Home Consulting Games Writing Personal Links

Why Online Games Suck

(And How to Design Ones That Don't)

by Greg Costikyan

This piece was published two years ago in The Cursor magazine (now defunct); although the analysis of the state of online games is now badly out of date, its discussion of design principles is as pertinent as ever.

I have designed two online games: MadMaze (on Prodigy, JUMP: Madmaze) and Reinventing America.

They suck.

I've been on MPath and TEN and Meridian 59.

They suck.

Every once in a while, I do a web search for online games and see what's out there. Lots of dumb little Shockwave games, many of them ripoffs of Asteroids or Space Invaders. Lots of people promising games that sound wifty but upon exploration prove to be drivel. I've never found anything that doesn't suck. Well, with one exception: I quite like Stellar Crisis . And Conquest (#Conquest on dal.net) is mildly entertaining.

And okay, to be honest, Gemstone doesn't suck. And MUDs are cool. But the suck/cool ratio is dreadfully high.

I've spent a lot of time over the last year talking to people who are in a position to fund online games. And to the degree that they are capable of grokking games at all, what they want is going to suck. And most of the announcements I've seen for forthcoming games are clear dogs.

Why? What makes this stuff suck? And what can we do that won't suck?

It's all pretty obvious, actually. It's just that you need to understand something about games. And most of the people who're funding this stuff are clueless MBAs, technogeeks who think cool technology will sell regardless of whether the product is fun to play, or Hollywood types who think the Internet is just like TV.

Why My Games Suck

My boss would rather I not say what I'm about to say. His point is, why should we advertise to the world that we do games that suck? He has a point, of course. But if I'm going to criticize other games, shouldn't I be coming clean myself? And, of course, these are games people have played and enjoyed, that have been successful on their own terms. If they're flawed, well, there's no such thing as perfection this side of the grave. And more importantly: if we see the flaws in our own work, that says good things about us and our ability to learn.


MadMaze sucks. MadMaze sucks because it's a solo graphic adventure. You wander through mazes, every once in a while encountering a "Place of Power" where you must solve a logic puzzle. To win, you have to get through all the Places of Power and mazes.

Okay so far. But here's the thing; even when it was designed (1989), it sucked, because you can deliver far better graphics and a far better interface in a computer game distributed on disk. So as a graphic adventure, it is inferior to other stuff out there.

And it's solitaire! The whole point of having a network is that it allows multiple players. In MadMaze, everyone's in his own little world, nobody can talk to anybody else, nobody can help or hinder the others.

This sucks.

Indicentally, more than a million people have played it.

Reinventing America

Reinventing America (hereafter ReUS) ran on the Internet twice (ReUS I and ReUS II). It was sponsored by the John R. & Mary Markle Foundation, a public policy non-profit interested in using new technology to help educate the public on policy issues.

Over the course of six months, ReUS examined every major program of the federal government--everything costing more than $1 billion, and many programs costing a lot less. It provided background information and links to articles all over the web from every political viewpoint. And it asked the players to vote. Votes were tabulated once a week by the game engine, and the game results posted on the Web--vote tabulations and "in game" news stories ("Congress Legalizes Heroin").

As an educational tool, ReUS worked. We provided people with a lot of information, and made them think about a lot of issues. But as a game qua game, it sucked.


Although it had thousands of players, it did not do enough to promote communication among them. Yes, we had bulletin boards and chat; but they were ancillary to the game itself, and used by a small minority. The purpose of a net is communication. The purpose of politics is debate and deliberation. ReUS did nothing--well, not enough--to encourage that.

And it was like voting in the real world: your decisions had only a marginal impact on the outcome. That's not much fun.

And if you wanted to make informed votes, you had to spend a fair amount of time reading our background materials and researching the issues. And after all that work, you got to answer a poll. No strategy at all, no real gameplay.

This sucks.

Why Their Games Suck

Enough self-flagellation. Everyone else's games suck, too.

Latency Isn't the Problem

Start with TEN and MPath. Both companies were founded with great hoopla, and an enormous amount of venture capital money, because they claimed to have "solved" the "basic problem" of online gaming.

The basic problem, they claimed, was latency.

What's that? Just this. When you do something on your PC, it responds in microseconds. When you do something on the Internet, the delay can be a second, or two, or ten, before you get a response.

You can't simply take a game like Doom, say, and expect to play it over the Internet. If I'm in Singapore and you're in Sheboygan, the ping delay between us means our computers get out of synch, and show us different things. You can play Doom if you're "close" to each other in Internet terms--or if data goes over a particularly fast line.

That's what MPath and TEN provide: speedy lines.

In other words, they allow you to take existing computer games, and play them over the Internet.

This sucks. Why?

Because all of these games are designed, first and foremost, for solitaire play. Designers and programmers and testers spend a lot of time, energy, and money ensuring that they are fun to play by your lonesome, sitting there at your glowing phosphor screen. As they should: that's the market they're selling into.

Most PC games have a network play version, these days, but it is an afterthought, tacked on at the last minute as a freebie. They aren't designed from the ground up with multiple players in mind.

And what it takes to make an interesting network game is very different from what it takes to make an interesting solo game. To get people to play games on the Net, you have to give them a compelling reason to want to do so. You have to create games that are so great, and so uniquely suited to the medium, that playing them any other way just isn't the same.

Ten and MPath claimed they had solved the technical problem of latency, and online gaming would now blossom.

They had, and it hasn't. Because the problem isn't a technical one.

It's a problem of design.

Download Times Aren't the Problem

Then we have hybrid games. What's that? It's a game delivered on CD-ROM, but played on the Net. And what's the point of that? The CD-ROM can contain all the graphics and audio and animation files you need to play.

Hybrid game producers claim the basic problem with network games is bandwidth. If you've got a modern computer, it can pull megabytes of information off your hard drive in a matter of seconds. But we all know how frustratingly long it takes to receive even a 20K image over the Internet.

And it's true, to provide a high-end, glossy experience of the same quality as a modern stand-alone PC game, you need a lot of data. Graphics take space, audio gulps it, video devours it in heaping gobs. You can't provide the same level of media over the net, not now at any rate.

So, say hybrid producers, you give it to people on disc. Then they play via the net.

But you know what? You've just shot yourself in the head. Everyone on the Internet can play a game that requires only a web browser. Everyone with a game disk can play a game at home. You've just limited yourself not to the union of those two markets, but to the intersection. No one on the Net can play your game until they go to the software store. And no one who walks out of a software store can play your game unless they have net access.

Hybrid producers claim they've solved the technical problem of bandwidth, and online gaming will now blossom.

They have, and it won't. Because the problem isn't a technical one.

It's a problem of design.

This Isn't Hollywood, You Blinkered Turd!

There, I've said it. You don't know how many times I've wanted to say that in a pitch session.

A lot of technically smart companies have decided that the problem of creating compelling entertainment on the Net can only be solved by people who understand entertainment. So they've hired a bunch of people from L.A. Used to be you had to deal with people who understood software and were clueless about games. Now, they're just clueless.

My favorite was the people who wanted a thirteen week pilot season with an option to renew. Straight out of TV-land. We're talking games, right? Think about it: Chess. Thirteen weeks, option to renew. Monopoly. Thirteen weeks, option to renew. Dungeons & Dragons. Thirteen weeks, option....

I mean, people make false analogies to older media all the time, but you'd think anyone who spends five minutes thinking about what a game is will realize that it isn't a sitcom.

These bozos aren't even addressing a technical issue. They're trying to solve the problem of creating compelling Internet entertainment by making it like TV.

But it won't work. Because the fact that the Internet isn't like TV isn't the problem; it's the opportunity, for Chrissakes.

The problem is design.

3D Isn't the Solution

Hey, here's another big cliff for a million lemmings to dive over. And watch 'em go. Not a week goes by without another announcement of how cool 3D is or is going to be.

Okay, so 3D is cool. I'm using it in a game now--a game designed for home play. But realtime 3D on the Internet? Why?

The technical problems on the Net are latency and download times. So -- let's use a technology that chokes on latency and requires even longer download times than flat graphics!

Is that cool, or what?

I've looked at a lot of 3D chat environments. You spend a lot of time wandering around a 3D space looking for people. It's hard to find them. And if you do, it's hard to get more than two people involved in a conversation at once. Because you all have to be able to see each other to tell what they're saying, and your field of vision is limited.

It makes no sense. IRC is better. 2D chat environments like The Palace are better. Just because the technology is cool does not mean people are going to use it, unless it provides a better experience than less cool alternatives.

3D isn't even a solution to a technical problem; it's a technology pushed for the sake of technical coolness. It creates additional technical problems.

The problem isn't technical. The problem is design.

It Ain't the Motion, It's the Meat

Gaming Is Social

Long, long ago, in the depths of prehistory--that is, before Pong--games were a social activity. People got together to play Bridge or Poker or Monopoly or Dungeons & Dragons to chat and socialize and have a fun activity to do together. Even head-to-head games that require concentration and discourage table talk--like Chess or board wargames--were social, because you'd invite a friend to play with you, and the game was an excuse to get together.

Solitaire games existed, but they were the aberration, not the norm, and everyone viewed them as a poor substitute for the real thing, something to kill a little time when bored.

Then came the PC. PC games had to be solitaire; one screen, one keyboard. Sure, there were "hotseat" versions where you'd take a turn and let the next player sit down, but nobody played them much.

And PC games were fun, because you could play them anytime, you didn't have to get friends together. And for other reasons, too, but the point is this:

Because of a technical problem--the inherently single-user nature of the personal computer--the development of computer gaming has been warped, in ways computer game developers only dimly comprehend.

Once upon a time, our image of gamers was a bunch of people sitting around talking and playing cards. Now, it's a solitary adolescent twitching a joystick in front of a flickering screen.

That's going to change. The Internet is going to change it. Because networks allow games to return to their natural form.

On the Net, it ain't the motion on your screen that makes a game better than what you can play off the net; it's the meat on the other end of the wire--the people you compete against and ally with and talk to.

The Internet is Social

People's perception of the Internet has been warped by the nature of the World Wide Web. The Web is a publishing medium; what's up there is up there, and it's static. Yes, CGIs can customize, but HTML itself is non-interactive. Or rather, it permits only the most modest of interaction: click this link, go somewhere else.

But the Web is not the Net. The Net is the Web, plus email, chat, ftp and gopher, push technology, Java, and everything that is to come.

The Net is a medium that allows everyone to communicate with everyone. It is not a publishing medium, although it can be used for publishing. It is not TV, which is inherently one-to-many. It is many-to-many.

The single Net service people use most is not the Web. It's e-mail. And chat, newsgroups and other bulletin board systems come close.

In other words, what people use the Net for most is socialization: communication with friends and strangers, who often become friends.

The Internet, in other words, is ideally suited to gaming--gaming as it was historically and shall be again, gaming as a social activity. Its technical limitations are not problems to be solved, because its very existence solves the gravest technical limitation of computer gaming.

What Wouldn't Suck

Okay. So if this is true--if socialization is key--how do we create online games that don't suck?

By playing to the medium's strength, instead of trying to solve its technical limitations.

By promoting socialization, of course.


You want players to communicate. You have to give them something to communicate about--something more than "hah hah, I got you!", which is about all Doom players say to each other.

Every paper game designer knows how to do this, because they've been dealing with the issue forever.

You allow and encourage diplomacy.

That's simple: you build in ways that players can help as well as hinder each other. And you make the game sufficiently diverse that there are times and reasons for players to do so. Then, they need to negotiate, to make deals, to backstab and betray. Ideally, you construct the game so that permanent alliances are rare (Stellar Crisis is weak on this point) so that players must always continue talking, to keep their allies in line and persuade their opponents to betray one another.

Social Structure

In a game with more than a handful of players, give them a social structure. Give them neighbors or liegelords; create groups to which different players belong; allegiances and hierarchies. In short, make talking to other players vital to success in the game; and build into the game a reason for each player to talk to specific other players.

You can build a whole game about a social structure, if you want. And people have: the Society for Creative Anachronism is a whole meta-game of social climbing among people with faux Medieval titles. It would work on the Net. And will, if I can get the budget.


Build in facilities for communication. That means in-game chat for realtime games. Messaging. Bulletin boards. E-mail exchange. Whatever you can provide to allow and encourage talk.

That means the bulletin board shouldn't be external to the game; bulletin board participation should be part of the game. Chat shouldn't be external; chat should be built into the client or screen.

And it means building the game so that people have the time to chat. You could add a chat line to Doom, for instance, but people would never use it: the game is too fast-paced, too edge-of-the-seat for a player to stop what he's doing and type in a line of text. He'd just get killed.

You need to design for the demands of communication. Not incidentally, that implies slower paced games. Which is good: you solve the latency issue at a single stroke.

Virtual Community

You want people to play the game more than once; you want them to return again and again, to make friends with other players, to discover new nuances to the game, to assist you with further developments.

You need to create a virtual community, a place where people hang out, socialize, and play. That implies continual participation by the game operators. It implies openness to your customers. It implies bulletin boards and scheduled chats and leaderboards and tournaments. It implies a whole slew of things: and enough has been written about virtual communities that I don't need to go into it here.

But it's vital to get out of the "Gold Master to shrink wrap, then sit back and cash the checks" mentality. A successful online game is a game that is continually refined, from a company that's open to and engages its customers, and allows them, nay encourages them, to engage each other.

The Internet is about communication, and not just among players.



The game of Baseball is played by two opposing teams of nine in nine innings. The meta-game of Baseball is the season; what goes on in an individual game is far less important, to the team and to the fans, than what transpires over the course of the year. The goal of the game is to make more runs than the opposing team; the goal of the meta-game is to win the World Series.

The game of Magic: The Gathering is played by two opposing players, each with a deck of cards. The meta-game of Magic is played by an entire group of friends who trade cards and spend hours thinking about their decks and how to improve them. The goal of the game is to reduce your opponent's life points to zero; the goal of the meta-game is to really, really impress your friends with what a clever player you are and what great decks you've constructed.

Meta-games, when they work, are compelling. The draw together an entire community of people, far beyond the participants in a single game. They provide a reason to obsess about the game, to think about strategy, to study it.

That means using more than the tools you use to create a non-game virtual community. It means thinking about creating a structure around the game that allows all players of the game, anywhen and anywhere, to interact with other players.

How do you do that?

You could use a sports-team metaphor. You could use a military metaphor (the game is the mission, the meta-game the war). You could make the social heirarchy the meta-game, and social climbing performed in individual games.

There's a million ways to do it. But you need to think it through from the game's inception: it can't be an afterthought. And no, bulletin boards and leader boards are not enough.

Is That All It Takes?

You'll notice I haven't talked about a lot of issues: realtime versus turn-based, thousands of players versus a handful, server-based or client-to-client, client software or playable in the browser.

All of that has an impact on the kind of game you want to design. But they're technical issues, every one. Choose a model, I don't care; but if you want to make it work on the Net, you need to think about how players communicate. That's what's key, that's what's vital, that's what the medium is all about.

And I haven't talked about business model, either. No one has a model that's proven to work, yet, but I'm not sure it matters.

Create a compelling game, and they will come.

Start from false design premises, and you're doomed from the start.

Back to Articles page.

Home Consulting Games Writing Personal Links

Copyright © 1998-9 by Greg Costikyan.