[Suggestion] Behavioral pattern analysis to identify griefers/trainers in dungeons.
People aren't going to run dungeons if they just die all the time. They will specifically avoid dungeon activities resulting in death when its a solo player training mobs into every group in the dungeon, regardless of motivation. If GMs are not going to actively monitor solo player activity in the dungeons, look at their damage/healing output when they gain aggro from mobs, or look at pathing patterns of said solo player... Maybe look at automating it with alerts.
Conceptually, a system could be coded to collect the necessary data to identify players training in dungeons that result in other players deaths, and alert GMs to this activity. A lot of this info has been pieced together from MMO coding questions with multiple AI platforms to understand some player movement/logging/database activities in MMOs and the resources required to make that happen (I didn't want to post something if it was such a heavy programming lift it made it unrealistic), it kind of morphed into how to automate player tracking for potential griefers/trainers (categories and some of the behavioral definitions are fully credited to ChatGPT query responses)
1. As a group enters the dungeon area, they get assigned some identifier, solo players get their own identifier, associate the following data points with identifiers.
Patterns of player movement in dungeons:
Paths taken by players in real-time, logging player positions and if the player is moving or not.
Patterns of mob movement in dungeon:
Identify mobs that are aggro'd, mob pathing, and who mobs have targeted.
Interaction between players and mobs:
Log the damage/healing/abilities/buffs/debuffs
Proximity between players/groups:
Log the distance between groups in the dungeons. Are the players on the same level/elevation? What is the distance between groups? How often are players reducing that distance to zero?
2. Behavioral Rules (ChatGPT, 2025)
Define behaviors that indicate potential griefing. For example:
Aggro Hopping:
The player consistently pulls multiple mobs and redirects them toward another group without engaging in significant combat.
Target Switching:
Mobs frequently switch from the player to another group in suspicious patterns.
Mob Density:
The number of mobs in a group’s vicinity spikes abnormally, and the pattern repeats.
3. Pattern Recognition Algorithms (ChatGPT, 2025)
Use algorithms to detect griefing patterns:
Path Analysis:
Compare the player's path to the mobs’ aggro paths. If the player repeatedly leads mobs directly toward other groups, flag the behavior.
Heatmaps:
Generate real-time heatmaps of mob density around groups and correlate spikes with nearby solo players or groups.
Repetition Analysis:
If a player engages in this behavior multiple times within a specific time frame (e.g., 3 instances in 10 minutes), increase suspicion level.
4. Behavior Scoring System (ChatGPT, 2025)
Introduce a scoring system to assess the likelihood of griefing:
Weighted Metrics:
Assign weights to actions such as:
Pulling mobs without engaging them (+2 points).
Leading mobs into another group’s proximity (+3 points).
Lack of meaningful combat engagement (e.g., <10% of the mob’s HP dealt in damage, +2 points).
If the score exceeds a threshold (e.g., 10 points), trigger further action.
5. Automated Responses
Implement in-game responses based on detection:
6. Log and Review (ChatGPT, 2025)
Event Logs:
Automatically save logs of flagged behavior, including timestamps, player paths, and mob interactions.
Admin Alerts:
Send notifications to moderators or admins for further review if the behavior persists or is particularly disruptive.
I'm sure lots of people have talked about this stuff, I would hope Steven and his team are discussing this stuff as well, I just figured if not, a solution starts somewhere, and complaining about it isn't going to fix it.
