Why don't 3e and 4e use percentile dice for skills?

[ferratus]

Eh... I get this way every so often, but it goes away. It's why I can never stay with GURPs too long. After a while of just using d6's I see all my other dice in the pile and start itching to let them have some fun too.

I don't think I've used my d4 since I switched to 4e. d8's and d6's seem to be the most common for damage, d12 and d10 get used once and awhile for big swords and axes. d20 takes care of everything else.

It is actually rather odd that damage die are so varied. You only need 2 inches of steel to kill someone. One weapon isn't more lethal than another because they do more damage, but that they are more suitable for particular situations.

 

[ggroy]

I don't think I've used my d4 since I switched to 4e.

Magic missile uses 2d4 for damage.

 

[Scribble]

I don't think I've used my d4 since I switched to 4e. d8's and d6's seem to be the most common for damage, d12 and d10 get used once and awhile for big swords and axes. d20 takes care of everything else.

It is actually rather odd that damage die are so varied. You only need 2 inches of steel to kill someone. One weapon isn't more lethal than another because they do more damage, but that they are more suitable for particular situations.

Truthfully I think the could have made all the weapons in 4e do one basic damage, and then let the powers and feats influence what damage they do. The weapon choices would open up different power sets.

 

[keterys]

I don't quite understand, could you elaborate? Is it a matter of smaller numbers being easier to add?

Not precisely, but close enough. Take two examples:

Trying to climb something difficult:
Base Skill: +10
Climbing Gear: +2
Difficult DC: 25, if fail by 5+ then fall

1d20+10+2 is easy math, rapidly done. DM watches for DC 25+ or 20- as triggers.

Base Skill: 51%
Advanced Climbing Gear: +12%
Difficult Wall: -15%

Critical Failure if fail by 25% or more* (because of the awkwardness of critical failure math of this sort in percentile, most such systems might have it just be flat 91 or 96 to 100 instead, which has pros and cons, especially when you get your skill up to the 90% range)

51+12-15 is more difficult math (and requires the DM to inform the penalty) then rolling d100 to see if you got the 48% or less or got in the falling range, depending how you do that.

Basically, adding comes more naturally to people than subtracting and smaller numbers more easily than large ones. Also, the percentile roll in general has a more scientific connotation than the other dice types, so I slightly prefer not using percentile for fantasy games and slightly prefer for scientific games.

 

[Theo R Cwithin]

Does a DC 25 challenge equate to a 125% chance of failure?

IMHO, no. Logically, you can't get beyond "100%", whereas in a d20 system, DCs above 20 are routine!

As I see it, DCs aren't really percents, they're just arbitrary numbers.

Now if you ask the same question, but say "d100" instead of "d%", I don't think there's any real reason other than tradition? easy arithmetic? branding?

 

[ferratus]

I see what you mean now. Unless you want greater precision than 5%, there really isn't much point to using percentile dice instead of a d20, and if you have a degree of percentage greater 5% you have problems with the difficulty of double digit addition and subtraction. That explanation stirred my memories of figuring out success of climbing for my 2e thief.

I think the problems with approaching 95% success aren't really solved by a d20 though. After all, the d20 largely has a ceiling of DC 40 for the most difficult tasks, and as you go up in levels it soon becomes impossible to have critical failures there too on everything except opposed rolls.

However, it is clear that if I wanted to bring in percentile dice, it would largely be seen as an unnecessary annoyance. I still think the odds are easier to visualize on the fly if you use percentages, but perhaps it isn't worth the trouble. I also like the idea of breaking up the combat/non-combat mechanics superficially to make it seem like they are two different states of play, though the underlying principle of success and failure would largely still be the same mechanic. That is probably just inefficient vanity though.

 

[ferratus]

I was thinking that <= 0% chance of success (or > 100% chance of failure if you prefer) would translate into "impossible for the unskilled" or even "only possible for those with supernatural powers".

 

[Cadfan]

Well, a d20 is basically a percentile dice that works in 5% increments. So unless you have a pressing need to calculate with greater granularity than that, the d20 gets the job done.

As for why you'd pick it instead of the percentile, there are two possible reasons. First, it keeps things consistent with the rest of the game, and there's value in that. Second, the d20 system scales more easily as characters gain in level. The percentage chance of success at a given task changes as the character's skill level increases in comparison to the static task DC. In a percentile system you have to recalculate the math as you go.

Did that last bit make sense? Let me give a quick example.

In a d20 game, if a task is a DC 25, and you start with a +5 to your skill check, you begin with a 5% chance of success. As your skill check improves in 5% increments at a time, you get a 10%, then 15%, and so on over the life of your character. But what you do remains the same- roll a d20, add your skill check number, and see if you meet or exceed 25.

In a percentile game, things are a little different, and in my opinion a little more unwieldy. You generally have to roll below (or above) a target number that reflects your chance of success. So to model what we just modeled, you'd start out needing to roll a 5 or below, then a 10 or below, and so on as your character grew in ability. But how do you know that the target number needs to be a 5 at one point, then a 10, and so on and so forth? You need some additional algorithm to generate that target number. And that adds an additional bit of complexity that isn't evident in a simple "roll d20, add modifier from character sheet, see if you meet or exceed the target" system.

Of course, a percentile system can model things in ways a d20 cannot. If you created an algorithm to generate a bell curve, you could do that (3d6 would be faster and just as good though). Or you could do a logarithmic curve of some kind with the appropriate algorithm.

 

[ggroy]

I think the problems with approaching 95% success aren't really solved by a d20 though. After all, the d20 largely has a ceiling of DC 40 for the most difficult tasks, and as you go up in levels it soon becomes impossible to have critical failures there too on everything except opposed rolls.

One lazy way to get around the "95% success" problem, is to scale up the DC's by level. In the 4E skill check system, the stated DC's for level 1 skill checks are:

- 5 easy
- 10 moderate
- 15 hard

A lazy way to scale up the DC's by level, is to add level/2 to the DC's along with the baseline 5, 10, or 15 for easy, moderate, and hard skill checks respectively. For example, a level 22 skill check for a moderate difficult task would require a DC of 21 while a hard task would require a DC of 26. (The actual DCs from the 4E DMG2 are slightly different, but do scale up with level). In practice for 4E, the DC's should also take into account the player's ability score increases which happen every few levels.

This same lazy way of scaling can be done for AC and other defenses of monsters and badguys, traps, etc ...

 

[Jhaelen]

This reminds me of my thoughts when reading the 'Eclipse Phase' rpg. They're using percentile dice for everything, but almost everything comes in 10% increments and they recommend to use 5% steps when point-buying abilities and skills.

About the only advantage I could see is the fumble/crit mechanism they use:
If both dice show the same number it's either a crit or a fumble (depending on whether you beat the (adjusted) difficulty or not).

Otherwise they'd definitely been better off just using a d20.

About the only rpg where using percentiles made sense to me was Runequest. And even some of the various derivates (Stormbringer, Pendragon) condensed the system to use a d20, IIRC. Curiously, Call of Cthulhu didn't (again, IIRC).