#25: The Performance Stage in Action Resolution

[RPGnet Columns]

http://www.rpg.net/columns/roughquests/roughquests25.phtml

Summary:

Looking at different ways to combined attributes and skills in a resolution system.

Go to the column (http://www.rpg.net/columns/roughquests/roughquests25.phtml) for more information.

[torbenm]

There are a few questions you should ask yourself before choosing a way to combine attribute and ability into a roll:

1. Should attribute and ability affect the roll in the same way or differently?

2. Do you want degrees of success and failure?

3. If you have opposed actions, do the opponent combine attribute and ability in the same what that the active player does?

4. Should a person with low performance score have a (however slight) chance of beating a person with high performance score?

That way, you can better decide if any proposed rolling mechanism will be appropriate.

For example, let us assume that the answer to question 4 is "no", then dice pools where you count successes is inappropriate, as even the most able character can roll no successes. On the other hand, mechanisms where you add your performance score to a dN and must exceed the opponents performance score by a certain number T will be fine as you can never beat an opponent whose performance score exceeds your own by N-T or more.

Similarly, dice-pools lend themselves easily to degrees of success, while roll-under-ability mechanisms require more complicated rules to handle degrees of success.

[smascrns]

Four good questions Torbenm. I'll refer to them in future columns when I discuss dice mechanics. In any case my answers are:

1. They do in Rough Quests as I mention somewhere. The reason is that this is a pre-modern game where technical skills are comparatively unadvanced.

2. Yes.

3. Yes. I like simetry and consistency.

4. Yes. That's why there are resolution mechanics. Besides, the players may always decide that actions are automatic when they consider that the low performance character should not be able to beat the high performance one. They don't need to waste time and resort to game mechanics for this purpose.

[torbenm]

Four good questions Torbenm. I'll refer to them in future columns when I discuss dice mechanics. In any case my answers are:

1. They do in Rough Quests as I mention somewhere. The reason is that this is a pre-modern game where technical skills are comparatively unadvanced.

2. Yes.

3. Yes. I like simetry and consistency.

4. Yes. That's why there are resolution mechanics. Besides, the players may always decide that actions are automatic when they consider that the low performance character should not be able to beat the high performance one. They don't need to waste time and resort to game mechanics for this purpose.

As for 4, if there is even a remote chance of success against any opponent, you will have players insist on rolling. It requires a strong GM to say "No" in such situations. Hence, I prefer that the mechanics temselves set a limit for when you can succeed, i.e., when the difficulty gets too high or the opponent too skilled. Symmetrically, there should be tasks at which you can't fail if you are sufficiently skilled (and not otherwise impeded). But I agree that it depends on the style of game you want.

Now, there are several mechanics that can achieve your wishes. Given that you have attribute, skill and experience field all at a seven-step scale, adding them up gives us values from 3 to 21 (or -9 to +9). That probably rules out dice pools where the number of dice equals the value, as you quickly get to an unmanageable number of dice.

A simple addition of the value to a dice-roll will work, and to make opposed actions symmetric, you can use a zero-centered dice-roll. In order for a person of value 3 to have a small chance to beat a person of value 21, you need the dice to have a wider range than the range of abilities. Using d10 minus d10 (range -9 to +9) just about manages that, as 3+9 = 21-9. This gives a tie at best, but since this is the extreme case, it is not so bad.

You can get degrees of succes by how much you exceed the difficulty (or opponent's value). Depending on how many degrees you want, you can scale the difference down to fit.

You can also take a different approach: Let each of attribute, skill and experience field determine the size of a die (value N gives you d2(N-1). For example, value 4 (average) gives you a d6). This gives you three dice in the range from d0 (always 0) to d12. Now roll all three and take the median (middle) result. For example, if you roll 3, 9 and 8, the middle value is 8. Now compare this to the difficulty or opponent's roll (made the same way) to determine degree of success. Again, some scaling might be required.

If you don't like scaling the result to get a degree of success/failure, you can use a variant of the second proposal, but instead of taking the median result, you pair your three dice-rolls against the three rolls of the opponent: Highest to highest, second-highest to second-highest and lowest to lowest (like in Risk) and count how many more (or fewer) pairs you win (ties count to neither side, unlike RIsk). For example, asssume you roll 3, 7 and 8 and your opponent rolls 4, 7 and 9, then you pair up 8<9, 7=7 and 3<4, which gives you a result of -2 (or, equivalently, your opponent a result of +2).

This gives results from -3 to 3, which is suitable for a degrees of success. In particular, this gives you seven degrees, which fits your scale of attributes etc.

For unopposed tasks, you compare each of your three dice to the difficulty number (from 0 to 12) in a similar way.

As for whether this allows a worst-possible character to beat the best-possible, the straight answer is "no", as the worst possible will "roll" all 0s, while the least the best possible can roll is three 1s. However, if you have at least two d2s (and one 0), you can beat three d12s (though you are not likely to). So, I think this criteria is reasonably well fulfilled.

[torbenm]

If you don't like scaling the result to get a degree of success/failure, you can use a variant of the second proposal, but instead of taking the median result, you pair your three dice-rolls against the three rolls of the opponent: Highest to highest, second-highest to second-highest and lowest to lowest (like in Risk) and count how many more (or fewer) pairs you win (ties count to neither side, unlike RIsk). For example, asssume you roll 3, 7 and 8 and your opponent rolls 4, 7 and 9, then you pair up 8<9, 7=7 and 3<4, which gives you a result of -2 (or, equivalently, your opponent a result of +2).

This gives results from -3 to 3, which is suitable for a degrees of success. In particular, this gives you seven degrees, which fits your scale of attributes etc.

For unopposed tasks, you compare each of your three dice to the difficulty number (from 0 to 12) in a similar way.

As for whether this allows a worst-possible character to beat the best-possible, the straight answer is "no", as the worst possible will "roll" all 0s, while the least the best possible can roll is three 1s. However, if you have at least two d2s (and one 0), you can beat three d12s (though you are not likely to). So, I think this criteria is reasonably well fulfilled.

Instead of having constant 0 at minimal value, you could roll a d2-1. This gives a worst-possible character a chance to match (but not beat) the best possible, if all dice come up as ones. And if just one die is non-minimal, the worst-possible character can actually beat the best possible (though at very low probability).

Also, instead of using fixed difficulty numbers, you can translate these into dice the same way you do for ability numbers, so even very bad characters have a small chance of succeeding at very difficult tasks. So, you would have seven degrees of diffculty that are translated into seven different dice, of which you roll three. So the main difference between opposed and unopposed actions is that all three dice are the same in unopposed actions, while they may differ in opposed actions.

[smascrns]

As for 4, if there is even a remote chance of success against any opponent, you will have players insist on rolling.
Yes, at least as long as the costs of insisting to roll are not too high ("Ok, you can roll but that means you will stay there facing the giant that seems able to kill you with a small and light slash. Otherwise you can opt to run as fast as you can...")

What's more my system will have both random and non-random mechanics with statistically similar results. This means that it is clear for the players when the non-random result would be certain failure. In this case the GM can decide that it's a case for non-random action resolution.

Symmetrically, there should be tasks at which you can't fail if you are sufficiently skilled (and not otherwise impeded).
It will also be part of the system.

Now, there are several mechanics that can achieve your wishes. Given that you have attribute, skill and experience field all at a seven-step scale, adding them up gives us values from 3 to 21 (or -9 to +9). That probably rules out dice pools where the number of dice equals the value, as you quickly get to an unmanageable number of dice.
I'm actually going to use a dice pool, the same I've been working with for a long time in my columns. It has the advantage that it has a very straigth interpretation of results and that it reduces the number of dice that need to be rolled compared with plain vanilla dice pools. Furthermore I'm working the rules system in terms that the size of the dice pools will hardly go above 6 dice (this is the number of dice equivalent to an automatic success or failure in non-random resolution). Nice list of alternatives, though.

[Aranwyth]

>Second, skill category modifiers hardwire too much how attributes influence >skills. Let me give you an example, the Jump skill. It is subject to the Agility >modifier that is based on Dexterity, Strengh and Size. So far so good, but what if >instead of actually jumping you want to teach jumping?

It would be capped by a teaching skill.

[smascrns]

>Second, skill category modifiers hardwire too much how attributes influence >skills. Let me give you an example, the Jump skill. It is subject to the Agility >modifier that is based on Dexterity, Strengh and Size. So far so good, but what if >instead of actually jumping you want to teach jumping?

It would be capped by a teaching skill.
It would indeed. But then you may have someone that has an high value in the Jump Teaching skill and yet does not know how to jump (notice, I'm speaking about knowledge, not about ability to use that knowledge; some people may know how to jump but can't do it, like a former athlete that suffered a disabling accident). On the other hand, you may have people with an high Jump skill and nonexistant Jump Teaching skill.

This may make some sense in the contemporary world where we created a separation between teaching and execution (but even then people tend to think that the best teachers are those that are both good at teaching and have a lot of knowledge/experience in execution; how would the system work out this coupling, then?). On the other hand, in a pre-modern world like the one I'm desiging for Rough Quests that break apart is not there.

And there's another thing: What you propose basically amounts to doubling the number of skills. For each "executive" skill there is a correspondent "teaching" skill. What about researching the skill in order to find new usages? That's also a different ability. According to your take on the matter there should be a third set of skills, the "researching" set. I would not like to have this multiplication of skills at all.

No, I'll stick to my proposition: A single set of skills that can be used in different ways, each way calling for the intervention of a different attribute.