Now the movement system has been analysed, designed and implemented, lets look at the next logical step in stealth gameplay: Detection.
With detection, I mean the way that guards or cameras can see you and be alerted. How close can you get to a guard, how far away can he see and how wide is his viewing angle.
How close can the player get to a guard While a bit of the easier question to answer, it could still proof difficult to not let the player go out of control.
Within the Guard AI analysis document there is a piece on how player will observe and assume a guards movement pattern based on the surroundings and the give-aways of the guard. This document also goes into the importance of predictability. To make sure these assumptions from theory are correct, a paper prototype has been made an executed.
The made prototype The prototype is split into 3 parts, a part where guards have no indication where they are currently going, one where they do and one where there are 3 frames after each other.
Within this document there will be an analysis about how a generic asset pack can be created to satisfy the design challenge set in the concept document.
This is going to be a more technical document due to being about development. Code examples may be given for C# code.
Pre-existing knowledge Due to having made some NuGet packages and trying to split up my own personal workload into different applications and projects, I’ve had some experience working with making code generic.
The following document contains the notes I’ve made while playing Dishonored 2. Keep in mind these notes are just my thoughts while going through the game.
Easy patterns to understand:
Most guards stay stationary until they are alerted. When alerted they walk in easy straight paths and only alert guards around them. The city is in constant chaos so death and sounds are not unknown to the area. This makes it so when you get spotted by a few guards, it doesn’t alert everyone.
A decently high level overview of the stealth AI system to be used.
An simple analysis of guard AI in existing games.