Scale of RPG mechanics

[Roman]

What 'scale' of RPG mechanics do you prefer and why? By scale, I mean that most numbers used in the mechanics of the RPG will lie in that range. For example, d20/D&D uses mostly the low end of the double digit scale - most numbers - be they stats, skill points, etc will be in the low double digits - variation between them usually being in the single digits to low double digits. Of course, there are many exceptions but a strong tendency to low double digits is there. The advantage of lower scale is ease of calculations. The advantage of a higher scale is greater potential for accuracy and gradation.

What scale (averages, deviations from the averages, etc.) do you prefer? Please be as specific as possible to distinguish it from other scales.

 

[shadowlight]

Well, I'm not sure if it's what you're looking for, but I really like percentage systems (like the old Gamma World 3rd edition Action Table) - so a scale of 0-100 (or maybe a little higher depending on bonuses).

I also prefer non-flat distributions. That's usually done by using multiple dice, but it can also be done by varying success ranges (again, a good example is the old Gamma World 3rd edition Action Table).

I haven't found an in-print game that uses both of these (and the old Gamma World rules have other problems).

 

[MoogleEmpMog]

Generally, I don't see the need for higher-than-low-double-digits accuracy in an RPG. Does it really matter if your barbarian can jump a 20 ft. chasm but not a 21 ft. chasm? I doubt it. 19'3" but not 19'4"? One certainly hopes not!

I prefer mechanics to be as simple and easy to use as they possibly can be to achieve the desired effect, and smaller numbers can do that.

However, I do have one serious caveat about that: playing, as I like to, in a 100- or 255-level system, rather than 20-level, the numbers have to be scalable. Which usually involves larger numbers. In that case, of course, other aspects can be simplified to speed play.

 

[RFisher]

I like bell curves. Higher resolution where it's needed; lower resolution where it's not.

Unfortunately, I have a hard time finding a curve generated by dice that I really like. I'd like to experiment more with Fudge dice or a system like the one used by The Ladder. I can manage to get by with 2d6 in a pinch, though.

And mid20: Roll 3d20 & drop the high & low ones. If I were to ever run a d20 system game again, I might use that.

There are some other problems with curves. Such as the fact that it is then harder to apply a linear modifier. (But then, I've wanted to use modifiers on a curved scale when using a linear scale, which is also harder.)

I think d10s are fun because they can provide nearly limitless resolution on an as-needed basis. If you have a 64.35% chance of success, roll 1d10. If it's 5- you succeed; 7+, fail; 4, roll again. If it's 3-, success; 5+, fail; 4, roll again. If 2-, success; 4+, fail; 3, roll again. If it's 5-, success; 6+, fail. I've always wanted to build a system off that. (Of course, you could do it with any sort of dice, but d10s make it easy for us base-10 thinkers.)

 

[shadowlight]

And mid20: Roll 3d20 & drop the high & low ones. If I were to ever run a d20 system game again, I might use that.

Now that's an interesting idea... I may have to give that a try!

 

[Roman]

I too prefer using bell curves as often as possible and therefore favour using multiple dice for any roll.

As to scale, I am not sure what I like. Something about a scale centered on 100 appeals to me. Perhaps it is because human metrics such as IQ, EQ, LQ, CQ, etc usually use 100 as the mean. On the other hand, I also find a percentage based system attractive, even though it is not really compatible with bell curve rolls. Hmm...

Also, how much randomness do you like in a system? How much should success in general be influenced by modifiers and how much by the roll itself (for example rolling a d20 when average modifier is 10 is very different from rolling a d20 when average modifier is 100)? Also, how much variation do you like to see from the mean modifier? Suppose that an average bonus, modifier, stat, ability score, skill score, whatever is 10. How much of a variation do you like to see from this mean?

 

[Roman]

Also, I forgot to ask: Do you prefer a linear or an exponential scale of stats, skill ranks, modifiers, etc.?

By exponential I mean for example: Strength of 20 designates somebody 2 times as strong (say in lifting potential) as a strength of 10. Strength of 30 designates somebody 2 times as strong as a strength of 20. Strength of 40 designates somebody 2 times as strong as a strength of 30.

By linear I mean for example: Strength of 20 designates somebody 2 times as strong (say in lifting potential) as a strength of 10. Strength of 40 designates somebody 2 times as strong as a strength of 20. Strength of 80 designates somebody 2 times as strong as a strength of 40.

 

[GuardianLurker]

The problem with expoential is that it does a *great* job at the high-power end, but the low power end gets *very* squished. MEGS is the best example I can think of.

Linear flips the problem - highly detailed at low power, overwhlemingly detailed at high power. HERO and V&V spring to mind as examples.

It all depends on what you're trying to get out of the system. For fantasy, high or low powered, I find the d20 linear scale works just fine.

 

[Impeesa]

I think another word you're looking for is 'granularity' - the level of precision. For example, White Wolf's games have a very low granularity - the entire spectrum of human capability is basically rated from 0-5 dots. At the other end is Rolemaster, which uses a system similar in concept to d20 but with a d100 as the basis. I think D&D is pretty good in that respect - maybe slightly too low, but it's hard to scale it up with the dice we have without jumping straight to the percentile system (which is more precision than we need, really).

--Impeesa--

 

[Roman]

The problem with expoential is that it does a *great* job at the high-power end, but the low power end gets *very* squished. MEGS is the best example I can think of.

Linear flips the problem - highly detailed at low power, overwhlemingly detailed at high power. HERO and V&V spring to mind as examples.

Good points, but I think an exponential system could be done with such 'granularity' (to borrow Impeesa's word) that it would be also detailed at low power. Hence my question about what scale/granularity you ladies and gentlemen prefer.