Jump to section
Problem DiscoveryBackground ResearchProposed SolutionAI ModelKPI MetricsUX and Game Design GuidelinesConclusion
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Griefing in Online Gaming
Proposing a robust solution to mitigate anti-social behavior using KPIs and data analytics

Team

Kalyan Chandana
Amit Gaikwad

Role

Researcher
 Product Designer

Timeline

Dec 2023

Problem Discovery

Have you ever wondered why the virtual worlds in online games, places supposedly built for escapism and joy, can sometimes turn into arenas of frustration and conflict? Our study of the online gaming behavior known as "griefing" is centered around this paradox.
‍
‍In this report, we're attempting to understand the human behavior behind griefing and how gaming communities and developers have responded to it in order to propose a threefold solution consisting of:

  • An AI-driven model for moderating grief online.
  • Relevant KPIs that can be used to assess trends in griefing behavior.
  • UX and game design guidelines that can be employed to help mitigate griefing incidents.
Pinterest: Save Feature
Interaction Design for Deeper Emotional Impact

Background Research

History

LambdaMOO, a pioneering text-based virtual community, epitomized boundless creativity and collaboration within its language-driven universe. However, it faced a pivotal moment dubbed "A Rape in Cyberspace," where a user exploited the platform's tools to perpetrate non-consensual acts, sparking profound discussions on governance in online spaces. This incident underscored the complexities of regulating virtual societies and initiated a broader dialogue on ethics and accountability in digital interactions.

Meanwhile, in the visually immersive world of "Tom Clancy's The Division," players discovered and exploited flaws in game design, exemplified by Spawn Room Blocking, where safe spaces became arenas for dominance and control. Ubisoft's swift response highlighted the ongoing dynamic between developer intent and player behavior, prompting reflections on the balance between freedom and regulation in digital communities.
across time: lambdamoo and The Division (Source: Wikipedia)
The juxtaposition of the LambdaMOO incident and Spawn Room Blocking in "The Division" illuminates enduring challenges in online communities, transcending technological eras. Both instances underscore humanity's propensity to exploit virtual systems for personal gain or amusement, emphasizing the crucial need for collective norms and respectful engagement in shared digital environments.

Broader Implications

The incidents at LambdaMOO and "The Division" are not isolated anomalies but indicative of a broader phenomenon in the digital age. They serve as microcosms for understanding the dynamics of online behavior and the complexities of virtual community management. As we delve deeper, we uncover layers of implications that extend beyond the realm of gaming.
Anonymity
The veil of anonymity in online spaces often leads to a disinhibition effect, where individuals feel emboldened to act in ways they might not in face-to-face interactions. This phenomenon raises questions about identity, responsibility, and accountability in the digital world.
Ethical Design
As technology advances, so will kinds of griefing. There is an increasing need for ethical considerations in game design. Developers must balance creating engaging, immersive experiences with safeguarding the well-being of their players.
Empowering Players
The cases highlight the importance of community-led initiatives in establishing and maintaining behavioral norms. Encouraging players to take an active role in community governance helps in managing behavior and fosters a sense of investment in the game.

Proposed Solution

The Approach

Our analysis of the current solutions in place to combat griefing led us to explore topics like AI-based auto-moderation, community-driven moderation, enforcement and penalty systems, and game design guidelines that skew towards positive behaviors. 

From our extensive research so far, we believe introducing a more proactive data analytics pipeline between the identification of griefing behavior and the eventual modifications of the game systems to balance these issues based on insightful metrics may be the best way forward.

AI-Based Model for Anomaly Detection

We think AI-based anomaly detection in multiplayer games can play a crucial role in identifying and mitigating grief by leveraging machine learning models to analyze player actions and interactions.

The basic outline for implementing an AI-based model for anomaly detection and eventually deriving KPI metrics may include:
#1
Define Griefing Behavior
‍‍
Work with game designers and community managers to clearly define what constitutes griefing behavior in the context of your multiplayer game. Identify specific actions or patterns that are considered disruptive, such as intentional team-killing, spawn camping, or other unsportsmanlike conduct.
‍
#2
Data Collection
‍

Gather telemetry data on player actions, interactions, and in-game events. This may include player movements, kills, deaths, chat logs, and other relevant gameplay metrics.

‍
#3
Feature Engineering
‍‍

Extract relevant features from the collected data that may indicate griefing behavior that you have identified for your game. For example, unusual patterns of player movement, excessive team-killing, or abnormal chat interactions.
‍
‍
#4
AI-Model Training
‍
Train machine learning models, such as anomaly detection algorithms, using labeled data. Labeled data should include instances of both normal and griefing behavior. Consider using techniques like clustering, classification, or deep learning, depending on the complexity of the patterns you want to detect.
‍
#5
Testing Anomaly Detection
‍‍

Deploy the trained AI model to analyze real-time or historical gameplay data. Identify anomalies or deviations from normal behavior that may indicate griefing.
‍
#6
Adjustment and Optimization
‍
Regularly update and retrain the AI model to adapt to evolving player behaviors and new forms of griefing. Seek feedback from players and game moderators to improve the accuracy of the detection system.
‍
#7
Identify Relevant KPI Metrics for Defined Grieving Behavior
‍‍
This would include identifying all relevant KPI metrics that can be reliably captured based on the AI models as well the adjoining game telemetry data.
‍

KPI Metrics

The key performance indicators (KPIs) derived from AI-based anomaly detection must be relevant to your particula game and the defined griefing behavior. However, here are some broad examples of KPIs that can be established:

KPI #1
‍Griefing Rate

The percentage of games or player interactions exhibiting griefing behavior. Provides an overall measure of the prevalence of griefing in the game environment.

(Number of games or interactions with griefing / Total number of games or interactions) * 100.

KPI #2
‍Incident Severity

Measures the severity levels of identified griefing incidents. Helps prioritize responses and interventions based on the seriousness of griefing incidents.

Assign a severity score to each incident based on its impact, and calculate the average severity.

KPI #3
‍Player Reports vs. AI Detection

Compares instances reported by players with those detected by the AI system. Assesses the alignment between player perceptions and AI-based detection, helping to refine the system.

(Number of AI-detected incidents / Number of player-reported incidents) * 100.

KPI #4
‍Moderation Response Time

The time it takes for moderators or automated systems to respond to identified griefing incidents. This evaluates the effectiveness of the moderation system in addressing griefing promptly.

Time taken to resolve incidents from detection to resolution.
‍

KPI #5
‍Rehabilitation Success Rate

Measures the rate at which penalized players cease engaging in griefing behavior after undergoing corrective actions. Assesses the effectiveness of penalties and interventions in deterring future griefing.

(Number of players rehabilitated / Total number of penalized players) * 100.

KPI #6
‍Community Sentiment

Measures player sentiment and satisfaction with the game community environment. Gauges the overall impact of griefing on the community atmosphere and player enjoyment.
‍
‍

Surveys, sentiment analysis of player communications, or feedback scores.
Finally, we would like to make some recommendations for game design and UX design to help mitigate griefing incidents. Each of these guidelines can be assessed objectively by monitoring desired changes in one or more of the KPIs we have outlined so far. We will discuss each recommendation along with pointing out the relevant KPI to consider.

UX Design Guidelines

Game Design Guidelines

#1
Clear and Accessible Reporting and Feedback Systems‍

Ensure that reporting mechanisms for griefing are easily accessible and well-promoted within the user interface. Provide an easy-to-use reporting system that allows players to report griefing incidents. Track cases and provide feedback on the status of reports.

KPI alignment: Compare Player Reports vs. AI Detection to evaluate the effectiveness of the reporting system. Use player feedback to refine detection algorithms.



#2
Player Education with a Clear Code of Conduct‍

Educate players about the consequences of griefing during the onboarding process. Establish and communicate a clear code of conduct that outlines acceptable and unacceptable behavior within the game.

KPI alignment: Monitor the Griefing Rate and Incident Severity to assess the impact of clearer communication measures and code of conduct on player behavior.
‍


#3
Transparent Penalty and Rehabilitation Systems‍

Implement a clearly communicated, graduated penalty system for griefing, ranging from warnings to temporary bans. Develop rehabilitation programs to guide penalized players toward positive behavior.

KPI alignment:  Track the Rehabilitation Success Rate and Moderate Response Time to assess the effectiveness of penalties and rehabilitation.



#4
User-Friendly Mute and Block Features

Offer intuitive mute and block features that allow players to manage their own interactions and reduce exposure to potential griefers.

KPI alignment:  Enhances individual player experience, indirectly impacting incident severity and Player Retention Impact.
‍
#1
Balanced Game Design‍

Strive for a balanced game design that minimizes opportunities for griefing. Consider game mechanics that discourage unsportsmanlike behavior.

KPI alignment: Analyze the Griefing Rate and Incident Severity in correlation with changes in game design.


‍
#2
Player Rewards for Positive Behavior

Introduce reward systems for positive and cooperative gameplay. Encourage and acknowledge fair play and encourage a more supportive gaming community.

KPI alignment: Assess the impact of reward systems on the Griefing Rate and Player Retention Impact.



#3
Personalized Matchmaking Preferences and Dynamic Adjustments

Allow players to set personalized matchmaking preferences based on their desired playstyles and tolerance for competition. Allow skill-based matchmaking to ensure that players are matched with others of similar skill levels. Consider implementing anonymous matchmaking to reduce the likelihood of targeted harassment between players.



#4
Dynamic Spawning Mechanisms

‍Implement dynamic spawning mechanisms to prevent spawn camping and create fair starting conditions for all players.

KPI alignment: Can positively influence Griefing Rate and Player Retention Impact.
‍

Conclusion

Our Vision

  • A robust framework for any scale!
    ‍‍
    We were hoping that the insights put together on this report not only serve as a blueprint for a large studio to implement a data analytics infrastructure, but also perhaps compels relatively young independent studios to start laying down the groundwork for anti-griefing design by using actionable insights from the KPI and guidelines section! Hopefully, the interconnected of the three sections also allow for the company to decide which direction they want to go in term of granularity.
    ‍
    ‍

Related Projects

UMD iConsultancy UX Research

Finding the main stressors related to finances international students face in the US

Star Wars Battlefront II: Menu Redesign

Redesigning the main-menu to generate a more efficient user flow