View Full Version : Questions on Randomness

2015-05-17, 02:52 AM
I have two questions on game design philosophy for the forum to chew over. They're largely unrelated on a detail level, but they both reach back to the core question of how random the game should be. I'm asking them because I'm picking at a class-based system with a d20 as its primary RNG.

In D&D, especially at high levels, characters tend to "fall off" the RNG in one sense or another - they hit or miss automatically on every roll against most non-epic DCs (including typical AC), and the same is true of the DCs they generate, with a few exceptions. In this case, rolling the d20 becomes a formality where you check for the occasional 1 or 20; with skills you don't even roll. Or they just go around the DC system entirely with no-save-just-suck style spells. I think this is pretty good for certain kinds of checks - skill checks that were hard when you were near level 3 should be trivial at around level 11, and those that are hard then should become trivial by the time you approach level 19. The same is true for certain kinds of attacks and spells - your level 14 fighter should be able to kick the stuffing out of three or four level 1 orc warriors without really trying (or getting hit). But what about the hypothetical wizard vs fighter situation? Disregarding what typically happens in D&D, if the fighter happens to catch the wizard with his pants down, should she be trivially able to punch him into next week? Or the reverse: if the fighter can't render herself impossible to target, should the wizard just be able to say "sorry, you're a bunny now"? If you say no to those last two questions for equal-level characters (and I do, I think), how much of a domain advantage should a character get in their area of expertise? At level 15 (this is the beginning of the highest power level bracket) in my system, as things stand, an expert in a particular domain (a certain kind of spell, or a class of weapons) will have a +15 bonus representing that expertise for offensive purposes. Against an attack from their domain of expertise or a similar one, they would usually have a +12 bonus to their defense, or possibly the same +15. Someone with minimal training at all related to that domain would have a bonus of +6 to attack and defense with that domain, and they would be missing key abilities needed to use it well. This is, of course, before ability scores, which would figure in on both sides one way or another. Is this reasonable? I was planning to resolve this problem with degrees of success (at 5 point increments). How do you feel about that?

Second, and largely unrelated:
What do you think about fixed weapon damage? In D&D, again, weapon-wielding characters usually manage to amass enough damage bonuses that the actual roll for damage doesn't really matter. A typical charge-barian deals something like five times his weapon's maximum base damage before he even rolls. Is that appropriate? How random should weapon damage be? 1d12 is too random for me, but nd6 seems a little boring, if damage is rolled. There's also the fact that rolling for damage adds an extra roll to the game. I was already planning on doing away with hit dice for HP; is it reasonable to do the same for damage?

2015-05-17, 05:27 AM
The problem in the d20 system does not originate from the randomness, but from vast difference in results (60% death 40% noting?)

There are too few rolls that does too much for the randomness to even out.

Beyond that you got silly scaling.

A second level is so much stronger than a first. A mage makes a massive leap in power every other level and some levels are completely irrelevant (say dead levels of non casters)

And to top it off you got the stacking nature of the system, where stacking beyond the point of dice relevance is easy enough to do without any real price or limits.

The system is simply broken due to content, rather than the randomness itself. Had the actions made less extreme with weaker and less game swinging spells and no absurd combos, it would work fine.
The leveling system still requires a total overhaul though on order to make a really solid system mechanically.

2015-05-17, 08:33 AM
You don't have to use a d20. 3d6 produces a curved distribution, which can make bonuses matter more. You're probably going to roll a 9 or 10, and a 16 or higher will be rather rare. That makes your random +2 from bard song or whatever more important. It also makes a +15 versus a +12 a larger gap, which might be what you are looking for.

You could also keep the d20 but make the bonuses scale more from level, so that your specialist has more than a +15 versus 'only' a +12 for the other side's defense.

Degrees of success are also interesting, and can make things less binary/swingy. Mutants and Masterminds has a different damage system you might want to look at, Legend of the Five Rings has 'raises' which are sort of like degrees of success, and I'm sure I am missing other examples too. Go read some more systems.

Weapon damage is largely up to personal choice; some people want to roll the dice and take their chances and others would be fine with the average or a static value. You could even have some people roll and some take average at the same table and my gut says the balance would be fine. So, do whatever fits your system better.

2015-05-17, 01:13 PM
A lot of this will be dependent on personal tastes.

Scaling defenses against the attacks in a manner beyond 10 + armor bonus + dex modifier, definitely helps make the attack rolls matter more at higher levels. If your 'fighter' archetype has +1 to BaB per level, and there's a bonus roughly equal to +1 per level to Armor Class, then a 20th level fighter has roughly the same chance to hit a 10th level foe as a 1st level fighter does against a 1st level foe. This is probably a good thing.

I've been working on a class-less level based D20 system where the basic mechanics include a flat +1 per level to Armor Class, Skills, Saves (which are handled as passive defenses like in 4E and Saga Edition), and Attack Rolls.

This scales really simply and intuitively. The power curve is completely straight and predictable. Since armor can boost AC quite a bit (plus granting damage reduction in my system), I also felt comfortable doing a couple things like turning proficiency into a bonus (similar to 5E). Also by recalibrating the Skill DCs to match the rest of the system, I think I've avoided a couple of the classic pitfalls.

One thing a lot of people seem to forget is to consider 'what is average'. If your average trained athlete rolls 1d20+10 for their athletic skill, well... an 'average' DC for an athletics check should probably be higher then DC10. I call this 'calibration'. Generally in the system I'm using, most skill DCs increased by around 5 from what they were in similar systems.

On the issue of 'rolled damage', you should consider some other things. If your system has some form of damage reduction from armor, I would really encourage you to use rolled damage... just to allow for the possibility of a hit getting through despite the DR.