Milsim Creator Framework-Core

MILSIM CREATOR FRAMEWORK - STEALTH MODULE

EARLY VERSION - the rules run and have been checked on a server without
players. Nobody has sneaked past a real patrol with it yet.

In vanilla an AI soldier that sees you a little is on you. Here it has to make up
its mind. You give off two numbers, what the eye gets and what the ear gets (0 to
100). Every AI near you keeps its own awareness of you and vanilla's combat only
gets you when it reaches 100.

WHAT YOU GET
- Two bars for the player, worked out on the server twice a second: the eye
  (vanilla's own visibility, which already counts stance and light, times how
  fast you move) and the ear (movement, stance and the ground you walk on).
  Sprinting in a vehicle counts as running.
- Awareness per AI per player. It rises while the AI sees you and when it hears
  you inside a range that depends on your noise; it falls when it loses you. A
  bullet past an AI or a hit is 100. A shot heard raises the whole group to
  "investigate".
- Bands: nothing, "hm?", investigating, alert, then combat. Between "hm?" and
  alert the group is told there is something and goes to look; at 100 it fights.
- It only concerns a player's character. AI against AI stays vanilla.
- Two small bars on the player's screen, an eye and an ear, bottom left. A third
  piece, an eye in the middle of the screen that fills when someone notices you,
  is available and OFF by default (setting noticeEye). Only the band is sent to
  the client, never the number.
- Other modules can change your bars through Core's modifiers (stealth.visual and
  stealth.audio), for example a perk. Stealth does not name any other MCF module.

ZOMBIES AND OTHER AI MODS
Some mods (Bacon Zombies is the one this was written for) run their own behaviour
and never ask vanilla who to shoot. Stealth finds such a man by behaviour, not by
name: one that has not used vanilla's target pick for ten seconds. Nothing is gated
for him; instead his perception factor is set from the nearest player's bars and
distance, and put back when nobody is near. It is coarser than the bands above.
If your zombie mod sets that factor itself, Stealth overwrites it while a player
is in range.

FOR SERVER OWNERS
- One settings file, written with defaults on first start:
  $profile:MCF/stealth/settings.json. Switches: enabled, gate, fallback, hudBars,
  noticeEye. Then reach (radius, tick), the bar values, the awareness rates, the
  fallback numbers and the HUD place. A value out of range is corrected for the
  run and logged.
- Log lines start with [MCF-ST].
- Switch gate off, or enabled off, and vanilla behaviour is back after a restart.

MEASURED, AND NOT
- Checked on a dedicated server with no players: the bar maths, the awareness
  rise and fall, the three places where vanilla's target pick is gated, the
  fallback for an AI with its own brain (with a switched-off vanilla man), and the
  cost of a tick.
- The rate at which an AI gains awareness at 150 m was measured on the probe.
- NOT tested with players: the bars drawn on screen and whether they overlap the
  stance or stamina icons (move them with hudLeft and hudBottom), a walking
  player's movement classes, the surface factor, a real player as a target, a group
  of three or more, a vehicle, 30 players with 150 AI (the cost), and Bacon
  Zombies itself, which was not loaded on the probe.
- The numbers are first guesses. Feel these first: visStill (a man standing still
  shows 55 to the eye), audioRange (40 m at a noise of 100), radius (also the
  longest distance a fight can start from) and decay.

KNOWN LIMITS
- The numbers are not tuned.
- The fallback for zombie mods is coarse: perception is scaled, not gated.
- Needs a connected player to see anything. In a mission with no players it does
  nothing.

DEPENDENCIES
- Milsim Creator Framework - Core
(Works beside MCF Character: a stealth perk can plug in through Core, nothing is
needed.)

MCF - Stealth

AI notices you over time instead of at once: stance, light, movement and noise decide how fast. Early version.

MCF - Stealth

0.0.5

0.0.4

0.0.3

0.0.2

0.0.1