I must create a system, or be enslaved by another man's. I will not reason and compare: my business is to create.

- William Blake

Tuesday, October 04, 2016

The fallacy of unified mechanics

In the classic TSR-era, D&D famously had a big number of task resolution mechanics. Some things used 1d20, roll high. Others, 1d20 roll low. Thieves skills used both a d% and 1d6 roll low - same as most other tasks. Reactions and morale used 2d6. Initiative, 1d6 roll high. And so on.

The number and type of dice were not the only thing the varied: attacks with a weapons would be rolled by the attacker, while spells would grant a roll to the defender, etc. A number of tasks would require having the necessary resources (spells, equipment, HP, etc) and no roll. Some stuff seems to be decided by GM fiat (a witches reaction to a high-charisma PC, for example).

And, of course, many old school enthusiasts believe most challenges would be resolved without using the dice, just by describing the PCs actions. Some will even rage against thieves skills as detrimental to the whole game.

Nowadays, I'm somewhat used to it. But at some point in the nineties that irritated me greatly; the "new" D&D stuff that I liked (AD&D 2e and the Rules Cyclopedia) insisted in the same confusion as the original versions, when there were more streamlined games around. Why not just use one single method of resolution?

As you might know from my last post, I eventually started playing GURPS, and when D&D 3e hit (to supposedly fix such problems, but we all know how that went) I already thought D&D was too complicated. One of the things that attracted me in GURPS is how everything was "roll 3d6 under your attribute or skill". So simple!

People call this "unified mechanics" - all "modern" games do it. And, of course, D&D eventually followed suit in the year 2000. 1d20, roll high, for everything! Quite elegant!


It is easy to see how old school D&D might look messy at first. Let us take what, in my opinion, is the simpler, cleaner, and easier version of D&D ever (Moldvay´s) as an example:

- Attack and saving throws use a d20 - you want to roll high.
- Well, there might be a saving throw for falling, if the GM allows it. It uses a d%.
- Thieves skills use d%. Well, except hear noise. It uses 1d6, like most other tasks anyone can do (forcing doors, etc). Roll low.
- Oh, initiative also uses a d6. You want to roll high.
- There is also some uses of 2d6, but the GM rolls them. Monster reaction uses 2d6 - rolling low means the monster is aggressive. Morale (optional) uses 2d6 - you want to roll low, because rolling high means surrender.
- Other tasks use a d20. Except you want to roll low, lower than your ability score. A "20" is always a failure. Climbing a rope, for example. No, climbing sheer surfaces use a different mechanics. Yes, most thieves might have better chance at climbing sheer walls than climbing ropes...

Compare this to 5e (and most WotC D&D):

- Attack roll? Roll a d20, and roll high. Saving throw? Same thing. Skills? Yup, d20, roll high. Thieves too? Yes. What about ability checks? Same deal. Initiative? d20. Morale (optional)? That is a d20 check. Roll high.

Pretty good, huh?

However... Skills and abilities become quite unreliable. Your dexterity 18 gives you 40% extra chance in succeeding in some crazy acrobatics test in B/X (if you use "there is always a chance"), but only 20% in modern D&D. And a 3rd level thief can often succeed where a 10th level thief has failed.

Same with initiative - you could be quite sure of acting first in B/X, but with a d20 your dexterity means very little. Forcing doors with Strength 18? That would be a 5/6 chance in B/X, more than double the usual 2/6 - but in modern D&D your Strength bonus will mean nothing 80% of the time. Likewise, morale becomes quite the coin toss.

The problem isn't numbers; the problem is my players that don't even care about the math realizing the dice is more important than their characters when resolving most tasks (and they do).

Well, at least everything is resolved the same way now, right? From fighting to, say, climbing? Really straightforward! Not quite.

When you climb, you roll a d20. If you roll high enough - success!

When you attack, you roll a d20. If you roll high enough - success! But wait, roll for damage. No, no d20s here - a d4, d6, d8, d10 or d12, depending on the case. Rolled a 1? Yeah, not such a great success after all. What if you rolled the maximum number? Well, then your enemy loses some HP. Success? Not quite, you must roll the d20 again next round to attack again. Did we roll initiative? Yeah, you must roll a d20 BEFORE attacking, to see who goes first...

I see no honest way to say that combat and skills use the same mechanics in any version of D&D. And, if they don't, using a d20 for both seems forced - because there are many more effective ways that have been used before.

Even when 5e tries to fix the d20 problems - but giving reliable skills to rogues or reliable strength for barbarians - it does so by adding NEW mechanics, often more complicated than just rolling a different kind of dice.

Back to GURPS then! At least 3d6 creates a good bell curve, making skills reliable, and all damage uses d6s. Not 3d6, but, still...

The problem is that "reliable" and "combat" don't mesh too well. Fighting can become quite predictable, and depending on the rules you're using, you can spend rounds and rounds with nothing happening because both foes have more than 70% chance of blocking an attack! There are some critical hits, but they are hard to come by... and then you roll in a critical hit table where "nothing happens" is what you 37% of the time!

You can imagine "reliable combat" got a bit boring after a while.

Now, let us make a few things clear. B/X. GURPS, and D&D 5e are some of my favorite RPGs.

And I quite LIKE the idea of unified mechanics - as one possible goal, not as a straitjacket. Yes, I have proposed some for B/X in this very blog and I use them in my own RPG, Days of the Damned.

But, here is the point:

The most important thing about a game mechanic is that it is FUN and it WORKS.

The rest is secondary. The fact that mechanics are similar to each other is a nice addition, but not a requirement. Games are meant to be PLAYED, not to be admired by the reader ("oh, how elegant, armor class is equivalent to difficulty class"...).

If the price for fun is using some different types of dice, I'll pay it gladly.

ADDENDUM: how to solve all my problems with GURPS, B/X and 5e, "unified" style!

These ideas are old, but they work well enough.

- Combat uses 1d20. A "bell curve" effect is created by dividing it in many rolls. Saving throws use the same because the risk is exciting. Skills uses 3d6. "Natural" bell curve.

- For 5e: use 3d6+mods, 4d6 if you have advantage, 2d6 for disadvantage. DC 30 is actually "nearly impossible" now. Or just fix the DCs accordingly (maybe 10/14/18/22/26).

- For B/X: everything but combat and saving throws is now 3d6 roll under, 4d6 if the task is hard, etc. Thieves have no skills, but they get a bonus equal to their level when trying to hide, climb, open locks, etc. It is not hard to remember: roll high with 1d20, roll low with d6s. Natural 20 good, 666 bad. Morale and reactions stay as they are, because they work.

- For GURPS: use a d20 in combat and high-risk scenes. And no more "nothing happens" results in the critical hit table!


  1. I very much agree, simplified mechanics are only good if the overall game plays better. Simplification for its own sake is not elegent, its merely simple. A bit of variety in mechanics keeps things interesting and runs better maths wise.

  2. I conclude - if I'm interpreting correctly - the thief(=rogue)skill solution with d% makes more fun than the d20 solution (which is, in fact, nothing else as a d%/5)? Oh, come on, there must be a solid argument for the skill d% chart oder arbitrary d6 discrepancies on top of that.

    1. That is... not what the post says. I actually think using a d% for that is a bad idea. I proposed one solution for this in here, but you can also just roll 1d20, add thief level, and try to beat 20 (delta's "Target 20" mechanic).

  3. You could solve the unreliable Skills/Attributes by making the bonus bigger. If a 14 Strength would give you q +8 instead of a +2 then the 14 Strength fighter with 14 Dex would beat the 12 Strength Rogue with his +4 Bonus.

    Making the numbers bigger would also lessen the influence of the d20.

    1. Yes! I've been playing with this idea. Currently, I'm using +2 for 12, +4 for 14, etc. It doesn't solve the issue completely, but makes things a bit better.