10×10 Room

[This is part two in a series of posts on probability and roleplaying games.  You can begin with part 1 here.]

Who else remembers these?

RPG designers haven’t always looked at the role of probability in action-taking in the same way.  In fact, over time, those designers have made it more and more likely that players will succeed at their actions.

What’s the most common thing players do in RPGs?  They try to whack a baddie on the head.  So let’s start from there and see how likely, historically, you’ve been to successfully whack that baddie.  And let’s narrow it further and look at fantasy RPGs:  what happens when a fighter-type tries to smack a goblin with a sword?  How often does he hit?

If we jump in the Wayback Machine and head to the days of 1st Edition Advanced Dungeons and Dragons, said fighter’s chances ain’t looking so good.  A 1st-level human fighter with Strength 16 – average for a fighter using the recommended method for rolling character attributes – has only a 30% chance of hitting a lowly goblin with his sword.  That’s right:  just 30%.  Now, there were a lot of quirks to AD&D’s system are worth exploring in their own right, as they show other ways in which the genre has evolved; for example, all 1st-level characters had exactly the same chance to hit our poor goblin, but fighters were vastly more effective at higher levels, which is very different from how most modern games handle level progression.  Still, for now, we’ll just stick with our simple number.

In other words, back then, a character could be expected to fail at this core action over and over.  Fast forward to AD&D’s 3rd Edition, and the picture looks somewhat different.  Take the 1st-level human fighter again, still with Strength 16.  He’s got +4 to hit – +1 from having a level of fighter, +3 more from Strength – and he’s facing a goblin with an AC of 15.  This gives him exactly a 50% chance to hit.

Now look at the 1st-level human fighter of today.  He has a Strength of 18 if we use the standard score array and assign his +2 racial bonus to Strength.  This gives him +4 to hit from his strength, +3 more from using a long sword.  We presume he chose the one-handed combat style, for +1 more.  We assume no other bonuses from feats, and that he’s just using his basic attack, instead of a power like Sure Strike.  There’s not just a single goblin for him to face, but most goblins have an AC of 16.  Now he has a 65% chance to hit.

What we see in these numbers is a direct and dramatic climb in the chances of success over the years.  This change can’t have been accidental:  just take a look at the notes on variant rules in the 3rd Edition Dungeon Masters Guide to understand how sensitive the D&D game design team was to the impact of much more minor rules changes than than these.  The designers made a conscious decision to have players succeed more and more often at their actions.

Okay then:  why has the number changed so much over time?  I think there are four interrelated answers:

1. Success is fun.  While letting players succeed with their characters’ actions all the time takes away the benefits of dice-rolling, players will nevertheless have more fun if they succeed more often than they fail.
2. Inaction is boring.  Failure usually results in nothing happening; a miss in combat, or a failed skill check, is usually wasted time.
3. Wasting limited resources feels frustrating.  If I can only cast a certain number of spells each combat, or can only use my special power twice per day, I’m going to save it up for when it matters; when the time comes, I want it to be likely to count for something.  (You can see 4th Edition D&D take this a step further by introducing a number of daily powers that are guaranteed to have at least some effect, albeit a reduced one, even if they fail.)
4. Failing takes time.  Assuming that the rate of foe failure is similar to that of characters (not always true, but close enough), introducing more failures means that the time it takes to resolve a conflict in a game is directly lengthened by the chance of failure, without changing the eventual outcome.

Now, you can have too much of a good thing.  The previous post already explored a bit of why succeeding all the time isn’t necessarily good.  Consider also what happens if the time allotted to a conflict is compressed by a very high rate of success:  that leaves fewer opportunities for player decisions, fewer chances for tactics and dramatic roleplaying, fewer moments where a gamemaster or computer can spring a surprise on the players.

But the point is that the people in the know – the Dungeons & Dragons designers, reacting to ever increasing amounts of data – steadily hiked up the chance of success, because they saw reasons such changes would improve their game.

Now, another factor started to appear in the ’80s, and has proceeded to become ever more significant, which is the rise of computer RPGs – games which began as followers in the trends set by paper and pencil RPGs, but have since switched roles to become leaders.  More on the impact of CRPGs, and the evolution of Conclave’s own use of probability to decide the results of actions, in the next post.

A great roll... but why are we rolling in the first place?

One question a game designer must ask is how often players should be able to succeed at the actions they take.  In many games, success is automatic:  you can’t fail to use a capturing move in chess, say, or to buy a property in Monopoly.  In fact, the majority of non-electronic games are based off of automatic success.  This is not to say random elements – the roll of a die to determine movement, say, or the cards you draw from a shuffled deck – but most such games limit the actions you can take through randomness, rather than leaving up to chance whether or not you will be able to successfully take your actions.

RPGs are one major exception to this rule.  Another is wargaming, roleplaying’s ancestor: Dungeons & Dragons evolved out of Chainmail, which took the mechanics of wargaming and applied them to the swords and sorcery genre.  As part of its inheritance, Dungeons & Dragons relied heavily on dice to determine if player actions succeeded.  (In fact, early editions of Dungeons & Dragons offered options to use dice to handle almost anything you might want to do in the game, from creating dungeons to determining which of twenty forms of insanity a character might develop if rendered insane, to figuring out what might happen if you mixed a potion of invisibility with a philter of love.  More on this topic later in this series of posts.)

Later RPGs questioned the centrality of dice in the game, with many seeking to reduce randomness, and some eliminating it entirely in favor of some mix of gamemaster and player dictate.  Often this came from the desire for stronger storytelling:  both gamemasters and players rebelled at having a story shredded by a particularly ill-timed lucky (or unlucky) roll.  But most RPGs kept dice.  Why?

One reason is that dice can be exciting.  Randomness – uncertainty – creates tension and variety.  This is pretty obvious!

But a second reason is that randomness helps enhance the sense that the RPG is a simulation of reality.  RPGs inevitably seek, to varying degrees, to simulate some version of reality, some cosmos.  In the real world, we are used to the idea that our actions will not always succeed or have perfectly predictable results.  The abstraction of the die roll provides a simple path to creating the same situation for our characters and their foes.  It’s easier to imagine ourselves into the bodies of our characters when we can’t know if they will succeed or fail within the larger world.

Additionally, as simulations, RPGs must handle a huge variety of possible actions, situations, and outcomes.  The RPG must be able to provide appropriate results depending on whether your warrior is trying to hit a cowering kobold, a veteran swordsman, an ancient dragon, or a deity (!).  Probability is a great mechanism here, expanding success from a simple digital yes-no to an analog range.  Boardgames cover a vastly smaller set of situations; with a simple possibility space comes the possibility of using simpler tools for resolving actions.

So die rolling is good, right?  Hold on there, sparky!  If there’s one thing any game designer, in any genre, knows, it’s that too much randomness – too much of that uncertainty mentioned above – can spoil the game-playing aspect of a game.  Players want to have control, too; they want to feel that their skill has an effect on the outcome of the game.  RPGs are funny beasts in many ways, and one is the tension they experience between giving players the unpredictability and simulationism enabled by die-rolling, and the desire to exert control.  Mess up this equation, and your game will become less fun.

What’s more, dice rolling takes time.  Every die roll involves a wait, and potential distraction from the flow of the game as a player digs for dice, adds up the results, and the roll gets checked against various charts and sheets to produce an outcome.  Such is not a concern with computer games, but you can bet that card and board game designers worth their salt take this sort of thing into account.

Having looked at some theory behind randomness in RPGs, we can next look at how randomness has evolved in RPGs over time – and what it looks like for today’s CRPGs.  That will be the topic of the next post in this series.

On her blog Creating Passionate Users, game developer and author Kathy Sierra has often written about the need for products to create an “I rule!” experience for their users. And at their best, the web’s most popular apps do just that: Facebook makes us feel more connected, Twitter more popular, Basecamp more productive.

Those products are designed primarily for utility. For products that are meant to entertain, however, the intended reaction is less “I rule!” than “That rules!” We might get lost in the action of a movie or empathize with the protagonist of a novel, but the experience doesn’t need to leave us feeling more powerful for it to be compelling.

Part of what makes game development so challenging, I’ve recently realized, is that games can provoke both types of reactions. Intuitive and responsive controls, a steady increase in difficulty, a satisfying conclusion: these elements add up to an “I rule!” experience for players. Story and aesthetics, on the other hand, are key to provoking a “That rules!” reaction. The difficulty isn’t just that we need to worry about both goals when creating games; the goals themselves are sometimes at odds. Games that spend too much time in cutscenes, for example, risk making the player feel like a passive observer no matter how well written and rendered they are. And satisfying the player’s desire to rule often places constraints on the stories that we can tell or how we tell them: an unexpected death can be a powerful moment, but not if it causes the player to quit the game in frustration.

Considering the need for games to strike a balance between “That rules!” and “I rule!”, it’s not too surprising that some have trouble seeing games as art. Done well, a game is both art and application, art and not-art — and that’s one of the things that makes creating a good one so rewarding.