until I frame it in the perspective of stakes that are akin to a FF encounter. Very interesting stuff.

    I still don't like how roll-under in D&D clones shines a spotlight on the inequality of PC attributes, but there are hacks for that if I were determined.

    I think the big question in my mind is whether an OSR game can duplicate the FF tension. In FF, you're rolling 1 die to solve the entire enounter. That whole thing goes North or South based on a single roll, and that's just not how a TTRPG plays.... although.... I'm getting some inkling of an idea of a tug-of-war mechanic where dice rolled through an encounter push or pull it toward one of two versions of whatever outcome occurs: the dire or the victorious version. It might work in a PbtA game, but I don't see how in a more traditional game. It seems a bit gamey and arbitrary.

    1. Ugh, I meant to say 16 stat, but I guess my mind was on the fact that rolling a 15 or lower is 75%.

    2. I am happy to see that the FF is doing similar stuff to your brain that it do to mine. :)

      Yes, stakes really is important to the frame. I am not leaning towards PbtA stuff here, but there are simple ways to raise the stakes, such as eliminating trying again, or other characters trying as well, by simply declaring that the initial die roll applies to all successive efforts (the roll is, *cough cough*, the Difficulty Class of the task, so to speak. *gags*) and making the Cost of a failed roll transparent to the players.

      FF frontloads the cost of a failed ability roll. "Turn to 263 and immediately suffer the consequences of your shitty roll".

      Similarly, I would say the DM should visualise a tangible cost to a failed roll before asking for an ability check to make sure the stakes are present. Soon enough, players should gain an awareness that ability checks means they are in turbulent waters here.

      As JB from Blackrazor puts it, there is nothing worse than a skill check merely to "find out you suck at this today".

  6. There is a variant of the Whitehack roll under your ability but over the difficulty mechanic: use a challenge die for the opponent/target. The number to exceed with your d20 roll is what is rolled on the challenge die: 0/d2/d4/d6/d8/d10/d12. This is simple but exciting, as you never know exactly how hard the roll is going to be. You can make character classes by adjusting how quickly their challenge die goes up for each ability. As a baseline, the challenge die should equal the character level or monster hit dice. This die could even be rolled with advantage or disadvantage, as the situation calls for. The players or the GM could roll the challenge die for monsters/obstacles, as your table prefers.

    To be fair to the players and the GM, ability scores in a roll-under game system should be rolled as 2d3+6, giving a bell curve from 8 to 12. This is the equivalent of modifiers from -2 to +2. Nobody is great right off the bat, but nobody sucks at anything either, and every character is different enough to be interesting.

    1. This is interesting, I do something similar in my game but without the blackjack mechanic. Checks are made rolling 1d12 + d4/6/8etc... The d12 is always rolled by the player and the "difficulty dice" is rolled either by player or me based on the situation.

      If they try to break a door open I say "d8 on STR", if they fail I like them knowing exactly "how much" ( since the Char "feels" how much the door was sturdy )

      If they search for traps I say "roll on WIS", but I roll the variabile dice without telling them what dice is it. In this way they still have an idea of what to expect: in a "3.5" system if they roll a DC of 23 they know no easy spottable traps are in the room. In my way if they have WIS 15 and roll a 12 they are just pretty sure there should be no evident traps in the room. Also if they have WIS 10 they are often clueless, as it should be. I consider the fact that in this way some checks are useless a feature, not a bug; if you are very strong and the door is made of wood you should just kick it out of the way. In the same way I like having unsurmuntable tasks which must be circumvented by taking the less obvious approach. But hey, we are all engineers in my group so maybe we're missing the fun in sometimes just tossing a coin and killing the dragon with a crit ! :)


Post a Comment

Popular posts from this blog

Fantasy Map Review IV: Forgotten Realms

Fantasy Map Review II: Greyhawk

Fantasy Map Review III: Dragonlance

Review: Five Torches Deep

Comparison: Five Torches Deep vs Into the Unknown