ForumWarz Postmortem: Interviewing the Game's Creators

ForumWarz is my newest obsession, a web-based game like nothing I’ve ever played. In short, it’s a parody of Internet culture in the form of a real-time role-playing game. You play as one of three Internet archetypes — the camwhore, emo kid, or troll — and try to disrupt message boards any way you can, using your sexuality, bad poetry, cross-site scripting attacks, or simply banging your head on the keyboard. In the process, you’ll meet a large cast of strange characters who will send you on missions in a very funny microcosm of the Internet.

Among those parodied: Furries, Google, script kiddies, Boing Boing, Apple Computer, ricers, 4chan, Ron Paul, gamers, Bill O’Reilly, Tubgirl, otaku, and the Church of Scientology. Also, it’s almost certainly the only game to include a text-adventure minigame based on R. Kelly’s “Trapped in the Closet.” This game isn’t for everyone.

Before reading any further, I’d highly recommend trying the first two or three levels. Warning: If you’re easily offended, this game is not for you. And don’t worry about getting stuck with the Jimmy character during the tutorial; you get to choose a username, avatar, and class when you hit level 2.

I interviewed Robin Ward aka “Evil Trout,” ForumWarz’s developer/designer and only full-time employee, to learn more about the history and making of the game.

Andy Baio: Listen, I want to start by telling you that I’m absolutely blown away by ForumWarz. Brilliantly executed, addictive, sheer fun.

Robin Ward: Thanks, that means a lot to me. When I first came up with the idea, Jalapeno and I spent a little time looking at other web RPGs. And I couldn’t believe how obtuse some of them were. There was one, I honestly forget which one, that said “The best way to get started is to read our wiki.” And I thought that was ridiculous, as if someone is going to sit down and read through a wiki before playing.

Is it just you and Mike “Jalapeno Bootyhole” Drach?

We actually incorporated recently, and there are four of us in the corporation. But the majority of the work is done by Drach and I. I work on it full time, and Drach has taken vacation time to work on it. The others are more casual, we have weekly meetings and we bounce ideas off each other.

I read you worked on it for over a year, and then three months in beta? I saw the original announcement on your blog from just over a year ago.

Yup. I first came up with the idea in September 2006. I was working full time at that point, and started learning Ruby on Rails. Drach is a great friend of mine, as is Jason Kogan, and we’d meet weekly to discuss and work on it. And quickly I started to notice that I was having a lot more fun working on it than I was my real job. And I’d lie there at night, thinking “Wow, wouldn’t it be awesome if I could work on Forumwarz full time?”

And then I looked at my financial situation — I’d managed to save up a bunch of money working as a developer, so I crunched some numbers, figured out a way to live really cheaply, and quit my job. So I did that Jan 1, 2007.

I worked on it full time throughout the year, with Mike’s help when he took his vacation in the summer, then released Episode 1 in Beta on Halloween. Then we launched proper a couple of weeks ago, in early February.

You’re the primary developer/designer and Drach writes… What do the others do?

We’re all friends, who I’ve known since high school (and in Jason’s case, grade school). So it was mostly an excuse to get together and throw around ideas. Sometimes it’s nice to just have people to bounce ideas off. Jason was great at giving UI feedback, like why did you put that button there, that makes no sense, etc. And the meta-game, Domination, is basically his baby. The fourth member, who isn’t so active any more since he had a child, is David Kalechstein. When he was laid off, he wanted to learn Rails and wrote a little bit of code for ForumWarz. (Mainly in displaying our leaderboards.)

We used to drink beer and go out for dinner at every meeting, but after a couple of months that got expensive and unproductive

Beer and meetings don’t mix.

Hah, generally not.

Your influences are all over the map. The name’s derived from classic BBS doors like TradeWars and Pimp Wars?

Yup, very few people get that actually. In terms of the turn-based gameplay and visits per day idea, that came from a door called LORD, Legend of the Red Dragon.

We wanted a name that reflected that. It was actually Drach’s idea to get ForumWarz, he’s the one who looked it up.

Were you involved in the BBS scene?

I never ran a BBS, but I was super addicted to them growing up. Of course when the Internet came around I jumped ship, but I’ll always have a soft spot in my heart for them. When I was young, I’d come home from school and just start dialing up.

LORD started in 1989, so I’m going to guess you came into the scene the same time I did around the late ’80s, early ’90s?

Yeah, that’s about right. I think I got my first modem when I was 8 years old, which would be 1987.

A lot of the influences in ForumWarz are from that time period. I was in love with LucasArts games, and we stole a lot from them, I think. Like breaking the fourth wall, the conversation trees. Also, my favorite parody RPG of all time, Superhero League of Hoboken, came out then.

There’s also quite a bit of text adventure influence in ForumWarz. In the writing, but most obviously in the Trapped in the Cupboard minigame.

We knew very early on that we couldn’t afford to make a visually flashy game, so writing was really important to focus on. Mike Drach actually writes for children’s cartoons, and I’ve known him since high school and have always thought he was hilarious but had never had an opportunity to truly apply it.

Trapped in the Cupboard exists more because I thought it would be a fun project to port Z-machine (the virtual machine that runs the old infocom games) to an Ajax interface. I started doing that for a day or two as a break from ForumWarz coding, and quickly realized it was far more complicated than what I needed. So I scrapped it, and came up with my own Ruby pseudo-DSL. I wish I could brag that I pulled that off. 🙂

You’ve also mentioned the obscure Sierra game Jones in the Fast Lane as an influence.

Ah yes, I wouldn’t advocate piracy, but it’s not too hard to get from abandonware sites. It’s basically impossible find in stores, even online ones.

One question I am asked often is why would you make a game about a mundane thing like the Internet? Jones is my biggest influence for that, because it’s the first real life simulator I played.

When I first started trying to describe ForumWarz to friends and on my site, I realized how futile it was. It’s impossible to explain what it is.

Yes, that has always been a problem. I would say 9 out of 10 people I’d explain it to just had this glazed-over look on their faces. I’m quite neurotic, and during those first few months when I was developing it like mad in a black box, it would bother me a lot that I was working on something that people thought was a totally dumb idea.

When we launched in Beta though, I felt a lot better because it resounded so well with people. It’s really better off played than explained, that’s why on the homepage we just want people to start playing right away. Don’t sign up, just click the big button.

Are there any other web-based games that are doing anything similar? I’ve seen people mention Kingdom of Loathing as the closest cousin.

Kingdom of Loathing is the closest thing we have to a competitor, in that it is a web based parody game. The interface and universe are quite different. You know, ours is a little fake Internet, theirs is a fantasy world of stick figures.

I think they did a really good job establishing that there was a market for web-based RPG games. Coming into the genre late, we had the advantage of new technologies like Rails and Ajax to create our interface.

You’ve also spent a lot of time developing strong NPCs to interact with, through the fake email and IM clients.

Yeah, I always loved conversation trees like in the old LucasArts games. One of the first things I started prototyping was the sTalk interface, because I knew it was going to be important. The interface is basically the same as the first version I came up with, but the tools for building the chats went through many iterations.

I actually wrote a more technical article about the back end for stalk a while ago on our forums if you’re interested in peeking into it.

The characters sometimes parody archetypes (obnoxious gamer, steroidal jerk, creepy pedophile) and occasionally real people, like Boing Boing’s Cory Doctorow as the pop culture/privacy-obsessed Doctor O.

Yes, we love parody, and fortunately there’s a basically endless amount of things to parody on the Internet.

My favorite NPC so far is Fr4gGingR1teZ, the parody of Fatal1ty. It was the last conversation Drach created for episode 1 and I think he’d really gotten the hang of it. Not only did he nail the parody, but he gave him some heart. (Especially since he had no idea who Fatal1ty was when I suggested it.)

Clearly, not everyone’s going to get every cultural reference. The game’s still enjoyable without it, but it just adds a deeper level of appreciation when you get the in-joke.

No way, I don’t even get them all. It’s important that the casual player finds it funny, but the huge Internet nerd finds it hilarious.

I mentioned in that Digital Journal article you linked to that I love Futurama, and it’s great for that. They even have a name for it, something like the 1% joke, where only 1% of the audience gets it.

I was disturbed that I knew 20 out of the 21 memes mentioned during the signup process. (Everything but Rabbit-Chan.)

Ah, Rabbit-Chan is quite obscure. I’m kind of obsessed with Internet culture. Things like Rabbit-Chan fascinate me, because it’s the kind of thing that just couldn’t exist anywhere else. She’s a teenage girl who posted some pictures on 4chan, and they became super obsessed with her. She became a “meme,” in particular how people claimed to be in love with her, how every minute of every day they thought about her and had no idea how to cope.

What’s weird though is hundreds of teenage girls have posted their pictures on 4chan, and yet she’s the only one who got that kind of treatment. It seems that popularity can be quite random.

Speaking of the meme list, I was struck by the signup and tutorial process. Very intuitive, and you can play for 15 minutes before even setting a password. How did that come about?

Thanks, we spent a LOT of time on it. It goes hand in hand with what I said earlier about the idea behind ForumWarz being a tough sell.

Joel Spolsky talked about barriers to entry, and it always stuck with me. Like the fewer barriers to entry you have, the more people you’re likely to reach. So I brought that up with the guys, and I was like, screw signing up! Let’s let them play right away. By the time we require them to sign up, they’ll know whether it’s the kind of thing they’ll enjoy.

The other thing was the choice of a quiz. Originally it was going to be used to choose your class, like in other RPGs, but soon after we started writing it we realized the whole tutorial would be a lot simpler if you just had a couple of attacks, hence the Jimmy the Re-Re character was born.

We thought it was hilarious that no matter what you answered in the quiz you always received the same class. As it turns out, that was a bad call. We’ve received many complaints about that, some people have even left the site because they’re like, “Screw this, I wanted to be a troll!”

I noticed many people get fooled by that, including some friends of mine. I thought it was obvious that everybody rolled the same character.

Especially since when you click “No,” it says, “How can I say this nicely — I REALLY think you’re a Re-Re.” We’ve changed that message to say “Don’t freak out, you can choose your class later, buddy.”

Don’t feel bad, everyone who freaked out was a re-re.

[laughs] Well, websites are neat in that we can change them and screw around with them at any time. We keep a lot of statistics on how many people left, and at what stage. So even though it was a mistake, we will come up with a smoother version of the same joke. It just might take a few tries before we get it right.

The game does an amazing job of teasing you further into gameplay, never leaving you confused about what to do next.

Yeah, I wanted to make sure we held their hand all the way into it. A lot of console and PC games are great at that, but few web games bother. Like I said before, “Read the wiki” is a terrible intro. We knew we could do better.

In Portal, they do a great job of training the player, helping them build a mental model of the game universe one piece at a time. As I played it, I was trying to think how that kind of effortless training would work on the web… When I first played ForumWarz, my immediate thought was that you’d figured it out. Every Web 2.0 startup could learn something from the first 10 minutes of ForumWarz.

Thank you, that’s very flattering. I’m a huge fan of Portal, and Valve’s work in general. Immediately after finishing Portal I started it again with the commentary, and I was like, oh, if only we had the resources to develop with constant testing and feedback.

You did three months of playtesting. Did you ever physically sit behind players to see where they got stuck? Or was it all watching the stats and reading the feedback?

I never once sat behind someone’s back, it was all based on emails and forum discussions, which isn’t the ideal way to do it.

How much did ForumWarz change in that beta period?

In terms of the intro, it’s almost exactly the same as what we went into beta with. If anyone tested that extensively, it was us. During the beta, the game improved tremendously. We had a lot of little nuisances that we took out. Some things that come to mind are allowing people to carry forum visits over days they don’t play, only counting a visit upon their first attack.

We did notice that people seemed to love the game, played through episode 1 in a week or two, then dropped off. So instead of going right into episode 2, we made the conscious decision to work on aspects of the game for people who had already finished it, hence ForumBuildr and Domination were born. I mean that’s still our major flaw, in that you can get through all we have to offer fairly quickly, but it’s better than it was before.

ForumBuildr is one of the most innovative elements of ForumWarz, allowing users to write and design new levels and vote the best ones into production. It really playing off the strengths of the web and couldn’t be done in any other medium.

I developed the “madlib” engine fairly early on, so that we could generate random postings on the forums. Then we’d have sessions where a bunch of us would sit around in a room with laptops, generating the posts for a particular forum, and it was quite fun. I think it was Drach who suggested initially that we give players the ability to use the tool.

At first I was against it, but I think it’s because I didn’t really understand how we would give it to them. Then early in beta, someone posted a thread saying it would be cool if they could write their own posts, and Drach replied saying we’d been thinking of a voting system for that, and all sorts of people were like, wow that sounds awesome.

So we pushed it to the front. I spent a week making the tool more user friendly, maybe another week adding the Digg-style voting controls and we threw it to them. I was quite happy with how well it went over, even with our limited user base at the time.

The user-generated levels are playable from the ForumBuildr section, but not part of the story… Are you going to introduce the user-created threads into the next episode?

For Episode 2, we’ve already created some forums and will create more, but our users are now creating material much faster than we are. We have plans to integrate their forums into our storyline, as well as a generic mission system that can come up with random missions every day with a specific reward. Those forums will be very useful for that. In the meantime, anyone can play the community-created forums.

They can even be used for levelling up, although getting to them [through ForumBuildr] is a little odd. Sometime soon I plan to redo the forum list page, to add things like sorting. I’ll be listing them along with the Episode 1 forums… I think it will make the Internet feel less empty. (The fake internet, that is.)

The real ForumWarz forums are just as strange, with people role-playing their characters.

[laughs] Yes, that was something that surprised us. It came out the very first day of beta. In the “Whiny Bitches” forum, people started posting as their characters and enjoying it, and then someone suggested I give them their own forum to roleplay in, and it took off like crazy.

They can be hilarious at times. I think smart people love to pretend that they’re stupid. But occasionally they take it too far and bring it into other forums and stuff, which has to be cleaned up. You hear these stories of people developing software, then being completely surprised when their users use it in a way that wasn’t intended. Our RPG Forums are that for me. We’ll totally enable them to do that, and I’m glad they enjoy it.

I noticed there was some vandalism on the wiki, with people removing core documentation pages. Do you think people are getting too carried away with pretending to be retarded trolls? Or were they actually retarded trolls?

Hah, I’m not sure. We’ve banned a couple, and hopefully we can stay on top of it. I think we might need to add some moderator support soon. I put a note in the edit screen for the wiki that says “If you vandalize this, you will be banned.” Hopefully that will scare some people off.

Good move. So, let’s talk a little about cash money. You’ve decided to go with paid upgrades but with a couple novel twists, allowing players to cheat extensively, buy new characters, or screw around with other players anonymously. How did you decide on what your brownie points can buy?

A lot of it was stolen shamefully from the Something Awful forums, in terms of being able to buy avatars and prank people. New characters just made sense because people often wanted to play the game as another class to see the difference.

How about the cheats? They’re extremely powerful. (Instantly killing a thread, virtually unlimited money, and no forum limits.)

The cheats are a relatively new idea. People often complain about the forum visit limit when they first start playing, because we suck them in for an hour or two of gameplay, then bam, they have to stop and come back tomorrow. So people had been asking if there was any way to buy visits for a while, and we always said no because it screwed with the competitive aspect of the game.

But then i was thinking about it, and realized there’s probably a lot of players who don’t care much about competition, who just want to breeze through the story. So I coded up a way to flag accounts as cheated, and disqualify them from leaderboards and Domination. It’s been really successful. I had no idea how many people just wanted to get on with the game.

And you’re also running ads, which can be disabled if you donate. Anything other revenue models I’m missing?

Nope, that’s it so far. It’s likely that Episode 2 will require some kind of payment to play, but we haven’t fully sorted that out yet. And the idea of selling t-shirts or posters has been thrown around, but both of those concepts are up in the air right now.

How’s it growing since the public launch two weeks ago?

Actually, first off, I should thank you for your link, speaking of growth. You, and the Wired article that followed, sent us a huge amount of users.

You’re welcome! Glad to help.

Growth has been great since we launched. The beta had about 1,200 accounts, but by the end only about 150-200 were active. Right now we’re closing in on 10k accounts, which is pretty good for a few weeks. The days following your link and Wired, we signed up over 1,000 each day.

I have no idea how high we can take that number. I have read that Kingdom of Loathing has 100k members, but they’ve been around since 2003. We typically have about 2500-3000 accounts active in a given day.

I’m very happy with the response, although I don’t have much to compare it to. I know of a local venture that had a team of 10 working on a social networking site for a year, and after 3 months were celebrating their 1000th user.

Now that the framework’s in place, how long do you think it will take to release new episodes?

Because the framework is in place, the other episodes should be easier to develop, but we’re also being ambitious about it. We want to add new ways to play that complement the existing game.

Having said that, we really need to get moving on episode 2. I’d planned to be working at it much harder than I am right now but the sudden business of the site has meant a lot of work staying on top of bugs and stuff. Drach is going to take some vacation time soon to help out again, which helped tremendously last time.

What would it take for the site to be self-sustaining and for Drach to leave his job to work full time on it?

I can’t speak for him, but I’d guess it would require some regular income. The current influx of users and income is great, but who knows how long it will last?

As a developer, I didn’t feel the risk of quitting my day job was a huge one, as the market for developers seems to be pretty good. As a writer, it’s much harder to find gigs. I’d think if we showed steady income over a period of months he’d be willing to do it.

Wrapping up, I want to talk a little about the technology. You’ve built it all on Rails, and the site is snappy as hell. Any concerns about scaling?

I have no doubt that Ruby is slow, but I tried to design the site in the best way I could, and so far I’ve been really happy with the performance. I’ve been doing web development for years, and I spent about 5 years doing J2EE stuff, and let me tell you I could never go back. I just feel so much faster in Rails, and I think that’s worth the performance loss. You know what they say, servers are cheap, programmer time is expensive

I think scalability is a problem you constantly have to face, no matter what technology you use. We’ve had some bumps and I think there will continue to be them in the future, but I feel confident that they can be surpassed.

How’s the Haml templating engine been to work with? It looks elegant, but I’ve never tried it.

Haml is a great product. I should let you know that I know the creator personally, he lives around the corner from me and he’s well known in the local Rails community. It is not perfect for all situations. I have the odd page that uses ERB for formatting, but for 90% of the pages I write it makes me faster, simply because I don’t have to close tags.

I should point out that we’re running everything off one server now (rails, memcached and mysql), so we haven’t even begun buying additional hardware yet. Up until thursday night we were running on 1GB of RAM too, but that wasn’t enough so we added another GB.

I saw on Netcraft you’re using nginx?

Yes, I use nginx as a proxy to a pack of mongrels. There are many different ways to configure a Rails stack, I’m not sure how I decided on that one, really. It just kind of turned out that way. I read a lot of blogs and stuff. I’m no huge IT person, and it seemed within my limits to configure it.

I read a blog entry that benchmarked nginx against mongrel and lighttpd, and nginx/mongrel came out on top.

I think I read a few benchmarks like that, yes. I remember reading that Apache2 and FastCGI was the fastest configuration, but that it ate a lot of RAM. And since our hosting company charges a fair bit for RAM, I went with nginx.

I wanted to ask about The Unintelligencer, the text parser you built that simulates progressively stupider commenters. Are you unintelligencing user-contributed text in ForumBuildr? Or only the stuff you wrote?

We do a little. It can be tuned with various percentages, but since we want people to quickly recognize what they wrote, it’s very low on their submissions. We could probably tune the actual game back some too.

Right now, some of the text is so mangled it feels greeked. (Lorum ipsum d0l0rr.)

Yeah, it was meant to be very occasional, but I think we took it too far.

I think it works as placeholder text. If the fake comments were more coherent, I’d start reading them as conversation expecting to see replies.

Yeah, and unfortunately we can’t do that. Some people have suggested we take it further, create Markov-chain based replies to your attacks and stuff. I liked that suggestion, but now I’m in that sucky period where everything awesome just reeks of TIME, and I have to just pick and choose the ideas.

You’ve built an extremely solid foundation and a brilliant first episode. I think it’ll be a big success. Thank you, Robin!

Thanks again. It really took an enormous amount of effort to get to this point, a lot of it without any feedback at all, so it means a lot to me (and the team) every time someone has something nice to say about it.