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

[Mustrum_Ridcully]

How do you factor in difficulty and skill in a percentile roll?

Typically, difficulty will be modelled by a substraction. Substractions typically seem to be "harder" for people than adding stuff.

The alternative would be to roll a d%, add skill, and compare to a difficulty number. But then, we can't really see the success rate directly anymore, and the advantage of a d% is gone.

 

[(Psi)SeveredHead]

For example, that you have a 50% chance to scale a rough wall without scaling tools, while only a 10% chance of scaling a wall made of ice. Or that you have a 70% chance of convincing a gullible child that you are a great hero, but only a 10% chance to persuade the sultan that you are a sheik from the land of Ababwa.

My experience with Warhammer says this is a bad idea.

DMs seem to forget they can modify DCs (well, the equivalent), and even when they do, the DC mods are usually pretty small.

Although I could just be bitter because modifying your own percentages was hard (for most skills in Warhammer), made worse because starting percentages for humans were only 20-40 percent per skill.

If you have a stealth skill of only 35%, that's hardly useful. And then there's roll-offs...

 

[ferratus]

Yeah, I remember being a 2e thief, and how you wouldn't do anything with your thief skills until they got over 50% chance of success. With 3e and 4e you generally start out able to to the basics with your skills fairly reliably right off the bat.

I think I've been convinced as to why the 3e D20 DC system for skills is better. I'll save my % die for DM stuff like weather, encounters and random treasure tables.

 

[ggroy]

How do you factor in difficulty and skill in a percentile roll?

The percentile dice system used in the form of the resistance table in Chaosium's "Basic Roleplaying" (BRP) ruleset.

Mathematically the "DC" percentage to roll the percentile dice under or equal to, for a success in the Chaosium BRP resistance table is:

50 + (Active_Characteristic - Passive_Characteristic)*5

where any

- (Active_Characteristic - Passive_Characteristic) >= 10 is an automatic success

- (Active_Characteristic - Passive_Characteristic) <= -10 is an automatic failure

(ie. Any "DC" percentage greater than or equal to 100 is an automatic success, while any "DC" percentage less than or equal to 0 is an automatic failure).

The percentile dice system used in the form of the resistance table in Chaosium's "Basic Roleplaying" (BRP) ruleset, can be converted to d20 relatively easily.

 

[Bedrockgames]

I think for a few reasons:

They wanted to stick with a one roll mechanic and keep the d20 because it is strongly associated with d&d.

They wanted something that seemed relatively simple. A d20 plus modifier, is very simple. A percentage roll draws attention to the math aspect of the game, and invovles two dice.

Percentile is asocciated with deeply engineered games like Role Master. Especially with 3E, they wanted to move away from old school and my sense is, percentile dice still have an old school vibe.

 

[ggroy]

Here's an equivalent d20 mechanic that reproduces the same probabilities as the Chaosium BRP resistance table, where rolling higher on a d20 is better.

The attacker's roll is:

d20 roll + Active_Characteristic

against the DC of the defender of:

11 + Passive_Characteristic

(This assumes there are no critical failures and successes, on natural d20 rolls of 1 and 20 respectively).

 

[karlindel]

I think the main reason is the added difficulty of adding or subtracting larger numbers. It is a lot easier to add single digit numbers to a d20 roll than it is to subtract double digit numbers from a percentile roll.

This is a particular problem if you want scaling critical successes and failures, as every roll will require adding or subtracting more numbers. Degrees of success based on how much you beat the difficulty are also more complicated when using percentiles.

The main benefits of the percentile system are that the odds of succcess are clearer to the player, and you can use increments of success that are less than 5%. However, most people are not that concerned with that extra level of granularity, particularly as it makes the math much more complicated (the math for percentiles and d20 rolls are similar in complexity if the percentiles are only changed by increments of 5%, although percentile is still the more complicated of the two).

 

[cattoy]

Same reason we don't use percentile dice for saves and to hit rolls.

 

[Minicol]

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.

Wasn't it what Warhamster 1e was all about

Cheers

 

[TerraDave]

Good topic, to which I am late. Oh well.

To summarize: roll under perecentages have great intuitive appeal, but are complicated when not all tasks are the same:

-double digit math for modifiers
-variable difficulty that undermines the intiutiveness of the system.

AD&D actually did have a pretty good compramise on this: use d20s for attacks, which would have modifiers and changing difficulty (in AC), and use % for skills, and then keep the chance pretty much the same for those.

For (non-D&D) systems that used % for attacks, you might have damage reducing armor, and maybe a parry or dodge check. So you keep the base %, and now have some more realism (but it will go slower...)

Another way to do a % based system with mods is to have the % chance be the difficulty, and then give the charecter some standard bonus, and keep to 5 and 10% modifiers. So a 50% chance to climb a tree, and your charecter has a +25% bonus from strength, level, etc, for a 75% chance. The key is to mod the % chance, not the role.

But d20 does work pretty well.