Game type

Game Play & Game Design
#1

I don’t know if this is just me, but I’ve found I’m slowly moving toward writing games with more AI because it ends up being the most interesting thing to program. Is that happening to anyone else? At first it was all about story and I wrote adventure games. Then the puzzles got more complex and had mini-games, and then I got sick of adventure games. Now I’m on this kind of rpg thing that I shelved a while back and decided to pick up again. It’s not really that I like rpg’s that much, it’s just that there’s more AI behavior and I know I’m just scratching the surface.

I think something like this article would be fascinating:

#2

I’m stuck to FPS because I like blood, action, killing cops and burning banks ;D

#3

I think it really depends on the person. I also find AI very interesting, but I’ve found that in 99% of cases the really interesting AI techniques are unusable. For example, I tried to use a Q-Learning algorithm for the guards in a platformer I made, and they turned out to be so good that the game sucked. I changed it to predictable turn at edges and shoot if they see you type behavior and it became much more fun. Adaptive learning guards would do things like jump off edges at the exact right height to avoid dying and then plug you full of lead before you even walked anywhere. They were more “seek and destroy” than “serve and protect.” Obviously I could tweak parameters so that they were rewarded for waiting until they had seen the player or something like that, but the state space eventually became so complicated and I spent so much time tweaking that it just didn’t make sense in the end.

In a multi-player game like an RTS, though, I could see an adaptive or really good AI being worth having. That way you can have a really really difficult human-like AI to play against if you want.

Anyway, I also really have a lot of fun just exploring tiny gameplay changes to squeeze out the “fun.” Lately all my games are gravitating towards how I can make the most group fun. So, that means multiplayer.

#4

[quote]Lately all my games are gravitating towards how I can make the most group fun. So, that means multiplayer.
[/quote]
I think multi-player would necessitate ai because someone could easily quit in the middle of a game and the ai would need to take over. Maybe not, I don’t play many multi-player games. Your ai is light years ahead of anything I’ve done which so far is very simple fsm. I think it could go too far. In the end, it’s still the game that matters and how fun it is. Predictable behavior can increase strategy, like in chess.

[quote]I’m stuck to FPS because I like blood, action, killing cops and burning banks
[/quote]
Yes, but ai is very important there also. If they are just dumb cops that always do the same thing, the game will become boring.

#5

IMO, multiplayer is always better than any AI. You can really out think your opponent, and that just isn’t possible with AI. Eg, in Street Fighter, I knock you down, then I walk up to you as you are getting up. I could do a wide variety of moves (or no move!), but instead I throw you. As you are getting up again, I walk up, and I throw you again. As you are getting up again, I walk up, and at this point you are totally ready for the throw, I expect this, and so I uppercut you, which beats your throw tech. Point is, I can really get inside your head, and that is awesome.

#6

[quote]IMO, multiplayer is always better than any AI. You can really out think your opponent, and that just isn’t possible with AI.
[/quote]
I don’t play much multiplayer, so I guess you have me there. The games I have played, I ended up getting my butt kicked in about 10 seconds. I was talking more about the programming aspect, though. Programming becomes much more complex and interesting when these bots have to be good enough to provide some sort of game play. I started programming puzzles, and it’s completely different and more rewarding, especially because you can actually play your own game. I am just scratching the surface, though, so who knows, maybe it will get old also. I’m a laid back gamer and game writer, so I’ll probably not do much beyond turn based gaming anyway. I guess it’s just something I’ve found to be more interesting in programming, personally. I know multi-player can be very challenging, but also expensive to release a game that way except maybe p2p.

#7

I find I end up using pretty much the same AI techniques in every game now. Which isn’t necessarily a bad thing - I still find AI really fun. But, for me, multiplayer is much more complicated to program.

#8

I think of multi-player as just passing packets of information back and forth. It doesn’t seem like it would be very interesting. It doesn’t even really seem like a game design problem, more like something underneath.

#9

You should try it, it’s quite challenging to get (close to) unhackable and integrated into the game in a good way.

#10

It’s much more complicated than that. You’ve got to figure out how to deal with lag, how much trust to give to the client, how to make it all seamless, who stores what information (between client and server) and more. It’s a different sort of problem from AI, but definitely not simple.

#11

I’m working on a multiplayer game and I have a server (VPS) from http://www.linode.com/ for ~$17/month. I have yet to see how many players I’ll be able to handle with one server, but I hope the server load will be pretty low. They have west and east coast datacenters in the US, and a datacenter in London, so for ~$54/month you could cover a pretty good portion of the world. In addition, you could provide the server software for users to host (ala Minecraft, Urban Terror, FPSes, etc).

The big difference between losing to real people and losing to AI is that the former is fair. They beat you from experience, skill, practice, etc. IMO, it doesn’t mean much if an AI beats me. However, I do think AI is suitable for some games, and is good to introduce beginners to a multiplayer game before they mauled online.

There are lots of moving parts for a networked game, and that can be interesting and challenging in itself. As you said, that isn’t game design, but it does enable new kinds of game design. It can give more meaning to hard to acquire items. It allows for intra-game social aspects like an economy, trading, marketplace, etc as well as extra-game social aspects like Facebook, GPS, etc. Multiplayer also means you can leverage teamwork, which can be really amazing as it allows you to play with your friends and to learn and grow together. Possibly the biggest reason to go multiplayer is that it can be a HUGE multiplier on replay value. I’ve played Street Fighter IV more than 1000 hours because the online play is fantastic. I would have never played so much against the AI.

#12

[quote]IMO, multiplayer is always better than any AI
[/quote]
+1.94

multiplayer immediatly make “life expectancy” of a game infinite

even nowadays AI are always stupid, even in AAA games, and when it is not, they lack “human” reactions, AI are too hard or too easy but… never surprising