Mobile Game Postmortem: Ngame's Chop Suey Kung Fu

When Ngame moved into the WAP market in early 2000, we initially concentrated on three big, innovative multi-player titles: Alien Fish Exchange, Merchant Princes, and DataClash. However, it was immediately apparent that, successful as these titles were, they were almost too big for a mobile phone. After talking with carriers about what they wanted, we realised that we were going to need a suite of smaller titles to bulk up our portfolio. The design challenge was to think of something that could be written by a single programmer in a couple of weeks.

The entire spec for Chop Suey Kung Fu was written in half an hour, on less than one side of letter paper. It is based on a simple traditional dice game, and here we really must acknowledge Reiner Knizia’s excellent book Dice Games Properly Explained. There were only three things that mattered: First, it was multi-player. Second, it was a popular genre. And lastly, it was really, really easy for the player to understand. This punched every carrier’s buttons. The marketing guys loved it.

Production started in early August, and we spent weeks trying to find a good artist who could work in WAP – using just 96x44 pixels, and monochrome, we needed strong characters, exciting martial arts moves, and a distinctive style. When the samples arrived from Martin Wheeler, we knew we had a winner. We just had to build the game. We gave the job to our rawest recruit, Alice, and we had a prototype inside a week. Just over a month later, we were running live, and notched up over 100,000 users.

Halfway through development, we needed to rethink our digital TV strategy. The ambitious plans we had for really exciting content looked as though they were going to be scuppered by the crippling limitations of the set-top boxes. We needed to develop something that would run on a bottom-end set-top box, using very limited HTML. However, we were determined to create multi-player games for TV. Chop Suey came to the rescue. We realised that with very little effort, we could convert the game from outputting WML to outputting HTML: all we needed was to change the graphics, and add some animation. This would not only be a TV title, but a fully convergent game, which could be played on both TV and WAP simultaneously. The TV version was more or less completed in December, and went live in April 2001 on Ntl in the United Kingdom.

This was a baby project, even by Ngame’s standards, but was very important in showing what can be done with very little. Since then, Chop Suey Kung Fu has become one of Ngame’s biggest successes, frequently outperforming the much bigger games, and is now one of the world’s most popular wireless games.

What went right

1. Small game, small team. Building Chop Suey Kung Fu for WAP was just like back bedroom coding, a return to 1980’s values. The full-time team was one programmer. There was almost no design or production work required, and between us, we fitted everything else in between our other jobs. No heavy management, no major milestones, no publisher breathing down our necks, no endless committees all trying to add features, and no arguments between teams. Communication problems weren’t an issue – who’s to communicate with? Alice just built it.

Even when we moved to TV, we only added in a couple of artists and an HTML programmer for a few weeks. This doesn’t exactly bring you into the “big team” stakes. We briefed the artists for a couple of hours, then left them to get on with it.

Best of all, with such short development times, nobody had time to get bored. There wasn’t a death march, or even all-night bug-fixing sessions. It was frequently frenetic, and there were plenty of mad panics, and at times we started to wonder who was in control, but overall everything went pretty much to plan, and everybody in the company got involved. The nearest we ever got to creative differences was the recurrent argument about whether you should be able to see Kim Chi’s underwear when she did a Chop Suey Special, whether this was sexist, and if it was, what effect it would have on player numbers. By keeping it fast and light, and doing something well within our ability, it was immensely satisfying to build. The effect on staff morale when you build something quick and fun cannot be underestimated. It brings back a lot of the fun to the job – and isn’t that why we all joined the games industry in the first place?

2. Purpose-built tools. Most of the credit for the ease of development has to go to our in-house toolkit, Sin. It was designed to make coding client-server games as easy as possible, and it really proved its worth. It allowed us to create convergent content that could simultaneously handle clients on several different flavours of WAP, HDML, and HTML, giving us access to pretty much the complete range of wireless, Web and digital TV devices. It was easy to localise the content for different carriers and different languages. Our development costs were tiny as a result of using Sin.

3. Building an online property. Online-only properties are very nice indeed to work with after the constraints of CDs. For a start, you can concentrate on distributing the software, and not worry about box art, printed manuals, or any of that rubbish. Just the game. The time between finishing the game and getting it out live is also minimised, which gives your creative team instant feedback, not a long wait between gold master and publication.

It also means that you can easily patch or upgrade games. Your users don’t need to download anything, you haven’t got buggy software in the shops, and you don’t have version conflict problems. It’s all very clean. About a month after launch, we hit a completely unexpected snag. One carrier in the US found the name Chop Suey Kung Fu to be offensive to people of Chinese origin. We changed the name of the game, on that carrier only, to Extreme Kung Fu. One altered .wbmp file, one text string changed, new version uploaded, and job done in a couple of hours. Imagine recalling a CD-ROM from the shelves to change the name of the game – ouch! While being able to patch running games isn’t an excuse for releasing sloppy code, it does mean that you can worry a little less – any problems that do slip through the QA net aren’t going to sink you completely.

4. Great artwork. A year ago, generating WAP graphics was a black art. Now, people are beginning to relearn some of those skills that most of us lost 15 years ago – if we can remember that far back. Every pixel counts, and eye candy is just as important on a tiny screen as on a PlayStation 2. The graphics can make or break a WAP game, just like any other video game. Martin instinctively knew what we wanted, and getting the graphics done was unbelievably painless. In addition, working with files of under a kilobyte makes for very easy working methods. Everything went back and forth by email, and the process was marvellously streamlined. A producer’s joy!

The TV and Web artwork was all done in-house. The animation was very simple, with just a few frames for each move, and working at TV resolutions meant that everything could be done quickly and easily. It was real comic book style art, and this in itself helped make everybody enthusiastic. The short timescale between concept and finished artwork – typically a day or so – meant that the artists could really let rip, have some fun, and experiment. This energy and excitement came through to the finished game, and gives it a freshness we were trying hard to capture.

What went wrong

1. Under-ambitious design. Chop Suey Kung Fu really is a simple game, and in some respects, it’s too simple. There’s no character progression, no levels, and no real incentive for the player to keep coming back. Very shortly after launch, we started getting requests for extra features. Maybe characters could earn different belts as they got better. Maybe there could be killer moves that players could work up to. Maybe characters should get extra hit points. There is much, much more that we could have added in, but didn’t. As a result, players moaned that there wasn’t enough to do, and would stop playing if we didn’t give them something extra.

We simply didn’t expect Chop Suey Kung Fu to be as successful as it was. We thought people might play a few dozen times, and get bored. We didn’t anticipate people playing for hours at a time, and fighting literally thousands of bouts. Still, it’s better to have been too successful and made a low-cost game that people enjoy than to have succumbed to feature creep! Sequels and patches are in the works, but why fix something that isn’t actually broken? It’s still getting the hits…

2. Didn’t cover enough platforms. Chop Suey Kung Fu was originally designed for WAP. At the time, WAP was all the rage. It hadn’t been released yet, and we really were working in the dark. However, armed with a copy of the WML spec and a Nokia WAP emulator, we designed something that could easily be done with our toolkit on the WAP platform. Unfortunately, WAP failed to take off, particularly in the US. Instead, most wireless users are still on HDML, which is basically the same, but without graphics. HDML also uses a different browser than Nokia. As a result, our first encounter with US carriers was a bit of a shock – all our graphics were gone, and we had to reformat for the HDML browser. In retrospect, we should have considered HDML much more seriously, and designed the game to have ASCII graphics.

The move to Web and TV was also less well organised than it might have been. We’d talked about doing versions for these platforms, and we knew that the underlying engine would cope with multiple platforms, but we didn’t develop them in parallel. Instead, we did the wireless versions, and then started on the rest. Our first error was immediately obvious: we’d commissioned WAP graphics, but we hadn’t commissioned colour graphics at the same time. We had to track down the freelancer who’d done the first set, and persuade him to produce more. It’s a beginner’s mistake, and one we’re not likely to repeat!

3. Localization. One thing we didn’t anticipate was the degree of localization that would be required. When Chop Suey was written, WAP was in its infancy (and arguably still is). We were planning on a cheap throw-away game that would get us in with American and British phone carriers. This year, Chop Suey is being translated into every major European language, and we’re having to look further afield as well. In retrospect, we should have coded to allow for multiple languages, but somehow, it just didn’t seem worth it at the time.

We also found ourselves faced with a mountain of customization. Every carrier has their own style, and wants the content heavily modified. Again, if we’d known this at the time, we’d have written it differently. Instead, we’re stuck with more of a maintenance and configuration problem than we’d like, but hardly an insurmountable one.

4. Collision with commercial reality. Chop Suey Kung Fu has been a great hit with the players. Despite all its limitations, they love it. But it doesn’t yet tell the commercial story we’d hoped for. It still has to see a truly convergent implementation. Telcos each want their own server, as do the TV companies. Instead of the global combat we were after, with any player on any device being able to play anyone else, it has fragmented into dozens of small games, where AT&T customers play AT&T customers, Sprint customers play Sprint customers, and so on. Convergence goes down well at shows and exhibitions, and everyone is keen to buy a convergent game, but nobody’s putting them out in the marketplace.