... now with 35% more arrogance!

Friday, January 9, 2015

Where Unified Rolls Go Wrong

I haven't had access to my computer for about 9 months, which has meant that I've had to put a lot of planned creative work on hold. It's pretty damned hard typing long articles on a 7-inch tablet, and sophisticated tables or other fancy layout is pretty much out of the question. But my computer is out of storage now, and I should have it set up and connected to the internet soon, so expect some PDFs of useful stuff before too long.

In the meantime, I wanted to rant a bit about the idea of unified game mechanics. Whether you use one mechanic to determine two different things, or a different mechanic for each, is pretty much a matter of taste, but there are still some people who insist that a unified core mechanic for all needs is the best way to go. They miss out on some good opportunities this way, since some mechanics are better for yes-no, hit-miss questions and others are better for multivalue results, or provide an easier, more intuitive way of separating ordinary success from critical success.

Take, for example, ability checks. If your game's ability scores are in the 1-20 range (or 3-18, for D&D,) then a roll under with 1d20 makes a lot of sense, even if some other checks like attack rolls use a roll over approach. Trying to switch to a unified mechanic means one of two things:

  1. If you opt for roll under, you need to change the attack rolls, perhaps using roll under Dex for ranged and roll under Strength for melee.
  2. If you opt for roll over, you need to derive new ability check target numbers for each ability, such as 21-Ability.

Or even worse, if you decide to go with roll over because "obviously" higher rolls should be better, but you want to minimize the effects of ability scores, you go with ability score modifiers such as the current D&D's (Ability - 10)/2, but that leads to people asking "Why do we even have stats? Why not just modifiers?"

And you lose the chance to make distinctions between easy and hard tasks in a simple manner. WotC D&D has been using DC tables for skill checks: Beat a score of 5 for Easy, 15 for Medium, 25 for Very Hard, with many other in-between points also defined. But do you really need that many? And even though it is simple, why not use the even simpler and more memorable system of "roll d20 under ability for most tasks, d100 under ability for extremely difficult ones"? You never need to change your target number that way.

I  was re-reading the Judges Guild Ready Reference Sheets recently, since they've been mentioned in a couple blogs. They use the d100 under ability for difficult tasks, like bending bars. But since a d100 is actually a two-dice roll, they were able to use another feature: on doubles, the character's ability is strained and unusable for several days. By using a different dice mechanic for one kind of action, they are able to exploit a feature unique to multidice mechanics -- rolling doubles -- to create additional detail with no additional die rolling.

Sometimes, the slavish devotion to unified mechanics means using the mechanic even when no mechanic is necessary, or complicating rules that don't require dice rolls. My last couple posts on ability rolls made a big point about not rolling at all, most of the time, because a character's high scores made that kind of action trivial. The new D&D 5, as already mentioned, has those DC 5 rolls, which seem kind of useless. It also has "passive checks", which do not require a roll, but do require claculating passive check scores for each ability (10 + ability modifier.) Again, wouldn't it have been simpler to make the passive check score the same as the ability score, so you wouldn't need to keep track of six extra scores? Passive check scores basically exist because the WotC way insists all mechanics must be d20 + ability modifier, so if you substitute a flat 10 in place of the d20 roll, you still need to calculate the final score.

And the sad thing is, it still is not really unified. The damage mechanic is still one or two dice of several different varieties, instead of being based on a d20 roll. It's OK to have variety there, apparently. Because, as I mentioned at the beginning, it really is all a matter of taste.


  1. I'm interested in this topic and the discussion that occurs because of your post.

  2. Anything can be made roll high is "better".

    d100 under ability can be d100+ability > 100. Similar d20 under ability can be d20+ability > 20.

    1. d100 + ability > 100 is probably one of the least fun dice mechanics I can think of. It's terrible for anyone even remotely dyscalculic, and much slower than roll and add with just about any smaller dice type.

    2. And MUCH worse than roll under, no adding necessary.

  3. It's funny, I used to be in favor of unified mechanics 'cause it seemed to make things easier, but I've come to realize that doing some things differently can actually be more intuitive. I also realized that some of my favorite stuff from 3.0 was actually those elements that didn't conform to the unified mechanic. Go figure

    One Mechanic to Rule Them All is a mistake, but it can be a good idea to recycle mechanics sometimes. One thing I loved in 3e was how flexible it could be (though it seemed as though nobody utilized its full potential). Swapping around ranks and abilities was a simple way to make ad hoc rulings and get a lot of mileage out of the skill system. All those synergy bonuses they added in 3.5 would've been better handled that way. Of course, you did have the issue of chance, experience or natural ability always having the same weight no matter what you were doing, and it wasn't always that unified when you consider some of the special rules they had for interpreting results

    Also, I need to find time to pore over those Ready Ref Sheets. I got 'em for Christmas, and it looks like there's a lot of cool stuff in there. Been working twelves, though, and my game-reading's been restricted to Tony Bath's Ancient Wargaming

  4. As you say in the beginning, both approaches have reason to exist and in most cases it comes down to a matter of taste. I tend to see this from another perspective, though. I believe a separation can be made between the rules important for the players and the rules important for the DM. Some of my players, for instance, are not very familiar with the rules of the games we play. I don't think they need to be, but that's not the point. It's rather that a unified approach for things that come up often during play is much easier to grasp and remember than lots of fiddly bits. So a rule of "high rolls are always good" has merit in that case. I think Cthulhu is a good example for this.

    The DM of a game, on the other hand, has a different approach to a system (for one, he should know it) and has different needs to make it work, so rules with a higher complexity in the results are of much higher value for the game from that point of view.

    (Btw., in our D&D game we used for ability and skill checks the formula (ability score [+skill bonus] vs. difficulty) for exactly the reasons I mentioned above, as "D20, roll high" applied to 90 % of what was needed at the table from a players perspective: checks, saves and attacks.)

    1. I agree with making a distinction between "rules for players" and "rules for the GM", although in my case, I prefer hardly any player rules, and a bundle of different reusable mechanics for the GM... but not necessarily more complex. More diverse in the results, definitely, which may be what you intended by "a higher complexity in the results".

      I want to stress that my intention, here, is not to argue the merits of roll high vs. roll low, but to say that the freedom to choose which is appropriate, combined with a desire for the simplest mechanic that gives the kind of results I want, is better than a slavish devotion to the unified core mechanic ideal. Especially when sticking to the unified mechanic results in unnecessary complexity.

      The passive checks in 5e, again, are a prime example. The game designers wanted a static number to compare to a difficulty rating, for detection tests. Characters already have ability scores, but using those as a static number didn't fit the preconceived notion that ability scores should add (Ability Score -10)/2 to some number, so they went with calculating passive check numbers, basically adding extra stats to keep track of. Yeah, a character with Wisdom 3 might do better with a passive wisdom check of 6, but you could also just use lower wisdom check DCs when designing adventures.

    2. Yeah, "diverse" is a way better word for what I was trying to say. But as a DM I can appreciate complexity, too. Like that text-to-dungeon idea you had some time back (which was inspiring, by the way ...).

      But it has to make sense in the greater context of a system (like you say, the system already has ability scores able to fulfill that function, so why not use them) and what you write about those passive checks sure sounds like bad design.