next »

[diegzumillo]

Hi all

One of my oldest and dearest projects is a first person stealth game. Among one of its features is a visual detection system that can actually spot the player from its silhouette, instead of just shadow/light. When I came up with this and the prototype was actually easier to make than I thought, I got very excited! However, the more I think about the more it seems like a bad idea. Thief and similar games have a less realistic detection system but the simplicity works in its favor. The player only has to look for shadows on the ground and walls to know whether he'll be well hidden or not there. In my system the player has to look for the line of sight of every AI around relative to where he wants to be. It's a lot more variables.

Sigh. I guess what I'm trying to do is to seek validation from someone else because I grew attached to an idea that deep down I know has to be discarded. It's not impossible to make it work, but my best workarounds turn the whole feature into a pointless gimmick.

[Tobers]

Shadow/light works because it's clear, intuitive and been done before. If you could wrap a clear and intuitive concept around that idea/code it should still work!

Maybe like, the player can create holograms of themselves to see if an enemy will spot them before moving.


Anyway, I think you can make stuff like this work, it will be much harder than copy/pasting existing mechanics of course.

[diegzumillo]

That was the plan. Give the player all these tools to 'simulate' what would happen if he were there. Could work, but I'm very skeptical about it. Seems suitable to a simulation of some kind, but not exactly fun.

[Tobers]

Yea I know what you mean. Could take a TON of time/resources to figure out how to make it work well.

[gimymblert]

Have better feedback? the gem in thief tells him when he is in the dark or not, but if you have a radial indicator that explain in which direction you are unseen?

[diegzumillo]

This reminds me of the first 3d platformers. The additional dimension creates more complications than payoffs.

[Mittens]

I've also had this idea in the past, the idea of a stealth game which is not based on rigid game rules/logic but is instead based on how well you can actually hide yourself visually.

I think you should go ahead with it because even if it's not fun, at least it's an attempt to innovate and therefore is marketable by it's novelty.


I personally hate most stealth games because they are not about being sneaky or clever intuitively. Instead they only dress themselves in this guise. Really stealth games are 2 dimensional puzzles. First you learn the routes of the moving pieces, you learn the timings, then you solve the puzzle over time.
I hate this, it's so tedious.

However I think the concept of stealth'ing around is really awesome, and I love games that execute it a little bit more intuitively then most games, but you can always still solve the logic behind it. The logic is always simple enough that it can be figured out and worked around, this is how the developer intends you to win!

Long story short, I say go for it because I would love a stealth game where the enemies follow the well known real-life rules that surround whether or not we can be seen by another person (instead of imposing tedious puzzle mechanics into this scenario)

[diegzumillo]

Really stealth games are 2 dimensional puzzles.

Yeap. That is true! I came to that conclusion as well but I love stealth games. But I disagree it's a guise, it's just what it is. Stripped out of its context most games will get very simple like matching threes, pointing and clicking at things, timing, puzzle etc. Most stealth games under the hood are timing and puzzle; you walk over the allowed regions, the regions themselves follow patterns which you have to identify. You seem to want to play a simulator of some kind. Simulators are fun, but that's not what I want to do.

[gimymblert]

Also you can hide this complication with level design anyway, it's like making an ai so intelligent it can flank you ... then make all level linear with no side alley from which they can flank ... You can contain any complication of game design with level design, be careful to where you place spot with less covering, be sure to teach the player those place, and there is no silhouette anyway in corner, so you can have safe spot in corner and silhouetting as a push your risk mechanics.

For example you could have a configuration where the player is at a T junction between corridor, the entering corridor have a guard that rotate his FOV, when looking straight at the T junction you have 2 big statue that serve as a corner in shadow, if you hide between them there is no silhouette from the traversing but as you move toward th entering corridor your silouhette is visible but still not to the guard looking down the entering corridor.

[diegzumillo]

That makes a lot of sense. But if you take this too far the whole mechanics becomes the definition of a gimmick. In your first example, saying the AI can flank you but render this functionality pointless by level design turns it into a marketing gimmick.

There's one more thing I didn't mention about my idea. The whole point of using this method is that the player can then change color to become invisible even in the light, if the guard sees him against a background of the same color. So how about a mixture of both worlds: completely hidden in shadows and silhouette mechanics in light.

This also feels like a gimmick though. I can't think of many situations where this is an interesting mechanic. A interesting case is a room with two guards, half the room is red the other half is blue and you are in the middle, each guard looks at you at different times so you have to keep changing color to avoid being spotted. But in most situations you just change your color based on the environment, which feels like something that could easily be automated. I don't know. I played a game recently that did that and it worked nicely, but it was a top down game with colors being very clear and well defined. It's a different scenario.

[gimymblert]

I was using exaggerated example to make it clear, it's obvious that the AI example is blown out if it's the whole game, however, say it's the first level it's a correct way to teach the player by decreasing the enemy agency so he isn't too overwhelm.

Don't feel bad about gimmick, mario game is all about the gimmick, as long as you expand the simple idea instead of just exploiting it. And you only do it by organically making unit case and their combination and rating them according to experience metrics (spectacular, difficulty, complexity, etc ...).

Else making a game is still better than not making a game, if you can't make it good ... also keep it short if you don't know how to extend, but generally playstesting will give you hint, also if you build a community interested in the game idea will flow through suggestion.