Volca on 30/10/2008 at 13:17
That seems fair enough - thanks to mention :thumb:
Beleg Cúthalion on 30/10/2008 at 13:39
On the other hand – if ghosting exists in versions in which people carry crates from one end of the level to another to prevent little spiders from spotting them, I don't think including these kind of realistic features for higher difficulty levels would be completely wrong.
[Carnage] on 12/11/2008 at 19:26
Quote Posted by hopper
I guess the no-lightgem idea would be possible, if awkward, in third person, but in first person there'd have to be a HUD icon in one form or another.
That would be the price of realism. Of course, it'd only be for some kind of expert mode, since I reckon it would make the game significantly harder. Either way, it would certainly be something I would like to try.
Springheel on 13/11/2008 at 00:08
Quote:
I read somewhere that the developers thought about the possibility of making Garrett's silhouette visible to AI if he was standing in darkness, but in front of a bright backdrop. Although it apparently wouldn't have been too difficult technically, they decided against it for gameplay reasons. Their sentiment was that players would feel the game cheated on them.
We had that same discussion while developing TDM. Although there are some technical hurdles as well (how does the engine handle the difference between an actual light and a bright window that's just a texture, for example), the main reason we decided not to do it was because your lightgem would no longer be an accurate indicator of how likely you are to be seen. Your gem could be pitch-dark but if you're in front of a light you'll still be easily spotted.
It might work if you were willing to do away with the lightgem, but I'm not sure how enjoyable it would be. While in real life you can fairly easily keep track of whether you are silhouetted, in the game it's much harder to do simple things like looking behind you, judging how much space your body takes up, etc.
Chade on 13/11/2008 at 00:36
The trouble with the silhoutte idea is that if the player wanted to hide in a spot that was not in a corner, he would need to spend minutes observing guard patrols.
Realistically, the player is not going to do this, and will instead lurk in corners, and use line of sight stealth. Or he will take pot luck and see what happens.
This is going to make the player feel much less like a master thief, and much more like a bumbling novice. Not to mention level design would have to adjust by giving the levels more dark corners to hide in and more obstructions ... again reducing the feeling of being a master thief.
You could actually risk making the game easier, because it would be difficult to give the player reasonable options beyond "lurk in the corner and wait for all the guards in the room to go away". The line between too easy and too hard would be more difficult for level designers to tread, and so to avoid levels that are too hard you'd probably end up with levels that are too easy.
At the end of the day thief currently has a nice feedback system which allows the player to scan a room and immediately plan his next action. A lot of people wonder why thief is so fun when the player sits in shadows and doesn't do much: but the simple shadow system ensures that the player can always be looking ahead and plotting his next action. Having guards notice Garrett's silhoutte would screw around with this.
I think it would be a terrible addition to the gameplay.
Beleg Cúthalion on 13/11/2008 at 09:10
Again, I think it's a matter of getting used to it. I would like to see Thief IV with these kinds of realism as higher difficulty features. More lights, more ways of being visible (like against a window etc.) and that kind of stuff. Not compulsory, but still implemented.
[Carnage] on 13/11/2008 at 14:30
Additional icons or interface features could help with the 'cheating' thing. Say, the system traces a line between you and the AI, then continues that line until it hits a wall behind you and determines its visibility/light level.
If you are visible from the AI's perspective, an icon shows up of you (being a black silhouette) standing in front of a bright background. This indicates that you are still visible to the guard, regardless of your light gem. The more visible you are, the more visible/brighter the icon.
I agree that it makes the game harder, especially for planning, but you will most likely already go to dark corners or places where you know for sure that they can't see you. Additionally, your silhouette doesn't have to give you away immediately. It pretty much depends on how much of your body 'sticks out' compared to the surrounding darkness. Depending on what else is projected on the wall, Garrett's silhouette might seem nothing more than a table, chair or coathanger. Thanks to his cape (at least in t1/t2), his form is not that human anymore and looks much more like an object, drawing a weapon might even help this cause.
Beleg Cúthalion on 13/11/2008 at 15:12
Apparently he almost never had a cloak, just because it's impractical (ever tried that with stairs or between cupboards with things on them which can fall down...?) and concerning camouflage it wouldn't help because it creates a huge mono-colored... blob. It hides outlines, though.
About the icons... well I guess this would surely undermine the purpose of implementing it, wouldn't it?
Springheel on 13/11/2008 at 19:16
Quote:
Say, the system traces a line between you and the AI, then continues that line until it hits a wall behind you and determines its visibility/light level.
Well, that gets into some of the technical hurdles I alluded to before. How many traces do you think it would take to get a reasonable sense of 1) how bright the area behind you is and 2) how much of you is blocking that bright area? Since every trace would have to be sent by every AI and updated at least every second, you're talking about some major performance implications. Another possibility is to take a rendershot from the AI's point of view and compare the amount of light and dark space around the player, but again that's quite expensive to do (we do something similar, once, for the lightgem, and it chops about 5-10fps off the average machine--imagine doing it for every AI who can see the player).
I suspect, if they truly did have a way of making it work, it wasn't true silhouette detection, but kind of a hacked system where lights had an invisible box projecting out from them, and if you're inside that box you are considered to be 'backlit' by the light.
d'Spair on 13/11/2008 at 19:47
Oh God, I've just discovered the original post.
Please tell me this is really a nightmare and not an insider's information, Digital :)
Though... the sound of it perfectly fits Eidos Montreal's approach to classical gameplay, judging by their 'innovations' in DX3.
