#26: Randomization of Performance

[RPGnet Columns]

http://www.rpg.net/columns/roughquests/roughquests26.phtml

Summary:

Different ways to use randomizers, and a decision for Rough Quests.

Go to the column (http://www.rpg.net/columns/roughquests/roughquests26.phtml) for more information.

[fmitchell]

Recently I've been reviewing different game systems, mainly on their handling of conflict resolution rolls, and I've come to markedly different conclusions from yours.

Part of the difference is that you regard attributes on a bounded scale (1-6, 0-100, whatever), and that a novice can do by accident anything an expert can do reliably. I assume the opposite: however good you are, there's always someone better, but someone far better at any level can do things you could only dream of.

So, here's a different take:

Single Die: After playing d20 for a little bit, I'm irked that a single die determines my fate. My character has a Dex of 18 (+4) and a Stealth of +3, yet sometimes I find myself with the lowest Stealth roll because I got a 2 while the lumbering half-giant got a 19.

Multiple Dice Added: I think you give bell curves short shrift. The GURPS and Hero ones are kind of weird, placing each adjusted skill level somewhere on a fixed bell curve; slide past the hump and your chances drop horrendously. On the other hand, systems using Fudge, PDQ, and in a way d6 Classic give you a bell curve centered around an expected level of performance.

In Fudge, for example, if you have a Good Stealth (Fudge uses adjectives, although you can tranlate each to a -4 to +4 or more scale), then you'll reach or exceed that level about 62% of the time; sometimes you'll do a level or two better, sometimes you'll have an off day and do a level or two worse. Rarely will you do extremely well or badly. If you have a Poor Stealth, or a Legendary Stealth, you'll have the same bell curve, but you can still expect to do at least Poorly or Legendarily on average 62% of the time.

Fudge uses funky dice (equivalent to 4d3 - 8), but PDQ uses 2d6 + bonus against a Difficulty Factor. PDQ's standard difficulty factors coincide with the different bonus levels, in multiples of two: Poor (-2 bonus, 5 DF), Average (0 bonus, 7 DF), Good +2, 9 DF), Expert (+4, 11 DF), and Master (+6, 13 DF).

As you can tell, this is my preferred method, so I'll stop here.

Dice Pools, Count Successes: The fact that an expert can get more successes than a novice is a win for me, although I don't like rolling a handfull of dice, checking each number, and setting aside the successes. Some games like Burning Wheel assume a fixed "difficulty factor", and there are horrendously expensive Burning Wheel dice customized so that successes and failures are obvious.

Dice Pools, Pick Highest Roll: John H. Kim analyzed one such game at http://www.darkshire.net/~jhkim/rpg/systemdesign/dice-icon.html and noted that after 3 dice, adding more doesn't help probabilities much.

Actually, I recommend Mr. Kim's full set of articles, starting at http://www.darkshire.net/~jhkim/rpg/systemdesign/dice-motive.html , as a starting point for considering dice mechanics.

[smascrns]

Recently I've been reviewing different game systems, mainly on their handling of conflict resolution rolls, and I've come to markedly different conclusions from yours.
Excellent, that means there is scope for very different takes on the issue.

Part of the difference is that you regard attributes on a bounded scale (1-6, 0-100, whatever), and that a novice can do by accident anything an expert can do reliably. I assume the opposite: however good you are, there's always someone better, but someone far better at any level can do things you could only dream of.
Actually I'm not that far off from your stance. I don't like that much bounded scales because they relativise what usually is strongly stated as an absolute. Say, the hero in fantasy is the best in his field, point. There's no level up his. Open bounded scales mean there is always one level more. Closed bouded scales work better but they put a cap on what is achievable. It's because of things like these that I discussed in past columns the concept of relative scaling. Check the next:
http://www.rpg.net/news+reviews/columns/Ruleslawyer_Sergio_9_1.html
http://www.rpg.net/news+reviews/columns/ruleslawmay00.html
http://www.rpg.net/news+reviews/columns/ruleslawmay00.html
http://www.rpg.net/news+reviews/columns/ruleslawmar00.html
http://www.rpg.net/news+reviews/columns/ruleslaw21feb02.html
I'll get back at this issue latter on.

Multiple Dice Added: I think you give bell curves short shrift. The GURPS and Hero ones are kind of weird, placing each adjusted skill level somewhere on a fixed bell curve; slide past the hump and your chances drop horrendously.
That's what I had i mind.

On the other hand, systems using Fudge, PDQ, and in a way d6 Classic give you a bell curve centered around an expected level of performance.
I had these in mind.

Fudge uses funky dice (equivalent to 4d3 - 8)
Something I tend to avoid.

but PDQ uses 2d6 + bonus against a Difficulty Factor. PDQ's standard difficulty factors coincide with the different bonus levels, in multiples of two: Poor (-2 bonus, 5 DF), Average (0 bonus, 7 DF), Good +2, 9 DF), Expert (+4, 11 DF), and Master (+6, 13 DF).
This is not that different from GURPS, HERO and the like. Actually the difference is that PDQ is more attentive to the distribution of results in a 2d6 dice mechanic. It basically attempts to optimise the mechanic found in those other games but the core is the same.

Dice Pools, Count Successes: The fact that an expert can get more successes than a novice is a win for me
I like this but it has a drawback: The novice can never do as well as the seazoned, he can never really surprise him. This goes counter real world experiences and fiction. This leads me to exclude such a mechanic.

You see, unlike you I like a game system where "a novice can do by accident anything an expert can do reliably"... provided that the accident is exceptional and not the norm. That keeps experts on their toes since they cannot take the novice for granted. It's all an issue of probabilities.

although I don't like rolling a handfull of dice, checking each number, and setting aside the successes.
Which is a looser for me (except in the case of very simple dice pools, count successes like in Prince Valiant).

Dice Pools, Pick Highest Roll: John H. Kim analyzed one such game at http://www.darkshire.net/~jhkim/rpg/systemdesign/dice-icon.html and noted that after 3 dice, adding more doesn't help probabilities much.
Well, I put the break at 5 or 6 dice. Even then it would be too short but you have to notice that I don't use a Pick highest DP. Instead I use a Pick Highest or Lowest DP. This means that my range of dice pools goes from 5 or six to 11. This has also to be compounded with some other concerns:
First, the granularity of the system. I prefer systems with low granularity.
Next, how many steps are there in the resolution mechanics. More steps mean that the relevant factors are distributed by several rolls. I prefer two to three steps.
If we combine both these points with a fine tuning of the system the low granularity of a DP, pick highest/lowest can be perfectly adequate to handle game resolution. That's my target.

Actually, I recommend Mr. Kim's full set of articles, starting at http://www.darkshire.net/~jhkim/rpg/systemdesign/dice-motive.html , as a starting point for considering dice mechanics.
Thanks for the link.

One of the things I like about DP, pick highest/lowest is that they ensure that each character will act in a reliable way as he supposed to act (the expert does expert things while the novice does novice things), while retaining the chance to misbehave. I also like the simplicity in reading results from the roll.

[Grop]

Sergio,

You are probably already aware of this, but La Méthode du Docteur Chestel uses a dice pool that sounds similar to yours, based on 2D6. Here is a link to a description of this game:

http://www.rpg.net/news+reviews/reviews/rev_5898.html

I have played this game: it is fun, but action does not matter much in such games.

[smascrns]

No, I didn't know about Dr. Chestel. Yes, the dice pool works the same way with the difference that Dr. Chestel introduces a bell-curve with its 2d6, thus changing the probabilities. Indeed, there's no new things under the sun...

Thanks for the reference.