Skip to content

Adaptive Endless Boss Mode with Engagement & Monetization Potential

vecnussy55 vecnussy55
in General Discussions

Adaptive Endless Boss Mode with Engagement & Monetization Potential


Dear Team at Behaviour Interactive,

I would like to submit a more detailed proposal for an experimental PvE mode in Dead by Daylight: an Adaptive Endless Boss Mode featuring a progressively learning AI Killer.

Core Concept

A round-based Endless PvE mode where four Survivors face an AI-controlled Killer that adapts after every completed trial.

The AI would analyze behavioral data such as:

  • Loop duration and pallet usage
  • Healing frequency and reset behavior
  • Generator efficiency and coordination
  • Perk activation timing (e.g., exhaustion perks)
  • Target prioritization tendencies

After each round, the AI dynamically adjusts:

  • Perk loadout synergy
  • Chase strategies and mindgame frequency
  • Tunnel/pressure decisions
  • Map control prioritization

The goal is not realism, but escalating difficulty and strategic counterplay.

Progression Structure

  • Round-based escalation (difficulty tiers every 3–5 rounds)
  • Increasing AI aggression and predictive behavior
  • Optional “Boss Mutation” phases (e.g., faster vaults, improved tracking logic)
  • Leaderboard integration (highest round reached)
  • Seasonal reset option for replayability

Engagement & Monetization Potential

  • Cosmetic-only rewards for milestone rounds
  • Unique Endless Mode banners, charms, or titles
  • Optional paid “Endless Challenges” track (similar to Rift structure)
  • Streamer-friendly format (high replay value, escalating tension)
  • Limited-time launch event to measure player retention metrics

This mode could increase:

  • Player session length
  • Solo-queue engagement
  • Content creation visibility
  • Skill development without PvP frustration

Technical Feasibility Approach

Rather than full machine learning, the system could initially rely on:

  • Pattern recognition thresholds (e.g., “High pallet drop rate” → Brutal Strength logic)
  • Decision trees with escalating complexity
  • Pre-built adaptive perk pools triggered by statistical triggers
  • Data collection during experimental release for iterative refinement

A limited-time beta release would allow performance testing and balance calibration before permanent implementation.

Strategic Value

An Adaptive Endless Mode would:

  • Introduce the first true PvE competitive experience in the game
  • Strengthen long-term replayability
  • Expand accessibility for players who avoid high-toxicity PvP environments
  • Thematically align with the Entity evolving through suffering and adaptation

Thank you for your time and for continuously evolving the Dead by Daylight experience. I genuinely believe this concept could add a unique, high-skill PvE dimension to the game.

Kind regards,vecnussy55

Tagged: KillersPerksSurvivors

Comments

  • GentlemanFridge GentlemanFridge

    I'm having trouble envisioning this. And I suspect this was written by an LLM.

    Would you have an example of how this is pulled off in another game?

  • vecnussy55 vecnussy55

    There are already good examples of adaptive AI in other games: Left 4 Dead 2’s Director adjusts spawns based on player performance, Alien: Isolation’s Xenomorph reacts to your patterns, and Resident Evil 3’s Nemesis tracks player behavior to adjust pursuit. The DBD AI Killer could work similarly—observing survivor actions and adapting strategy/perks—without needing full machine learning.

  • GentlemanFridge GentlemanFridge

    @vecnussy55 said:

    There are already good examples of adaptive AI in other games: Left 4 Dead 2’s Director adjusts spawns based on player performance, Alien: Isolation’s Xenomorph reacts to your patterns, and Resident Evil 3’s Nemesis tracks player behavior to adjust pursuit. The DBD AI Killer could work similarly—observing survivor actions and adapting strategy/perks—without needing full machine learning.

    L4D is not what you're looking for here. How it works isn't particularly applicable to DbD, since the latter relies on a map that once generated, doesn't change. L4D can manage enemy spawns, item drops, even break open the map to create shortcuts, on the fly. None of that works in DbD. There's only 1 enemy (maybe 2), items drop only from chests, and dynamic shortcuts make no sense in DbD. There'd need to be far more variables in the killer's abilities for a L4D-style director to work.

    With RE3 I have to assume you mean the original, 'cause in the remake all Nemesis appearances were scripted. And in the original, as far as I can tell, it was just random when he could show up. But I didn't play it so I wouldnt know for sure.

    Alien Isolation is probs the best example, but probably rather difficult to implement in a setting like DbD. Alien Isolation is a much, much slower game than DbD. And if you're found, you're immediately dead. If you want to emulate that game's systems, you need to overhaul DbD's main gameplay loop quite significantly.

    Also, why does it have to be an AI to begin with? Why can't it be a player?

Categories