5 ms·

Ask HN: How do I bootstrap an online real-time multi-player game?

11 points by jabo 16y ago

A friend and I have been working on a browser-based real-time multi-player game for over a year now. Right from the start, we considered the real-time multi-player aspect of our game to be its biggest strength, given that it is a browser-based game and it does not use Flash or any browser plugins (yes, so it works on Mobile Safari on the iPad too). Now, the problem we're facing is, boot-strapping the ecosystem within the game. We do not have a single-player mode, because it is a strategy game and we felt that writing a reasonably good AI engine would take a lot of time - we've already been in development for 1.5 years (working part-time on it) and we want to push out what we have ASAP. We've shown it to our friends, everyone says the game is very addictive. BUT, the problem is, to get the full experience of the game, several people (10 at least) need to be playing together in real-time. Achieving that critical mass of people, that initial push, to sustain the multi-player ecosystem is proving to be a challenge. We tried contacting tech and gaming blogs, etc. We were featured once by a tech blog, but what happened was, though we had a lot of traffic, people visited the site sporadically and none of them got the actual experience of the game. So well, no one came back. (It's a real drag to play the game single-handedly). Any tips on how to go about boot-strapping our multi-player game?

Charuru 16y ago
You should try Amazon Mechanical Turk. Having used it for a huge amount of projects, I find it very effective. Cheap labour > expensive automations usually. Edit: having seen your game, I don't think it's ready yet. You should have some tutorials, get social integration, and improve the look and feel before trying to get user traction.
- jabo 16y ago
  Um, I posted a link to the lite version of the game here. At the end of the lite version the user is shown a link to the main game which is at http://pixza.com/league http://pixza.com/league. In the main game, we use Twitter's @Anywhere platform. We have a wiki at http://pixza.com/wiki http://pixza.com/wiki. We have tutorials (even video) in there. Do you think we're missing something besides this? And thanks for the pointer to Amazon Mechanical Turk. Will definitely look into it.
  - Charuru 16y ago
    Sorry I must've been looking at the lite. Ignore that part then please.
Void_ 16y ago
In these games you can usually level up or something by yourself. And than you do fun stuff with friends. Is there any way of making your character better by things you can do alone? By the way I'd like to see that game, can I be a beta tester? :-)
- jabo 16y ago
  Yes, you can earn points, level up, trade points for special powers. We have all those elements in place. But, to earn those points in the first place, you need to play a game, for which we need several people, for it to be fun! And that's our problem. And sure the game is live at http://pixza.com/lite http://pixza.com/lite
  - davewasthere 16y ago
    Is it just filling in a grey B with blue boxes? I don't get it. Maybe you could have smaller levels that two-three people could play which demonstrates the strategy elements. Something that could be completed in a few minutes. Then, as more players are online - let them attempt bigger maps.
    - jabo 16y ago
      The "grey B" is what we call the watermark. With each game, your team is assigned a random watermark. And you need to fill it up with blue pixels. The opponent's watermark is hidden from your team's view and vice versa. The team which fills its respective watermark first wins the game. And then there are super pixels which can possibly change the entire course of the game! It's a strategy game. Here's how an actual game with atleast 4 people per team looks like: http://www.youtube.com/watch?v=kcRKXgy2rg4 http://www.youtube.com/watch?v=kcRKXgy2rg4 And, we actually had a bigger grid earlier! And then we brought it down to the current size. Any more reduction, and the game will become too easy!
jabo 16y ago
One thing that we are very confident of is that if we were only able to let this 'critical mass' of people know about the game, it would certainly be a big hit and grow exponentially. This conclusion is based on actual feedback about the game from people who we've shown the game to. Everyone unanimously agrees though that getting this critical mass is going to be difficult. Help!
znt 16y ago
I just checked the game out, it looks fun. You're saying the user earns points and stuff, but there's so little feedback about the point and xp gains. Maybe you should add little glowy point popups for that. Go check http://www.realmofthemadgod.com/ http://www.realmofthemadgod.com/ . It is a real time Flash MMO and it does a very good job at displaying user progress. Also if you want to go "gray" you can code some AI which mimics basic user actions at low frequency and display a higher online user count.
- grease 16y ago
  What/Where is his/her game?
  - jabo 16y ago
    Hi, the link is http://pixza.com/lite http://pixza.com/lite
- jabo 16y ago
  Thanks for the feedback. The link I posted here is actually the lite version of the game: No login, no history of points earned, etc. Adhoc type, to lower the barrier to entry. Once you finish the lite game, you're shown the link to the main game. The rationale being, if the player finished atleast one lite game, he/she will like the main game. The main game shows points, has badges, levels, more super pixels and lots more. The problem with AI is, this is a real-time strategy game and we being a two member team, it's going to take us a long time to come up with anything that's good enough to come anywhere close to a real person. And if we have dumb bots playing, new players might mis-construe it as a single-player game with terrible AI!
  - znt 16y ago
    You are preventing the users from making emotional investment in your game by witholding progess feedback. I don't think it is a good idea. Just display big flashy numbers when a user places a square on a tile. And if you really really want decent AI then maybe you should hire some people, teenagers, kids to play it for a while.
YuriNiyazov 16y ago
1) You might be able to implement sufficiently smart bots that seem like real people. Only use this technique in the beginning, and as soon as you get some real players, kill the bots 2) You can pay a bunch of people far away to play the game for two weeks, and during those two weeks ramp up your real-time presence (as in, actively tweet and hang out in gaming forums and just dedicate yourself to shameless self-promotion)
- jabo 16y ago
  1. Like I pointed out in another comment, we are a two-member team and getting a bot to play a real-time strategy game, is going to be time-consuming. 2. Good idea! Do you know anyone who might be interested? Or how I could reach them? Sites like freelancer.com don't exactly fit the picture right?
  - YuriNiyazov 16y ago
    1. I don't know anything about the actual gameplay, but I think you are not interpreting this right. I don't mean some highly involved complex AI, I mean just something that picks its moves slightly better than random. These things are perfect for a genetic algorithm, or an ant-colony-like simulation (Look up John Koza's books on how to do this right). Also, if you are a two-member team, and this is your biggest problem (which it sounds like it is - you have a product and no users!) then you should devote time to it. 2. I don't know anyone personally, but sounds like what you need to do it go on Elance and find a "virtual backoffice" team, and pay them to recruit local players.
jasonkester 16y ago
Sounds like you've picked the single hardest thing to bootstrap. Games are the hardest software products to make successful. Mmos are the hardest type of game to make successful. It will be hard. For some reason this George carlin quote springs to mind: "have you ever tried to start a path? It's hard. You have to hold down the grass by yourself for a while."
- jabo 16y ago
  Yeah, it's so hard!
  - jay_kyburz 16y ago
    Thats what she said.
jay_kyburz 16y ago
I should chime in here. I've been bootstrapping a small games business for about a year. My first game Neptune's Pride has seen some success and my second game Blight of the Immortals is due out in January. I'm on the iPad right now, email me and we can chat off line.
- jabo 16y ago
  Just sent you an email.
chipsy 16y ago
I just had an extremely hard time connecting to the game. 20 tries or more, but now it's going through. The game can be fixed to be more playable with few people. Just have random pixels teleport into the watermark, and don't award anything(or as little as possible) for the teleported pixels. Then, if you have only one player, you basically get a cool effect with pixels rapidly flying into place. With two or more players, the game starts to move along, but the pacing of pixel teleports is controlled by the number of people playing. At 10 players you turn off teleporting and it's all player-controlled. No real AI is necessary, the goal is just to make sure things are happening onscreen and the player has to fight to make a difference at all times. The dullness primarily comes from the screen being static and lifeless, and that is easy to fix. Also, add particle effects. They help.
- jabo 16y ago
  20 tries?! That's a lot. Could you let me know what browser/OS combination you were using? Your idea sounds brilliant. But my concern is, don't you think the teleporting will start to look like cheap AI? I mean, teleported pixels will jump walls of pixels and might evade any strategy the user is trying to adopt. It will essentially break the rules of the game and will give a totally different experience (in the negative sense) of the game. Imagine a strategy game like Age of Empires in which all the AI does it to fly across the map and attack your area directly! Or is it just me thinking this way?
ashraful 16y ago
Its an interesting game, but I was wondering if you have any monetization plans. The reason I am asking is because, even though you may be able to make this game very popular, it seems almost impossible to monetize, because there isn't any "virtual goods" to sell. Even advertising doesn't seem like an effective way to monetize this game.
- jabo 16y ago
  The link I posted on this thread is actually the lite version of the game, where the super pixels are limited etc. In the main game, we have levels, more super pixels (virtual goods) which we plan to monetize. But monetization is really not our goal right now. We are looking to first establish our user base and get them to like the game.
staunch 16y ago
AdWords. Use accelerated delivery and scheduling to deliver a bunch of traffic to your site within a one or two hour window. Pick good keywords so the people actually want to play a game. This method won't be super cheap, but it will do what you want. Whether your game will hook them is another question.
- jabo 16y ago
  Thanks for the pointer about accelerated delivery in AdWaords. We were also planning to advertise on Facebook, since we get to choose the demographic of the target audience.
jabo 16y ago
Another idea a friend suggested is to package the game as an iPad app and put it up on the AppStore. That would give us an additional distribution channel and Apple's user review system could benefit a lot. Are my premises and conclusions correct?
jabo 16y ago
Another idea was that we hold competitions within a community eg, a university setting, and give away cool prizes to top teams. This would not only get the word out about the game, but also get us a set of players who will take the extra effort to be good at the game.