I am currently throwing ideas around for a comic-style competitive third-person stealth game.

One player will assume the role of the infiltrator whose goal is to steal a certain asset or hack a computer while remaining undetected by enemy guards, defense bots and security cameras. At his disposal is a standard set of equipment including a stun gun and hacking darts. If he sucessfully completes his task(s) and leaves the map without getting caught, he wins.

The second player is in command of said guards and surveillance equipment and has to prevent the first player from completing his task. If he can catch the first player, Player 2 wins. However, while Player 2 has all the manpower and technology at his disposal, he can only see what his minions can see at any time or what the surveillance system messages back (lights out in room x, door opened at point y, camera out in hallway z etc.)

Both players would invest points in and set up additional equipment before the map starts. Player 1 might want to buy a camo suite to make him invisible for a short amount of time to get around guard posts and cameras or disguise himself as a water dispenser or flower pot (fun before realism ), while Player 2 might buy some additional/better guards or place thermal cameras which could even trace the residual heat of an infiltrator sneaking through. In a way, one player is playing Deus Ex while the other is playing Dungeon Keeper.

The problems I am stuck with right now is how to keep the game interesting for Player 2 while keeping it balanced for Player 1 who seems to have all the fun while his opponent is playing the waiting game.

I have already decided that there should be multiple entry and exit areas for Player 1 on each map to prevent Player 2 from just spamming all his guards and cameras around a single point or cram them all into one room to wait for the infiltrator to arrive (though a stun grenade might take care of that). Introducing multiple, randomly chosen objectives might be another idea to counter that playstyle. Or would it make sense to have Player 2 start all his guards' patrol routes on "his" side of the map to allow Player 1 to leave the entry area before everybody rushes at him?

Another idea might be a possible cooldown on guard orders Player 2 can give - he could go check every four sides around a guard by making him turn, but once he orders a guard to look/walk somewhere it will take a certain amount of seconds before he can issue another order. Or maybe guards who stay in one single spot for too long have a reduced perception (boredom!) or doze off completely.

I am thankful for any input you folks can give.

One thing you can give the second player to do, is have him be positioning and re-arranging the guards, rotating cameras, and things like that for a better security layout, while Player One is infiltrating.

The security could already be set up with a basic layout, but Player Two would want to re-arrange everything, and that takes time (for Guard A to reposition himself in Room B), and it takes time for Player Two to manually go through each guard (Player Two "making the rounds" at a high level of abstraction) to make sure they are in ideal positions.

Your idea sounds similar to the multiplayer in The Legend of Zelda: Phantom Hourglass, look it up and you might get some inspiration

I'm curious if you've done any prior research, or why you want to make this.

I think you want to focus on if this game's realtime or not, and how the assaulting force is discovered. Why, if the players know they're both in the game, would the waiting player not just find the infiltrator on his own terms by killing everything?

The idea originated from an online discussion about games people would like to see and the idea of a multiplayer Deus Ex, and that got stuck to the point where I lay awake at night fantasizing about gameplay. If that isn't a good sign, I don't know. It would also be a great opportunity to learn something new, like multiplayer implementation and a bit of AI.

Game would be realtime. The infiltrator could only be discovered when being spotted by guards or security cameras, though the defender would get some additional feedback from the game. If a door opens somewhere without anybody being around, it's a sure sign that the infiltrator must be around that area, though you wouldn't know if he just opened that door to set a false lead for the defender.

Basically killing/catching the infiltrator is the goal for Player 2, but I would limit control over his minions so it doesn't end up as a game of Cannonfodder.

One thing you can give the second player to do, is have him be positioning and re-arranging the guards, rotating cameras, and things like that for a better security layout, while Player One is infiltrating.

The security could already be set up with a basic layout, but Player Two would want to re-arrange everything, and that takes time (for Guard A to reposition himself in Room B), and it takes time for Player Two to manually go through each guard (Player Two "making the rounds" at a high level of abstraction) to make sure they are in ideal positions.

That would certainly give him something to do, yes.

What of making it possible for both to infiltrate each other or you could have missions for player 1 to be completing.

I guess you would have to limit control over your minions to avoid having to micro-manage them and your infiltrator, but that's a pretty good idea indeed. Might require more symmetrical level layouts, but it wouldn't seem out of place with the non-realistic approach and would introduce an element of time pressure to finish faster than your opponent. Maybe two competing companies sharing office space in the same building?

Your idea sounds similar to the multiplayer in The Legend of Zelda: Phantom Hourglass, look it up and you might get some inspiration

Thanks, I will have a look at it.

For some reason I'm thinking that getting immediately notified of a door opening is too much of an advantage, unless it's something that's reported by one of the guards or surveillance when they see it. Player 1 would have to consider if the time it takes to close a door is worth the risk to try decreasing his chances of being noticed.

I think it might be fun as player 2 to set up and changing patrol routes for guards. But knowing specifically what it is that Player 1 is targeting might make it too easy. But if there were say 10 different things player 2 is charged with protecting and player 1 is interested in only 1 of them, then player 2 has a reason to continually divert attention from each location.

What about this:

There is one "defending" player and there can be 3 - 7 infiltrators, and there are five control points on the map, if one infiltrator manages to reach 3 our of 5 control points he becomes the new defending player, the old defending player becomes an infiltrator, and all infiltrators are moved till they are a certain distance away from the nearest control point, all "control point reached"-stats are reset and any alerts are reset.

The defender needs to be capable of changing enough of the defenses to make sure the old defender has no usefull knowledge (at all)left of the defenses by the time he reaches the first control point(sooner would be better), the defender could for example send his guards in other directions to patrol.

A player could either score points through being in control for a long time, or by catching a lot of infiltrators.

It's a bit action-orientated, but it allows for a good flow of the game.
(and realy, nobody wants to be a defending player when the infiltrator goes afk, or decides to study that camera for five full minutes)

It would also need to be connected through a server, but that is pretty inherent on small multiplayer games, the game itself can be hosted on a player's computer
(You can consider letting whoever hosts the game be the defender)