In randomized action resolution the ability has to tie with a randomizer in such terms that the probability of success should reflect the level of the ability. Of course, this is dependent on the way abilities are quantified and on the randomizer used in the system. That's what we want to settle in the case of Rough Quests. As always, it is interesting to see how other games handle it.
Ability plus random value. The die or dice used may vary a lot but the more common are perhaps the d20 (D&D and derivations) and the d10 (Ars Magica, for instance). I don’t like this for a simple reason: I can never figure an interpretation to the randomizer. What does it stand for? What does it represent? I can’t tell. Furthermore, the link between the range of values of the abilities and the range of values of the randomizer are also poorly defined in most cases.
Ability vs. random value rolled with a single die. Mathematicaly there's no big difference between this option and the previous one. On the other hand, I find it much easier to figure what the die stants for. Take Pendragon, for instance. Your ability is a fraction of 20 where 20 stands for maximum ability and 1 corresponds to the lowest ability. The die provides random fractions of 20. Things glue together very well. The range of values provided by the die also corresponds to the range of ability values, so there's a perfect match between the two. Simple, effective and clear. (Of course, game systems can twist this. That's the case with RuneQuest that allows for abilities above 100. In this instance an excellent core has been defaced by a poor addition.)
Ability vs. random value generated with the sum of several di(c)e. We are in Hero, Tri-Stat, and GURPS territory. A lot of people prefer this alternative to the previous one since it provides a bell curve. I just can’t see the advantage to a straight die roll. Yes, it provides a bell curve, but so what? Besides, stright roles also have an implicit bell curve if we look at the character. I mean, a character is not created to perform once: he has a life and in the course of that life he will hopefully perform plenty of actions. Suppose that he does 100 possible things that in game terms require 100 die rolls. It's easy to see that the game system that uses 1d20 instead of 3d6 still incorporates an implementation of the bell curve, not at the level of a single roll but from the point of view of all the rolls that are to happen in the course of playing.
Ability maps into dice pool, count successes. I like dice-pools provided they keep within manageable sizes. The issue is in how they are to generate the outcome of the dice roll. Counting successes is an option I would only consider if it was as simple as in Prince Valiant, the original Storytelling game. Still, I’m not happy with. In most success counting systems one is able to get as many successes as the size of his dice-pool. Now, that means that the smaller dice-pool cannot achieve the same number of succeses as the bigger one. I like games where the less able party can perform as well as the more able party, no matter how slim the chances may be. Success-counting dice-pools don't allow for this.
Ability maps into dice pool, retain value rolled. If the reader has been following my RPGnet columns he will know that I'm partisan towards this type of dice-pools. I just need to point to what I said when I wrote about dicing conventions more than 6 years ago, and my unfinished games The Travels of Mendes Pinto and GlovE.
... and for Rough Quests we will have ...
In the previous paragraphs I narrowed my choices into two: Ability vs. random value rolled with a single die; roll dice-pool and retain one value rolled. The final question to ask is which of these will I keep for Rough Quests.
We know by now that Rough Quests will have abilities that are composed by attribute plus experience plus skill. We also know that each of these will be quantified into scales with seven values. Furthermore we have seen that they are going to be combined by a simple addition. So abilities in Rough Quests will have a range of 21 values. Interestingly enough this maps fairly well into a d20: Just suppose that attributes range from 1 to 7 while experience and skill range from 0 to 6. This gives us a total ability range that goes from 1 to 19 ...
If all we need to consider is how to map abilities into the range of values in a die roll, things would be settled by now. Unfortunately we also have to account for situational modifiers, differences between species of creatures, and the nuances of different types of outcomes for the the performance. These cannot be directly mapped into a range that goes from 1 to 20. Yes, I have a way to deal with it, a way that keeps the possible valus within that range and is mathematically sound. It’s based on marginal increases or decreases of ability due to the factors just mentionned. The problem with this solution is that it either requires complex (in game terms) calculations or a table. I want something simpler.
And that leeds me to my old dice-pool proposal, the one I referred to some paragraphs above. It provides exactly what I want: A direct mapping of the abilities and other factors afecting performance into the randomizer (just add all factors and, voila, you have the dice-pool); a direct reading of the values rolled (just pick the highest/lowest value rolled); a direct mapping of the values rolled into the level of success in performance.
By now things are fairly settled on how Rough Quests is to handle performance. Next column we will look at how it generates results based on the level of performance and other factors.
