Vko6)Ygbyp8͚dE(a0(U~a%%ٲiG0){x2!%>K&#: )H)4CiI̒yB%꛹Ҋ$(23;}HHnQ|`ITg /$\P0ngԱ:sde>?DA?/ꋳxy[,f#/Y7A#4rRчǦX,j< j/̅ ݗOgj ȿ,<24 "jD+g] HŖG\@vC }D{fiۮFײ0TQBnmpPbitҬԓ$1-kh,1]k$\mkٶ4OcD"` 3 BnQhgt9KlaIuO̗aeR>h&RiEPݲ~M9D]x4DF<뺥0xŸv۶{=ѝ>tGB4Huǀ} 8N'qY1QbPaFg7dDQ-`H.G D1n[ġܠoyP8i_]'Ab֞AK;4@uB Xeբ慾My QN)e"< "|tcџ\WvMcoS+@4 yD4̕„#U.k]sJ>aR`.""+5t }m8g"%KF7=9<Ȗ'y{V,U}"A|Ëf7^XV+.ܶZNguبL]{sX.KtQ~ Z6?;y]2e20U2zw@PȍB*F:^.7]WT%oc> {v %G{(I%~'mOdݖ+[l!fFݏ@4&M\|\a?++7!sOOӚх0,wП spM jFB0]V)Dxf'aNR>>:R/~ƩxOJV.ؔ 0gm|WiX7U, fϜv+SNw:Rwo缌isew,(nMۛ[UTaՊc=hp$*7bE'j4grgTwی%q.

A Change of Course




Oh well. This is only my second column and I'm already going to change its format. It just happens that RFF#1 highlighted some issues I did not anticipate:

  • The original format looks a lot like a game review, something that's completely different from what I want this column to be. In fact, some of the people that were kind enough to read RFF#1 did mistake it for a review of Alternacy.
  • Like someone pointed out, it looked like I was trying to provide alternate rules for a particular game system, instead of trying to discuss how to rule about a given role playing domain in a more general way.
  • There was no direction on where I wanted to take my discussions. Since what I would discuss would depend on the game system I would select for the column, I would be jumping randomly from an aspect of role playing to another aspect of role playing.

Given the above shortcomings, I decided to change my approach. From now on I'll select an aspect of role playing, and I'll discuss different ways to rule about it. I'll base these discussions on some games, usually selected from the Free Games we can access through RPGnet. For reasons that I'll explain in a moment, my discussions will usually start with Chaosium's Basic Role Playing rules (BRP). This may seem odd, since there's no where to find a free edition of BRP. In fact, BRP was edited in the 70s and early 80s, and since then it has been out of print. Yet, BRP underlies most of Chaosium's games, including Call of Cthulhu, and influenced so many games produced by other companies that I'm sure I can bet that most people are familiar with its basic concepts.


I also thought that I should say something about my life with RPGs, so that you may understand where I come from, and what shapes the way I look at RPG rules. (What prompted this was a recent message from Sandy asking me to send him my bio.) Here it goes:

1983: In a trip to France on holidays I was able to buy my first RPG. Based on the reviews in some French magazines I decide to buy... no, it was not that game, it was RuneQuest instead (which included as a bonus the above mentioned BRP, on which RQ is based). For the next years I played RuneQuest, and experimented other games, including CoC, AD&D, Paranoia, Bushido, TOON, Pendragon, Stormbringer + Hawkmoon, Elfquest, Ringworld. Of all the games I had the chance to read and play, the ones that impressed me the most besides RQ were: Skyrealms of Jorune (beautiful game with excellent rules, unfortunately out of print), and Malefices, an excellent "classic horror" French game, as good as CoC (also no longer in print, and also based on BRP). Yet I always went back to my first love, RuneQuest. As you can see, most of the games I mention above are either based on BRP (in fact, most of them were even produced by Chaosium), or are more or less distant cousins of this system (that's the case of SRJ and Pendragon, for instance). In fact, I still think that BRP is one of the most brilliant game engines ever created.

Not long after finishing my Law studies (which was ten years ago) I stopped playing, and for some years I had no RPG activities. What changed that situation was the net. In the last couple of years I got back to my hobby, getting as much info as I can in the net (may I mention the role played by RPGnet in this?), participating in discussion lists, buying or downloading new games, playing online. And I've been writing my own set of rules.

As you can see, there's good reason why I'll use BRP as the starting point to most of my discussions about rules. In fact, BRP shaped to a great extent the way I think about rules, and even if the way I conceive the "perfect" rule system today is very different from it, most of the time I got where I got while trying to find better alternatives to what BRP has to offer.



Well, now to the meat of the present column. As you can see from this month's column title, I decided to deal with dicing conventions. Since I'll try to deal with rules in a structured approach, moving from the basics of rules design to deeper levels of detail, I'll start with dicing conventions which are certainly among those things that we can consider "basic" about the rules of any game. First, let's clear some ground:

I'll not discuss whether there should or should not be any sort of dicing convention. Yes, I know this is disputable, I know that there are diceless systems (which may be diceless because they are randomless, or which may be diceless because they use other randomizers different from dice (n1)). I simply think that any RPG should present alternatives for random decision, alternatives that the players may drop if they decide to go randomless. And I think that dice are still the best way to introduce randomization in RPGs. After all, who doesn't know how to use a die, at least in Western culture?

I'll not discuss whether there should be one or one hundred different dicing conventions in a game. I have a bias here. I think that the simpler a game system is in its core mechanics, the easier it is to be learned and played, and the easier it is to develop rules that cover a wide spectrum of gaming situations. Anyway, most games have what we may call a "core dicing convention", the dicing convention that underpins how the game deals with the core gaming situations. This does not excludes the possibility of other complementary dicing conventions, of course. Given that, for simplicity sake I'll only discuss core dicing conventions.

Now, if we are talking about a core dicing convention, we must also ask ourselves what we use it for. In other words, if we have a core dicing convention (CDC), its because we also have a "core resolution principle" (CRC) that requires that CDC (or a set of resolution principles. Once more, I've a bias towards a single CRC, or to a small number of CRCs: why have ten different ways to deal with situations, when one ore two are enough?). A resolution principle is basically the rule that defines what, in game terms, is used to deal with game situations. It is how the game shapes the game world reality, the language and concepts that are used to describe it, and the way that this allows players to deal with the situations faced by their characters. It is the "how" and "what" of the game. For instance, BRP uses attributes (or characteristics) and skills. xD&D uses attributes, classes, and class descriptors. Since today I'm not going to discuss resolution principles, only dicing conventions, I'll just call the resolution principle an ability check. This simply means that whenever a character wants to perform an action, there's a corresponding ability that must be tested through rolling dice. (I'll leave the discussion of CRCs for future columns: after all, a single dicing convention can be used with different resolution principles, and a single resolution principle may be used with different dicing conventions.) index

BRP's dicing convention

RuneQuest used the percentile dice as its core dicing convention. This was complemented with other dicing conventions, but the d100 was the core. It inherited this from BRP. I guess we are all familiar with this:

  1. The resolution convention represents the different aspects of the game that may require random resolution as a percentage of ability to succeed;
  2. When making a fuzzy action, the player rolls 1d100 (in fact 2d10 where one of the dice is the tens and the other is the units, and where 00 means 100), and if the value is =< then the ability value, there's a success, while if the value rolled is above the ability, the action fails;
  3. There are qualified successes (special or critical) and failures (fumbles) which are a fraction of the ability to succeed (or of the possibility of failure) (n2) - suppose you have 60% chance to succeed in Elric!; you have a special action if you roll equal or below the lowest 10% of 60%, in other words, if you roll 1-6; but you may also fumble if you roll the highest 5% of the probability of a failure, in other words, if you roll 99-00.
  4. There's always the chance that a character with top ability may get a really bad result (fumble), and even a character with the worst possible ability can have a top result (critical).

This is a very elegant dicing convention. It is simple, easy to learn and memorize, intuitive. Yet, there are some drawbacks:

  1. I prefer even simpler conventions. I tend to prefer the d20 to the d100. This is another of my bias: the result should be readable on a single die, not in a combination of dice. But the way BRP deals with qualified successes makes this impossible (n3);
  2. The quantitative scale of results is not fixed. What I mean here is that, say, in RQ we know that there are five qualitative results, critical, special, normal, failure, fumble; but their quantitative thresholds will vary among different basic values of the ability (for 60% ability we have: crit<=3, spe<=12, norm<=60, fum>=99; for 45% ability the thresholds are: crit<=2, spe<=9, norm<=45, fum>=98). This is very hard to account for if abilities change in the middle of play (yes, there are tables with pre-calculated values, but that means we have to refer to those tables). What's fixed is the randomizer - we always roll d100;
  3. It strongly directs us to represent abilities in purely quantitative terms. I like to have abilities represented with qualitative scales, instead of quantitative scales (that's another of my bias)(n4)(n5);
  4. Its not very consistent the way it deals with factors that may affect the difficulty of the action and that are independent from the ability (in other words, with difficulty modifiers): we cannot deal with these as thresholds, since the resolution principle is focused in the ability performance; we can use fixed modifiers to the ability (+/- n%), which is very simple, but produces some strange and eventually highly distorted results (what about 30% ability with an -30% modifier?); we can apply relative modifiers (modifiers that are a % of the chance to succeed where chance to succeed is equal to [ability * +/- mod]), but that means that we are going to face constantly the problem mentioned above about changing the value of abilities in middle play (I'm personally not willing to use a calculator to determine the chance of success for each action my character's going to make).

All in all BRP presents us with a strong dicing convention, but its shortcomings should prompt us to search for an alternative dicing convention, one that keeps its advantages, but that avoids its disadvantages (n6). Let's do it, let's look at a couple of other dicing conventions. index

Other dicing conventions

The dice + ability + modifiers convention: In this case the result is equal to the sum of the value rolled in one (or more) die, the value of the ability, and situational modifiers. The final value is compared to a difficulty threshold. This is the convention that we may find in games like Alternacy (that I discussed in my first column), Fuzion, Luck 13, Ars Magica (n7). It is simple, but has several limitations:

  • It doesn't comply with A, since the final result is based on an aggregation of factors, and has the effect referred to in C.;
  • It may happen that a character with very low ability will never achieve top results if the sum of his ability, modifiers, and maximum rollable die is below the threshold for those results; conversely, high ability may protect the character from very low results if the value of ability is higher the threshold of those results. This means that such systems may not comply with point 4. The only way to avoid this is through "open ended" rolls. And here we hit another of my bias concerning dicing conventions: I just don't like open ended rolls. Here's why:
    • Open ended rolls exist to allow characters to achieve extreme results; but they do this without any direct relationship to ability; the fact that a character may achieve those results is linked only to luck while dicing; so, systems with open ended rolls don't apply point 3;
    • Open ended rolls go further against point A; they introduce more and more levels of dice rolling and computation.

A convention that's close to BRP's is the one that we can find in games like TOON or GURPS. These games work basically like the BRP system, meaning that we have to roll dice, and if the value rolled is <= the value of the ability we succeed. What makes them different from BRP is that those games don't use the d100, but the sum of several d6s (2d6 in TOON, 3d6 in GURPS). So, they're subject to some of the remarks above, specially A. and C. (n8).

The Window (n7) dicing convention. In it each level of ability is assigned a different die. The highest level is assigned a d4, the next level a d6, and so on (in other words, it assigns lower dice for higher ability). Each action is assigned a difficulty threshold, and in order to succeed the player must roll a value on the ability die that's lover than that threshold. I must say that I like this convention. It deals with point A, and to a certain extent with point C (n5). Unfortunately it fails on what concerns point 4: Yes, any character may get top results, and the chance to reach top results is linked to ability. But it does not work that way in the other sense: A character with high ability is protected from very poor results, since the threshold for those results is above the maximum rollable value in, say, a d4 or a d6. (A system that's close to The Window is Earthdawn (n7).)

Well, since it seems that none of the dicing conventions considered so far allows us both to retain the advantages present in points 2 to 4, and to deal with the disadvantages in points A to D, we will need to consider other options. index

My own dicing convention

When I started to think about alternatives to the basic BRP dicing convention, I followed more or less the next path:

I reversed the reasoning behind the quantitative scale of results, while keeping the qualitative scale. In other words, I designed a qualitative scale (Critical, Special, regular, marginal, failure, fumble). I made a correspondence between these qualitative levels, and a fixed 20-levels quantitative scale: Fumble = 1; Failure = 2-7; Marginal = 8-10; Regular = 11-16; Special = 17-19; Critical = 20. The result of an action is the value rolled on a d20 compared to this scale.

This raises a question: how to represent abilities? It must be in a different way from what we find in BRP. Inspired by Skyrealms of Jorune (n9), I defined levels of ability (an odd number of levels, for reasons I'll explain in a moment), each level corresponding to a certain chance to succeed. But, while SRJ assigns a success percentage to each level, I assigned to each one of them a certain number of d20s to roll. The number of dice to roll is based on the next conventions:

  1. There's always at least a d20 to roll. This is called the basic die, or BD;
  2. Middle level ability (called standard ability - statistically, the average ability) corresponds to the BD: If the character has standard ability, he only rolls the BD;
  3. If the character's ability is sub standard (lower than average), he receives one more d20 for each level below standard ability; this is denoted by +nd, where n is the number of levels below standard, and d denotes the fact that it is sub-standard;
  4. If the character's ability is above standard (higher than average), he receives one more d20 for each level above standard ability; this is denoted by +nD, where n is the number of levels above standard, and D denotes the fact that it is above standard;
  5. If the ability is composed of several character descriptors (say, an attribute and a skill), we just add the corresponding dice, keeping in mind that +nds and +nDs cancel each other (suppose that ability is based on two descriptors, one is +2d, the other +3D; the +2d reduces the effect of the +3D to +1D, which is added to the BD, giving a final number of dice that's 2D);
  6. When rolling +nd, the player picks the lowest die rolled as the result of the action;
  7. When rolling +nD, the player picks the highest die rolled as the result of the action;
  8. There are as much lower levels than higher levels, so the number of levels is an odd number (lower levels + higher levels + standard level = an odd number);
  9. If we want to introduce modifiers to the difficulty of the action, either these modifiers make the action easier to accomplish, or they make it harder to achieve; in the first case they are represented as +nD, while in the second case they're represented as +nd; we just have to introduce these to the calculus of the total dice to roll. (So, does that mean that there are no automatic, or impossible actions? No. The GM may always rule that an action is either automatic or impossible - for one or more characters. In that case there's no need to roll for success, that's all.)

How does this system compare to the points raised above about the BRP dicing convention? Well, we can see that:

  • It reverses the reasoning behind point 2: instead of having a fixed randomizer, and variable quantitative success levels, we have fixed success levels, and variable randomizers;
  • It keeps characteristics 3 and 4, since there's a set of different results, and any character may perform any of those different results (even if the chances to do it vary according to ability);
  • It complies to point A, the result is readable on a single die, and it's a d20: the player may have to roll several dice, but he only needs to keep one to calculate the result (either the higher or the lower roll);
  • It addresses the issues raised in B: the quantitative scale of results is fixed, so there's no need to introduce complex calculations whenever there are changes in the value of ability descriptors;
  • It also deals with point C, since we can represent abilities in qualitative terms: just assign a qualitative concept to each level of the ability (5);
  • As we can see in i, this system deals in a very simple way with the problem mentioned in D.

The curious thing is that I developed this system on my own, but other people were already using dicing conventions based on similar principles. In fact, this type of conventions became collectively known as dice pool systems (n10).

For instance, let's consider the famous Storytelling system (SS), developed by White Wolf (n11). They use the next dicing convention:

  • Ability is represented by a pool of dice (d10s) assigned to a skill and an attribute; that means that ability can vary between 1d10, and 10d10;
  • There's a threshold of difficulty assigned to the action; the result is equal to the number of "successes" rolled: the results that are higher than that threshold;
  • The nature of success varies with the number of successes rolled: 1 is marginal, 5 is exceptional (my terminology; not necessarily WW's terms, since I'm describing from memory).

How does this compare to my convention? Let's look at it in detail:

  • The first thing we may notice is that my convention is more economical on dice; for instance, a 5 level scale requires 5 dice in the SS, while it only requires 3 dice in my system; this makes it easier to use more granular ability scales;
  • The SS goes counter point 4: If you have a level three ability, you can only roll a maximum of 3 successes, which means that you cannot roll the whole set of possible outcomes. Unless, of course, we add further rules (forex, rules for open ended rolls, but these are subject to the points raised above) which goes counter point A;
  • The fact that I make a distinction between +nd and +nD is an added complexity when comparing my proposal to what we find in the SS.

Finally, there's a difference between the two systems on what concerns the respective experience curves. But that's something that I'll discuss in Ruleslawyer for Free # 3. index


I tried to find a dicing convention in which:

  • There's a scale of different action results;
  • Any action may fall in any of the levels in the scale of results;
  • The result is read in a single die, and requires no complex computations (no more than very simple sums);
  • The average quality of results will vary according to the ability of the character;
  • There's an easy way to introduce situational modifiers;
  • Allows for both quantitative and qualitative representation of abilities;
  • Abilities evolve along a well defined experience curve.

I think that my system allows for all of those characteristics. (Of course, some of these need to be proofed, something I can only make in future columns.) To make things simpler, I'll give it a name, so let's call it Double Sense Dice Pool convention (DSDP) (n12).

Given that, I eagerly look forward for your comments.



(n1) By the way, I include Fudge's dice among the "non-dice randomness" concept. Yes, they are dice, but they are not standard dice. r1

(n2) Chaosium used more than one alternative to deal with qualified successes. In RuneQuest we have both special and critical successes, where a special is 20% of the ability to succeed, and a critical is 5% of that ability. In Stormbringer/Elric!/Hawkmoon there's only critical results, which are 10% of the ability to succeed. Yet, I think that the best take on qualified successes based on BRP was not developed by Chaosium. It can be found in a French game that can be accessed through the net (unfortunately, I've lost the link to it, and don't remember its name). In this game there are special, normal, and marginal successes. A special is the lowest 10% of the ability to succeed, while a marginal is the highest 10% of that ability. r2

(n3) In fact, if it wasn't for the qualified successes rules, we can easily convert the % scale to a 1-20 scale, and use a d20 instead of a d100. That's exactly what Chaosium did with Pendragon (a game that's of a less pure breed of BRP). r3

(n4) True, we may super impose a qualitative scale above the quantitative one. That's what we find in Skyrealms of Jorune (and even in xD&D, if we consider the labels they assigned to character levels). In fact, it was the qualitative scale in SRJ that prompted me to think about abilities in qualitative terms. The problem is that this scale is marginal to the way situations are solved, so it is easily dropped by the players. r4

(n5) We will discuss this issue in greater detail in a future column. r5

(n6) There's a further problem with the BRP dicing convention, a problem that relates to the shape of its experience curve. I'll discuss this in my next column. r6

(7) These games can be found at the Free Games section of RPGnet. r7

(n8) I base what I say about TOON on the game as published, and what I say about GURPS on GURPS Lite, which can be found in the Free Games section of RPGnet. Also notice that, as I'll explain in my next column, these game systems avoid the problem with the learning curve I mentioned in note (n6). r8

(n9) In SRJ's basic resolution principle skills have 12 levels, and each level corresponds to a defined % of success (which may vary from skill to skill); these 12 levels are aggregated into a smaller number of qualitative levels. r9

(n10) I own no game which uses dice pools, but there's always the possibility that I was unconsciously influenced by reviews of such games, or by accessing discussions of such dicing conventions in the net. r10

(n11) To tell the truth, I own no WW game. I have no interest in their settings. It just happens that I was able to skip through Vampire: The Masquerade in a book shop. That's where I got knowledge of what I describe above. r11

(n12) I don't claim to have designed an original dicing convention. I seem to recall reading recently - either in a discussion list or in a review - about a game that uses a similar convention. The only thing I can claim is to have developed this convention without external inputs. r12

What do you think?

TQo0~^DҒt< ek&Ǿ$\۵ZFȃuwݝIŃU QYir2HR2.u3MFoعq]4#A`pP5(b& )b)ⰾp7(i<[-2gL#5[f g?*rVGf8*)s'+20ϟ̑F}KB<7wSL\gbvm9WiRބYŜvd y0'p2I_Fc2>#o A )VL[Qk?3`)<У[(*W.JH ?tXCt谙 X:@ \0w ~LqĤE-rFkYœj4q 5AQ6[AxG [>w|?( fХθY䝛$c=_qNĦoǸ>O_|&/_Mi7"宥CЧk0dӷLh;TmuCGU-!Ul{ h<\bQX.~"O2*yPcz!ŠGg [an error occurred while processing this directive]
[an error occurred while processing this directive] TQo0~^DҒt< ek&Ǿ$\۵ZFȃuwݝIŃU QYir2HR2.u3MFoعq]4#A`pP5(b& )b)ⰾp7(i<[-2gL#5[f g?*rVGf8*)s'+20ϟ̑F}KB<7wSL\gbvm9WiRބYŜvd y0'p2I_Fc2>#o A )VL[Qk?3`)<У[(*W.JH ?tXCt谙 X:@ \0w ~LqĤE-rFkYœj4q 5AQ6[AxG [>w|?( fХθY䝛$c=_qNĦoǸ>O_|&/_Mi7"宥CЧk0dӷLh;TmuCGU-!Ul{ h<\bQX.~"O2*yPcz!ŠGg