d100 Hate

There’s a debate going on somewhere about whether the mechanic of roll under percentile (as used by BRP, RuneQuest, CoC, etc.) is “disdained”. And on this rare occasion the forum debate is enough to inspire a post…

To clarify roll-under: you get your dice and try to roll equal to or less than your skill. This is compared to roll-and-add where you roll dice, add a modifier and try to hit a target number.

Some of the cited detractions of roll-under percentile are:

  • No bell curve (inferior to 3d6, which is also a roll-under for GURPS)
  • No range of interesting results (binary pass/fail)
  • Confusing bonus, e.g. is +20% a benefit (applied to skill before the roll) or a penalty (applied to the roll itself)?
  • Subtraction is harder than addition
  • Roll under a modified value is less intuitive than roll and add

I’m not going to talk about bell curves. Bell curves are useful in considering a range of performance over a group of people; applying them to individual skill rolls makes sense if you don’t know who is going to be performing the task (maybe your PC is a gestalt?), otherwise it’s a matter of taste, and less transparent than a linear roll.

The other items are worth considering, because they all represent some kind of cognitive load on the player. This includes the “fix” for the binary pass/fail in d% which I and plenty of other people house rule as

Pass = full %

“Success” = half %

“Expert success” = one fifth %

“Critical” = one twentieth %

Of course you need to attach meanings to Success, Critical, etc. (and add fails, fumbles). But, as long as the player knows the order of desirable outcome, everything’s fine.

Really there’s two objections:

  1. Whether d% is aesthetically pleasing to the players (and I think that has as much to do with percentiles, as contending with long lists of skills; see below)
  2. Whether d% actually causes more of a cognitive load than other systems.


Can d% really be aesthetically pleasing? A pleasing RPG design is going to be one where the player can visualise what they are going to do in the game.

  • At the iNtuitive end of the scale, there’s an appreciation for the character as a rounded individual and their sense of place within the game world. This is helped by pretty character sheets and meaningful names. This is probably not helped by long lists of narrowly-defined skills, particularly if they have obscure names.
  • At the Sensing end of the scale, there’s the effect the character can have on the game world, and therefore what they should spend most of their time doing. Having a system where skill aptitude is declared absolutely is beneficial here, and you don’t get more absolute than a percentage.

You can overcome the long lists issue by chunking the information. ORE games like Reign do this by clustering skills under headings. RuneQuest 3 did this as well. If you do this you may please your iNtuitive types but the Sensing types would rather have an alphabetical list, thanks.1

There’s one other aesthetic consideration: roll low may give less of a sense of achievement than roll high for a success. I think that’s down to the norms of the group. People learn that they want to roll low fairly quickly.

Cognitive Load

First, consider the granularity of success; if you apply the fix above to the % roll, you require the player to make some mental division to establish whether they’ve achieved an extraordinary success. A linear die roll with adds makes things easier.

Well, maybe not. On one side you have the effort of division, and on the other side you have the need to keep several factors in working memory. Typically roll and add systems involve adding two or more numbers together before the roll is made (e.g. Unisystem, Interlok), and once you factor in additional bonuses, time to scan the character sheet for the right stats, and the GM’s time to think of a target number, you probably have around the same amount of effort.

It may even be more effort. In a d% system where rolling under 1/5 of the skill is important, that calculation will be done after the roll and only when the player thinks there’s a chance they qualify.

Roll and Add Roll Under %
  1. GM declares stat and/or skill, may declare difficulty
  2. Players and GM negotiate bonus/penalty, apply as needed
  3. Player retrieves information from character sheet and rolls dice
  4. Player sums up result.
  5. GM compares result with their difficulty.
  6. Outcome is declared.
  1. GM declares skill for roll
  2. Players and GM negotiate bonus or penalty, apply as needed
  3. Player retrieves information (single %) and rolls
  4. If the roll is significant (e.g. near a point where under-1/5-skill is a consideration) player then makes calculation. Otherwise, dice stand.
  5. GM compares result with their difficulty.
  6. Declare outcome.

Let’s ignore steps 1 and 2 for now, and assume it’s no more of an effort for the GM to declare “roll $ATTRIBUTE plus $SKILL” than it is to declare “roll SKILL%”. Let’s assume any negotiation step will take the same amount of time (mostly dominated by players arguing if they’re in cover or not, etc.)

Let’s also ignore step 3, and assume it takes the same amount of time for players to scan their sheets and get the information they need to make the test. Not always true when the player needs to retrieve 2 pieces of information vs. just one, but we’ll forget that now.

The real load comes at steps 4 and 5. In step 4, on the left the player has to add several numbers together (easy, but harder with more numbers). On the right, the player has to divide (harder), but not always; when they do roll, it’s because they’re motivated to make the calculation because they might get a better result out of it.

OK, what about the load on the GM? On the right, the GM has to contend with a limited number of outcomes (“Fail”, “Success”, “Critical”) and the meanings of those outcomes will be established in the norms of the group and in the rules.

On the right, the GM has to go through an interpretive step (comparing how much better the player rolled than the number the GM has in her head; is it one point better, 3 points better, etc.). That’s a load of arithmetic you’ve heaped on the GM. The only way the GM can get away from this burden is to accept a binary pass/fail as the only outcome, something which d% is criticised for.

Arguably the d% system presents less of a burden than additive systems, even simple ones. Why then do some people react badly to d%? I propose that the strength of d% is exactly the reason some people can’t stand it: some players don’t want to know how likely an outcome is. If there’s only a 30% chance of success, why make the attempt? But if there’s an unknown but slender chance, choosing to blindly roll is seen as heroic, rather than tactical. Furthermore the promise of some open-ended rolling is alluring and optimistic, whereas the chance of rolling 5% or less is pessimistic.

So that’s it, then. This isn’t to do with cognitive load at all, it all swings back to aesthetics. Gamers who hate d% are incurable optimists and fantasists, wilfully ignoring realities and probabilities, and gamers who embrace d% are realists and pessimists. It would explain why d% works so well for Cthulhu.

  1. INtuitive types probably don’t like the Big List of Skills approach (myself included). Each skill is of very precise function, and there’s not much room to interpret experience from one skill benefitting another skill, and for some of us that doesn’t make much sense. But that’s not a problem, that’s the game. If you don’t like that play a different game with broad-brush characters like Everway or Over the Edge or Summerland or Ghostbusters.

4 thoughts on “d100 Hate

  1. I like % as I find it nice and clear cut and can specify in advance what the player is looking for.

    On 3d6 (which I appreciate you have left here), I once had a GM (who doesn’t read you) who chose a 3d6 system and their default success was “roll 12 or greater”. They were then very puzzled that said successes only turned up about 1/3 of the time (a quick internet check says 37.5%) rather than over half the time. A good reason for people to steer clear unless they really love 3d6!

    1. I’ve nothing against multiples of d6. In fact, I like d6 more than other dice (and spots! Proper d6 have spots).

      Also the bell curve has some advantage, e.g. in FATE/FUDGE where the mean roll on 4dF is zero, there’s a strong bias to get results only one or two steps away from the mean. For FATE this then drives the whole fate point economy (spend a point to re-roll, etc.)

  2. First I must say I biased towards bellcurves for purely math/statistical reason. From the game point of view I find d100 easier on the players (at least before they think to much). In order to granulate and have more interesting results from one roll we adapted two things from our WFRP game. The first is that double rolls (11, 22 etc) is critical rolls, under skill equals critical success and over critical failure, simple and easy to remember. The second is that the ten die shows “level of success”, when needed. Levels of fail I ditched as unecessary, Crit Failure and just Fail I was enough granularity in that end.
    I have also played around with difficulty levels by modifyng the skill number, but haven’t really found a good way to present it in a smooth and easy to use way yet. Probably going to drop that in favor of numeric bonus (-10% etc).

    1. Hi there. Yes, both the doubles-for-crits and creative use of the tens die (or indeed the 1s die, as Kromosome does) are good innovations that don’t require complicated post-processing, and it works neatly with simple +/- bonuses or penalties too. Levels of Success could be a bit more random though, given the range of numbers.

Comments are closed.