Fidelity Vs Breadth

granularityI’ve been thinking about fidelity and granularity in RPG systems again, specifically the virtues of  high fidelity, high granularity systems (such as, say, Basic Roleplaying) versus broad-stroke, low fidelity systems such as FATE.

For a long time the trend has been towards simpler and simpler systems.

Low fidelity systems are – generally speaking – quick and easy to play and fit better into the increasingly rushed schedules and limited time of an ageing audience for RPGs. There’s less note taking, less dice rolling, combat moves more swiftly and unified systems make the games easier to learn, all positives.

On the downside, low fidelity systems mean that there’s little granularity. A small difference in the game’s statistics makes a big difference to a result. If your game is based on a 1-6 roll, for example, then a +1 bonus makes a difference of almost 17%, which is huge. Low fidelity systems also tend to lack meaningful capacity for customisation and for differentiation between one character and another. They also tend to have limited capacity for growth, meaning that long term games tend to break down much more quickly.

High fidelity systems with high granularity allow for high degrees of customisations and meaningful – if small – differentiation between characters. In a percentile system – for example – a single point bonus is only a 1% difference, which is probably as much of a scale as you’d want to go for. Characters can fine tune their equipment and situations to squeeze out every possible bonus, however small, and see a return on that investment in results. For tech-heads, car fanatics and gun-nuts this kind of fine tuning can be immensely satisfying and knowing that recalibrating your sights between each combat makes a real difference can also encourage players towards more simulationist, ‘realistic’ play.

The problem with high fidelity systems is that they tend to be slow and fiddly. Even basic arithmetic slows down with larger numbers and the more realistic the system the more there is to remember, the more special conditions, special rules and so on you have to worry about. This can happen with low fidelity systems too as special conditions and abilities proliferate, but it is more of an inherent problem within high fidelity game systems from the get go. Even without any special additional rules or exceptions.

Another issue is that we’re bound to our dice. If we want a realistic probability curve then, ideally, we want to use two dice and to combine their numbers.

EG: 2d6 will produce the following curve of results.

400px-Dice_Distribution_(bar).svgWhile a single dice roll returns a flat probability level.

In general a single dice is a good bet for a game of heroic adventure where you can fail or win big at the roll of a single die, while a combined dice roll is better for more gritty and realistic games. FATE throws a spanner into that theory, but that’s more about characters being reliably awesome at what they do I think.

High fidelity comes with using large dice, but the largest dice available to us is the d100 and using combined results at that level simply isn’t practical. Computers are simply better at handling that kind of thing invisibly and out of the way while the usual role-playing, tabletop situation retains the edge for more improvisational and less fiddly gaming of a more narrative and heroic bent.

Given that we now live in a world where everyone is carrying computers around with them all the time, are we truly limited by the shapes of physical dice?

Could we use random number generators of custom values to roll and combine ‘dice’ with any number of sides to give us a truly random range?

Suppose we had a system with infinite, potential, range but based around a 1-100 scale and where a score of fifty was needed for an average success.

Suppose Joe is at the firing range with a 9mm pistol, but has never, ever used one before. He has a Hand-Eye of 62/100 and a Pistol Skill of 5 – he’s only played a few video games.

ce7b03660042063f0e52ec7e046f7ec5The Glock 17 he’s firing has an effective range of 50 metres (+50 skill) and he’s shooting at a range of 15 yards (-15 skill). He’s taking his time to carefully aim (+6 stat) and is using the iron sights (+1 Skill).

His final rolling dice are d68+d41. The results? 4+30=34. A miss. He manages, somehow, to miss the target completely. Probably clipping the edge of the target.

He takes a breath and takes a second shot, with instruction from a firing instructor with a skill/stat combo in teaching of 100, giving +10.

This time he’s rolling d68+d51. The results? 20+25=45, a miss again, but better.

A third shot. d68+d51. The results? 51+47=98. A good, solid hit, good enough to hit the ‘X’ in the centre of the target – though more by luck than judgement. His average, reliable shot at this range would be around 60. Good enough to hit the ‘8’ band on the target with some success.

For high fidelity systems, the possibility of rolling a dN and doing it in combination, could allow for a renaissance in systems with granularity and where small differences in weapons, equipment and training can matter. The trick will be stopping it getting too complex and creating paperwork that counters the possible advantages.


One response to “Fidelity Vs Breadth

  1. There is only one question that is meaningful in this context:


    And it goes all the way back to your original choice of descriptor, “Fidelity.” You choose high fidelity versus low fidelity, but fidelity to what? Real world results? Narratively appropriate results? Some Platonic ideal I have yet to imagine? You’ve already loaded the discussion by choosing “fidelity” as the descriptor, and I think I object up front to that being the implication. After all, what does a “low fidelity system” mean? That it doesn’t, by design, aim for results which are in accordance with a certain intent? Who would do that? “Low fidelity” is a descriptor that we can attach to a system design after the fact, and generally only pejoratively. So that axis isn’t even an issue for our concern.

    What we’re left with is granularity, and granularity is a huge concern, especially for management’s of the only important currency that is exchanged between players in a game: decisions. Good games were once described as “a series of meaningful decisions,” and I don’t think that is truly up for debate. There are 2 parts that statement, firstly that players make those decisions, and secondly that those decisions be seen by the players as meaningful. As granularity increases, decisions about increasingly granular portions of the system decrease in meaningfulness because they are overwhelmed by random result.

    For example: in your d100 – d100 system, look at your modifiers. Very few of them represent meaningful choice or situational creation by the player. Those that are, are overwhelming magnitude by a 200 point random swing. That’s horrible. There’s no other way to put it, that is a bad setup. The player has no meaningful input into that conflict resolution.

    Which leads to another problem with systems of granularity which are that extreme: they require that you pile on modifiers, that things that adjusts the results are accounted for in great detail, because they invite it. Every one of those modifiers that is outside of the players control – and I should probably add “immediate control” – takes the player further, and further out of the loop. They also increase bookkeeping, overhead, random mental cruft, and generally just make things un-fun.

    Cognitive research has shown that the human mind is really good at juggling somewhere between 5 and 7 factors at a given moment. 5 to 7 things are generally enough traits/descriptors/situational elements for human beings to deal with and feel like they have a grasp of at any given time. The last thing you want to do if you want human beings to play your game, and to play it meaningfully, is to go beyond that level. Increasing granularity does exactly that.

    There’s a reason that game design has been moving toward ever simpler means of representation overall for the last couple of decades, it’s because it’s better. It’s simply better. You can get a better gaming experience which is more focused and better representative of whatever narrative the designer had in mind with a simpler, cleaner, less granular system. It’s really that simple. It’s not even that systems have become “less complex” in the sense of books for the players to get their hands on and manipulate, because there are many very popular games which are extremely granularity/die simple with very complex interaction profiles. PDQ, FU, even Fiasco – which is, at heart, based around very small d6 pools, they all put the complexity over into the decision-making space and keep it out of the die mechanics.

    One of the biggest examples of highly (in my opinion, ridiculously overly) complex rules with fairly straightforward and narrow granularity is Hero. You know, the role-playing game which you can use for anti-HEAT armor on your personal tank. It’s huge, it’s ridiculously complicated, but it’s built on creating small pools of d6s (and sometimes large pools of d6s) with differing traits to resolve. ORE, while not as complex, does the same thing with the d6. The range of the d6 is one of its strengths, adding one d6 to a pool of less than 10 creates a meaningful but granular result. Giving between a +3 bonus and -3 malus to a d6 roll (or 2d6 roll) falls right into that number of elements that human minds find comfortable to work with and provides a small enough granularity that their decisions, their choices, and their positioning are actually, immediately detectably, relevant to the outcome.

    In a real sense, a d10 random element is the largest you will ever, in a reasonable sense, ever need. Most d20 systems are constructed such that the random element of resolution overwhelms most of the choice component of a given resolution, which is fine in the context of a much larger wargame – as D&D was constructed to be – and the loss of one element can be extremely random because you are expected to have a lot of those singular small decisions occurring at any given time for immediate, accumulative, significant decision results, but it’s just a terrible thing for most RPG design today. The FUDGE design, with 4dF is in a particular sweet spot when it comes to meaningful decision-making, and the way that it is rooted to your expected outcome and varies outwards from there is quite nice. That probably explains why it is been an extremely popular choice of mechanic.

    But we’re back to my opening question: why? Why do you need extreme granularity just because you can? As a model, which is all that RPG mechanical systems truly are, what is it that extremely high granularity is supposed to bring you, and more importantly bring the players, that is not better brought by increasing the amount of power that their decisions have directly? I think there’s even a case to be made that mechanical random granularity is probably an orthagonal methodology to your intended result. In fact, going the other direction might be quite profitable.

    Consider Kingdom, a game with no dice in all, wherein the players take on one of 3 powerful narrative roles, and they undergo challenge which depends totally and completely on their individual intent and decision-making, not just individually but as negotiated between them. I would put forth that such a game is far more nuanced, has a far greater range of possible outcomes, and provides a far more satisfying experience for the players in every case than a system which is based upon a compilation of extremely granular random external influences.

    I’m no anti-dice zealot, either. I love dice. But I love games more than I love a specific manifestation of one means of resolution, and I don’t believe that focusing on the granularity of that means is useful when it moves beyond a set of results that players can grasp meaningfully, affect meaningfully, and into it meaningfully.

    So – why?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s