Defining "Good" Games

It's almost four o'clock in the morning, beginning the last third of a noon-to-noon LAN party. Right now I'm surrounded by a bunch of hardcore gamers and geeks playing Team Fortress 2 and a variety of obscure Warcraft 3 maps. I've decided to take this opportunity to poll this sample of the hardcore gaming public about "what objectively makes a good game". After a lot of argument about the proper use of the term "objectively", I got an interesting set of answers.

Seung said that the measure of a game is the quality of its storytelling and pacing.

Jason claimed that pure and simple "fun" was the main criterion, but replayability (offering a different experience with each play) is important.

Tom M. said that a good game includes "original" game mechanics and a strong sense of progress through the game, represented through either new gameplay opportunities or an unfolding story.

Barnett's sole criterion was value; a good game keeps him entertained for a long time.

Tom L. listed gameplay as the most important aspect of a game. His example was the game Tales of Symphonia, which he characterized as having poor storytelling but fantastic cooperative multiplayer battles.

Derek said that replayability was his only measure of a game's value. A good game keeps you coming back. He hasn't played Portal, but he speculates that it's not "worth it" unless he can play through it several times.

David told me that a good game will be either a short and difficult test of skill or a longer, easier, more involving experience.

Jack plainly listed his three criteria: challenge, fun, and creativity.

And Jonathan said that a good game is defined partially by its "polish" (how well-implemented the game is) and mostly by its "concept" (the originality and quality of the ideas underlying the finished game).

So now let's try to draw deep meaning out of a tiny and undoubtedly flawed opinion sample:

I was surprised that so many of my friends emphasized replayability or game length as a measure of how good a game is. I tend to consider replayability a matter of quantity rather than quality. While quantity might influence purchasing decisions, it doesn't affect how "good" I consider a game. I think that the disconnect might represent a basic difference in how we think about games. My vision is often informed by movies, probably because I envy film's combination of popular appeal and artistic reputation as a medium. I tend to think of more narrated games like Deus Ex or Half Life 2 or Portal as the games that are closest to realizing greatness. Some of my friends tend to think more of games like Defense of the Ancients or Starcraft multiplayer or TF2 that can be played over and over again without losing any freshness. This conceptual difference mirrors that between the proponents of art-through-narrative and those of art-through-mechanics.

Several of my friends mentioned creativity and innovation, but in general this was in reference to gameplay and game mechanics. What most struck me was that nobody mentioned anything about the type of emotion or effect on the player that the game engenders, except mentioning "fun". As Warren Spector noted, other media are applauded upon creating experiences that are distinctly not fun but are otherwise enlightening or powerful. It could be that the focus on fun is why games are not widely considered a "serious" medium. Of course, most games are intended to be light entertainment, and few aspire to be high art, but it bothers me that the possibility of artistic potential doesn't even enter so many minds.

Game Length

Portal was a transformative experience for me. It was undeniably a short game, but it felt like it lasted just the right amount of time. It got all the good puzzles in, never felt repetitious, and it didn't overstay its welcome.

Further, I had a lot of other stuff to play, and it was really, really nice to be able to suck all the enjoyment out of Portal in a few hours. It didn't require a month-long commitment to appreciate. I didn't enjoy Portal "despite" its short length, but because of it.

But, then again, my favorite game of all time is Morrowind, and I often cite the game's depth as its strongest point. I probably spent over 200 hours in the game world, learning all of its secrets and exploring all of its hidden corners, and I really don't think that anyone can truly enjoy the game without spending a lot of quality time with it.

So I'm left with the interesting and important question: how long ought a game to last?

I'm immediately tempted to turn back to my earlier post about playing for the experience of the game versus playing for the joy of exploration versus playing for skill. And, certainly, I'd expect a game that offers a lot of exploration to last a long time (like Morrowind), a game that offers a test of skill to be endlessly replayable (like multiplayer shooters), and a game that offers a tight and powerful experience to be shorter (like Call of Duty 4 single player).

But this system breaks down with some examples. The Final Fantasy series is certainly an experience game (though mostly a passive one), but it relies on its long length for an engaging story and world. Psychonauts is an exploratory "new art"-motivated game, but it's linear and could succeed even if it were much shorter. And you might be able to describe Portal as a skill-test, especially if you include the "challenge" levels.

I think that ultimately a game designer must look at exactly what his game has to offer and then pace the game accordingly. If there's not much variety to offer in the gameplay, then a short and dense experience like CoD4's is a good idea. If the joy of playing is through immersion in an open world, as in Morrowind, then create that open world and don't limit the player's time in it. If you've got a few hours worth of fantastic puzzles and hilarious dialogue, then don't try to extend the game to an unnecessary length; just make Portal!

The important thing to remember is that game length does not determine game content or game value. Let the game decide its own length, and we'll be happy with however many hours result.

EDIT: Coincidentally, Tycho from Penny Arcade also blogged about Portal and game length today.

The Best Moment in Gaming

At least, my favorite moment. It came from Warren Spector's masterpiece, Deus Ex.

My character, JC Denton, was charged with finding and killing known terrorist leader Juan Lebedev. As I charged into the airplane hanger in which he was hiding, I was intercepted by my trusted brother Paul. He had switched sides and was now working with the terrorists. "Join us, JC. Talk to Lebedev. He can convince you."

When I approached Lebedev, he immediately surrendered. "Easy now, Agent. UNATCO has a policy against killing unarmed prisoners. We have much to learn from each other." He started to tell me that the conspiracy freaks were right; the government and my cohorts, he alleged, were the bad guys.

Just then, my partner Anne Navarre barged in. She had been a little aggressive in the past, demanding that we kill enemy terrorists instead of using non-lethal weapons. And now she was demanding that I complete the mission. "Terminate the prisoner, Agent. If you are too afraid, you are ordered to return to base, on Manderley's authority. There is a helicopter waiting."

I began to ask more questions from Lebedev, and he started to elaborate on the crimes of the UN anti-terrorist coalition that employed me. Navarre interjected, ordering me to kill him. "Leave us, Agent. Now.... You have disobeyed a direct order." She had made clear her intention to kill Lebedev.

The choice was clear: shoot Lebedev and be rewarded for a job well done, or disobey and let Navarre do it for me. Neither choice was satisfying. The man was a terrorist leader, and his people had been trying to kill me throughout the game, but he was convincing, and it was against our policies to kill him.

I got pissed.

So I shot agent Anna Navarre in the head.

It was brash, an act of anger, murder in the second degree. I had not taken any of the choices that had been offered. It seemed inevitable that Lebedev would die, but he did not, so I fully expected the game to break. Instead...

"I guess Paul must have convinced you."

Lebedev lived and kept talking. I learned the big secret and we both got out. I made up a story for my employers about hearing shooting on the plane in which he had been hiding.

It was an incredible feeling. I truly acted spontaneously and out of frustration. For the first time, a game let me take the "third option". I just hope that more games in the future offer the same.

Commentary: Chris Crawford on Interactive Storytelling

CCoIS, Chris Crawford's latest book, is partially an examination of the problems of storytelling and interactivity and partially a guide through the technical problems of creating a Storytronics-style interactive storytelling engine. The first part is full of insight and useful lessons for game designers and storytellers. The second provides an interesting glimpse into the technology that Crawford has spent years cooking up.

For those that don't know, Storytronics is Crawford's current approach to implementing true interactive storytelling, and it's what he sees as the first major step of the new medium he is founding. And he really does consider it to be a new medium; his book, along with most of his speeches and writings, drips with contempt for games. Video games, he proclaims, are adolescent nonsense, defined by simple and often violent themes. Interactive storytelling, on the other hand, is about people.

I'm not so sure that IS is so far separated from games, and I'm not convinced that games really are so limited, but I'd love to play what Crawford describes. He rejects branching-tree stories or foldback schemes in favor of a language-based playing experience and personality modeling. Storytronics provides an authored storyworld with characters and situations in place, and the player navigates this storyworld, forging a storyline.

Some parts of Crawford's approach still look like they might present problems. Consider, for example, the problem of interactive tragedy. The most dramatically compelling ending for Romeo and Juliet is the double suicide at the end. But the player is trying to win! Crawford's solution is to change the definition of "winning" to the dramatically compelling ending rather than the character's best interest. He thus proposes that the player be rewarded with "applause" upon completing the tragedy. I'm not sure that players will find that truly rewarding, and I worry that it breaks character identification. You're forcing the player to role-play rather than powergame, and that won't work with everyone.

Crawford also concludes that interactive storytelling cannot support a creative "third option" to problems. His reasoning is that, should he offer the choice to the player, it becomes obvious and no longer is a player-created solution. Within the language-driven context he describes, I understand this limitation, but I know that in Deus Ex my most powerful experience was a third option involving shooting a previously-friendly character. The game did not present the choice to me directly, but it existed in the background. I worry that accepting this large a limitation is cutting off a powerful playing experience.

But despite my few misgivings, the majority of the concept of Storytronics sounds solid, and I'm looking forward to playing it.

The primary reason I read the book, however, was for the insight it gave into the more general fields of storytelling and interactivity. Crawford offers plenty of useful lessons. Stories are about people, not things. Interactivity is a multi-agent cycle of listening, thinking, and speaking. Stories are not based in spectacle. The choices that the player is offered should not be black and white, or even black and white and gray, but ALL grays. Every design decision constrains future design decisions, so do the tough stuff first.

I still feel beholden to games, and I'm not willing to totally abandon them in favor of IS, but this book contains enough insight that I'm happy to have read it.

Role-playing vs. Powergaming

I've heard a lot of discussion about role-playing versus powergaming. In the context of a tabletop RPG, I understand the dilemma. Gamers are given an explicit goal. But a lot of the fun of the game comes from role-playing. Sometimes, winning the game and role-playing in an interesting way are contrary goals. For instance, finding a stranger in the woods, killing him, and taking all of his stuff is not interesting, but it helps beat the game.

The problem is that what you (the player of a game with an explicit goal) would do is different from what the character (a fictional person in the game world) would do. Role-players tend to enjoy the story that is created in a game, so they make an effort to inhabit the mind of their character. Powergamers are more interested in the competitive and game-mechanical aspects, so they just do what it takes to win.

A lot of games cater directly to the powergaming crowd by dropping most pretenses of a story and offering great mechanics. Team Fortress 2 doesn't explain, for instance, what these two color-coded teams are doing trying to kill each other. A much smaller subset of games cater to the role-playing crowd. Morrowind, for example, isn't a whole lot of fun unless you get into the lore and try to inhabit the game world.

Most games will offer at least a context for the role-player parts of us to enjoy. Counter-Strike, for instance, has the context of a terrorist vs. counter-terrorist battle. That mindset adds to the fun, as we peek around corners and defuse bombs with seconds to spare. But if a game like Counter-Strike encourages behaviors like bunny-hopping through its mechanics, then the role-players are generally forced out by the powergamers.

The ideal games, then, are the ones that have mechanics that encourage the character's intentions directly. Sports are the best example, since the player and the character are the same. Powergaming vs. role-playing even seems like a stupid way of looking at sports since the only story context is the real world.

Other games do a good job of this as well. Uplink, for example, casts the player as a hacker. Winning the game entails hacking systems, making money, upgrading your own computer, etc. There's no motivation to break the story context; you're not doing anything that your character wouldn't do. The party game Mafia is another good example; the mafioso's character has to lie, scheme, and deceive in the same way that the player must lie, scheme, and deceive in order to win.

I think that the issue might ultimately come down to character identification. The best games make us inhabit the character in every way. Our goals and the characters goals are the same, and the game world operates according to the character's world (i.e. if the character is a counter-terrorist agent, then bunny-hopping is simulated as being ineffective). Catering to both powergamers and role-players is just a matter of simulating the fictional world with enough depth that what we must do to win and what the character would do never diverge. Let nothing in the game be unrealistic from the character's point of view, and you've got a winner.