Johannes Pfau

Hey, I'm Jo!

I'm a researcher (currently: Assistant Professor @UU) passionate about (video) games, AI, data (and octopi). Scroll through my page to see some showcases of my work for industry and academia.
For questions or collaborations, don't hesitate to reach out!

Games Involved

Titles in which some of my work took place.

Guild Wars 2

(ArenaNet, 2012)

I am currently running Wingman, the largest endgame analytics platform for Guild Wars 2, which sparked a series of tools and studies anchored in a vast abundance of behavioral data and a close connection to the player community. 7,8,9,10,11,14,18
Aion

(NCSoft, 2009)

Working together with an AION Server, I was able to develop a custom daily dungeon that featured an individual DPBM replica end boss and DDA mechanics, before I utilized player models of a whole community for automated game balancing simulations. 2,4,5
Lineage II

(NCSoft, 2003)

To accumulate initial, low-level player behavior data and benchmark DPBM accuracy and replayability, I recorded action patterns of Lineage II players and trained individual replicas from these. 1
League of Legends

(Riot Games, 2009)

Being one of the most popular games of our times, LoL enabled a series of timely studies on play patterns, practice habits, and process visualization. 20,28,29
Baldur's Gate 3

(Larian Studios, 2023)

Damage mechanics become immensily complex with game titles rich in variety across equipment, actions, passives and other choices. Baldur's Gate 3 seemed to be an apt candidate to model, simulate and optimize in my work about co-creative damage optimization. 8
Borderlands 3

(Gearbox Software, 2019)

Damage mechanics become immensily complex with game titles rich in variety across equipment, actions, passives and other choices. Borderlands 3 seemed to be an apt candidate to model, simulate and optimize in my work about co-creative damage optimization. 8
Fallout: New Vegas

(Obsidian Entertainment, 2010)

Using the Virtual Personality Assessment Lab (VPAL) mod to Fallout: New Vegas, me and my students investigated how to visualize player behavior patterns, extract meaningful differences, and predict the impact of personality. 15,16,17,24
Anna's Quest

(Daedalic Entertainment, 2015)

During my work with Daedalic, I developed ICARUS, a Reinforcement Learning tool integrated into autonomous game testing routines for all in-house products generically. 11
The Pillars of the Earth

(Daedalic Entertainment, 2017)

During my work with Daedalic, I developed ICARUS, a Reinforcement Learning tool integrated into autonomous game testing routines for all in-house products generically. 11
Deponia

(Daedalic Entertainment, 2014)

During my work with Daedalic, I developed ICARUS, a Reinforcement Learning tool integrated into autonomous game testing routines for all in-house products generically. 11
Korona:Nemesis

(Nevermind Creations, 2019)

To assess whether players could differentiate heuristic bots from DPBM agents in online multiplayer games, I developed a projectile-based fighting game and published it on Steam. 3,5
Hades

(Supergiant Games, 2020)

What would happen if the end boss after a long run of Hades is not your father, but your former self? We used DPBM to do exactly that!
Super Mario Bros.

(Nintendo, 1985)

As an all-time classic for speedrunning and games utilized for science, we modified feedback mechanisms in Super Mario Bros. and studied the impact on performance improvement. 27
Mirror's Edge

(DICE, 2008)

Mirror's Edge parcour time trials offer exciting challenges to best yourself and others - We looked into how mechanical affordances of such trials can be modeled and generated procedurally.
DNF Duel

(Arc System Works, 2022)

Consisting of muscle memory, response selection and reaction pace, fighting game combos make up for a precise skill to study feedback mechanisms and training paradigms on. 30
Kitchen Clash

(University of Bremen, 2019)

Kitchen Clash is a human computation VR game I developed to record human decision making and movement trajectories for the robotic training database OpenEASE. 32,33,34,35,36
Adaptify

(University of Bremen, 2018)

To harness the capabilities of games to sustain long-term motivation, we developed several applied games that helped patients staying tuned to their physical therapy. 37
The Lux

(UC Santa Cruz, 2021)

Resilience is a core competence to a healthy and robust life, yet how to achieve or even measure this cleanly is still a challenge. Using ARGs, multi-modal measures, process mining and AI, we investigated how to quantify resilience in the wild. 22,23,25,26
Parallel

(Drexel University, 2017)

How to program in a proper parallel fashion is a challenge we tackled within the Open Player Modeling project. Recorded behavior from play sessions of students furthermore revealed explanations for successful strategies and how failure patterns form. 14,17,18
Maniacs

(Nevermind Creations, 2024)

A cooperative/competitive card game to escape from a mental asylum, fabricated with methods from generative AI.

Click here for action

Projects

Deep Player Behavior Modeling

In my dissertation project about Deep Player Behavior Modeling, I extended the current state of the art of Imitation Learning by a new approach and several evaluations applied in ecologically valid environments. Up to then, most (generative) player modeling techniques were encompassing larger sets of players to represent something believable, human-like or behavior that would emerge in other unique ways. For the application cases in the following however, I highly emphasize the value of replicating individual playing styles. To realize this technique, atomic state-action pairs are recorded for all symbolic player inputs, so that game state, player state, target state and action history are mapped to a distribution among predicted following actions1. This results in a player-like, continuously learning foe, of which game communities reported to recognize their own behavioral patterns and enjoyed the steadily spiraling challenge2,6. It went even so far that, in online multiplayer sessions, players did not notice when disconnected fellow players were replaced with their individual substitutes3. To scale it up, we showed that DPBM can also be used for large-scale automated balancing simulations to closely resemble a whole population of players4,5.

1

Towards Deep Player Behavior Models in MMORPGs.
Pfau, Johannes, Smeddinck, J. D., & Malaka, R. (2018, October). In Proceedings of the 2018 Annual Symposium on Computer-Human Interaction in Play (pp. 381-392).

2

Enemy Within: Long-term Motivation Effects of Deep Player Behavior Models for Dynamic Difficulty Adjustment.
Pfau, Johannes, Smeddinck, J. D., & Malaka, R. (2020, April). In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems (pp. 1-10).

3

Bot or Not? User Perceptions of Player Substitution with Deep Player Behavior Models.
Pfau, Johannes, Smeddinck, J. D., Bikas, I., & Malaka, R. (2020, April). In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems (pp. 1-10).

🏆 Best Paper Finalist:
4

Dungeons & Replicants: Automated Game Balancing via Deep Player Behavior Modeling.
Pfau, Johannes, Liapis, A., Volkmar, G., Yannakakis, G. N., & Malaka, R. In Proceedings of the 2020 IEEE Conference on Games.

5

Dungeons & Replicants II: Automated Game Balancing across Multiple Difficulty Dimensions via Deep Player Behavior Modeling.
Pfau, Johannes, Liapis, A., Yannakakis, G. N., & Malaka, R. In IEEE Transactions on Games, 15(2) (pp. 217-227).

6

Deep Player Behavior Models: Evaluating a Novel Take on Dynamic Difficulty Adjustment.
Pfau, Johannes, Smeddinck, J. D., & Malaka, R. (2019, May). In Extended Abstracts of the 2019 CHI Conference on Human Factors in Computing Systems (pp. 1-6).

Game Analytics

During my PhD, I discovered my passion for game data, which encouraged me to start the now largest platform for Guild Wars 2 endgame content analytics: Gw2 Wingman.
Primarly, I am interested in player-driven analytics, so tools, visualizations and reports that give insights about one's own performances, limitations, progress and more. For this, I compile manifold analytics from millions of combat logs directly derived in liaison with the player community that is longing for such feedback 7,11, deploy simulations and co-creative AI optimization 8,12, and assess and quantify tailored notions of game balance9,10. Such measures and insights are however equally valuable for game industrials13. Using higher-level company data from hundreds and games and studios, I developed an interactive tool to scaffold an understanding of success factors among companies and titles: https://ghs-gceko.ondigitalocean.app/ 21,22
And, together with my PhD student Zhaoqing "Jimmy" Teng, we brought process mining to game analytics, making sense of the contextual information between sequences of game actions or events, and how to interact, interpret and understand these visualizations: https://inspect.nevermindcreations.de/ 14-18.


 
🏆 Honorable Mention:
7

Player-Driven Game Analytics: The Case of Guild Wars 2.
Pfau, Johannes, & Seif El-Nasr, M. (2023, April). In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (pp. 1-14).

8

Damage Optimization in Video Games: A Player-Driven Co-Creative Approach.
Pfau, Johannes, Charan, M., Kleinman, E., & Seif El-Nasr, M. (2024, May). In Proceedings of the CHI Conference on Human Factors in Computing Systems (pp. 1-16).

9

On Video Game Balancing: Joining Player-and Data-Driven Analytics.
Pfau, Johannes, & Seif El-Nasr, M. (2023). In ACM Games: Research and Practice.

10

Balancing Video Games: A Player-Driven Instrument.
Pfau, Johannes, & Seif El-Nasr, M. (2023, October). In Companion Proceedings of the Annual Symposium on Computer-Human Interaction in Play (pp. 187-195).

11

The Real MVP: Quantifying Individual Performances in Multiplayer Online Games.
Pfau, Johannes (2024, August). In 2024 IEEE Conference on Games (CoG) (pp. 1-8). IEEE.

12

Automated Game Testing with ICARUS: Intelligent Completion of Adventure Riddles via Unsupervised Solving.
Pfau, Johannes, Smeddinck, J. D., & Malaka, R. (2017, October). In Extended Abstracts Publication of the Annual Symposium on Computer-Human Interaction in Play.

13

The Case for Usable AI: What Industry Professionals Make of Academic AI in Video Games.
Pfau, Johannes, Smeddinck, J. D., & Malaka, R. (2020, November). In Extended abstracts of the 2020 annual symposium on computer-human interaction in play (pp. 330-334).

14

Player Segmentation with INSPECT: Revealing Systematic Behavior Differences within MMORPG and Educational Game Case Studies.
Teng, Z., Pfau, Johannes, Maram, S. S., & Seif El-Nasr, M. (2022, November). In Extended Abstracts of the 2022 Annual Symposium on Computer-Human Interaction in Play.

15

Interactive Player Journeys: Co-designing a Process Visualization System to Video Game Analytics.
Teng, Z., Pfau, Johannes, Maram, S. S., & Seif El-Nasr, M. (2024, May). In Proceedings of the 19th International Conference on the Foundations of Digital Games (pp. 1-11).

16

Identifying Player Strategies through Segmentation: An Interactive Process Visualization Approach.
Teng, Z., Holmes, J., Dominguez, F., Pfau, Johannes, Escarce Junior, M., & Seif El-Nasr, M. (2024, November). In Joint International Conference on Serious Games.

17

Visualization-based Iterative Segmentation to Augment Video Game Analytics.
Teng, Z., Pfau, Johannes, & El-Nasr, M. S. (2023, August). In 2023 IEEE Conference on Games (CoG) (pp. 1-2). IEEE.

18

Mining Player Behavior Patterns from Domain-Based Spatial Abstraction in Games.
Maram, S. S., Pfau, Johannes, Villareale, J., Teng, Z., Zhu, J., & El-Nasr, M. S. (2023, August). In 2023 IEEE Conference on Games (CoG) (pp. 1-8). IEEE.

19

A Topic Modeling Approach Towards Understanding the Discourse on Religion and Video Games.
Maram, S.S., Pfau, Johannes, Kajar, M.R., & Seif El-Nasr, M. (2024, October). In Proceedings of the ACM on Human-Computer Interaction, (CHI PLAY)

20

“Trust the Process” Examining the Impact of Process Visualizations for Self-Reflection in League of Legends.
Kleinman, E., Xu, J., Pfau, Johannes, & El-Nasr, M. S. (2024, October). In Proceedings of the ACM on Human-Computer Interaction, (CHI PLAY)

21

Predicting Success Factors of Video Game Titles and Companies.
Pfau, Johannes, Debus, M., Juul, J., Lundedal Hammar, E., Canossa, A., & Seif El-Nasr, M. (2022, October). In International Conference on Entertainment Computing.

22

From Teams to Games Connecting Game Development to Game Characteristics.
Lundedal Hammar, E., Canossa, A., Debus, M. S., Pfau, Johannes, Seif El-Nasr, M., Juul, J., & Azadvar, A. (2023, July). In International Conference on Human-Computer Interaction.

Psychology of Play

What good comes from data alone if it does not push us to further understand ourselves?
Together with my PhD students of the Lux project, I investigated in how far we can measure resilience, the core competence of bouncing back from life's struggles and adversities, using in-game activity and communication data. We built multiple Alternate Reality games to derive collaborative problem solving solutions 23, investigate the applicability of different stressor types 24, attributed for differences in personality 26, and envisioned intervention mechanisms based on emotion regulation and coping strategies 27. On another note, I revealed the often unhealthy practice patterns of esports players together with my PhD student Ioannis Bikas 29-30, which besides physical strain and afflictions can also cause training-inhibiting mental fatigue, before looking into alternative feedback mechanisms and schedules 28,31. Eventually, player personality can not only be predictive 26, but also being predicted by in-game choices and play styles 25, which we furthermore utilized to tailor game mechanics to the very player, as a catalyst for increased intrinsic motivation 32.

🏆 Best Paper:
23

A Data-Driven Design of AR Alternate Reality Games to Measure Resilience.
Habibi, R., Maram, S. S., Pfau, Johannes, Wei, J., Sisodiya, S. K., Kashani, A., Carstensdottir, E., & Seif El-Nasr, M. (2022, June). In International Conference on Human-Computer Interaction (pp. 586-604).

24

Under Pressure: A Multi-Modal Analysis of Induced Stressors in Games for Resilience.
Habibi, R., Pfau, Johannes, Maram, S. S., Li, J., Larsen, B., Xu, J., … & El-Nasr, M. S. (2023, April). In Proceedings of the 18th International Conference on the Foundations of Digital Games (pp. 1-10).

25

Modeling Player Personality Factors from In-Game Behavior and Affective Expression.
Habibi, R., Pfau, Johannes, & El-Nasr, M. S. (2023, September). In 2023 11th International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW) (pp. 1-8). IEEE.

26

Assessing the Impact of Personality on Affective States from Video Game Communication.
Kashani, A., Pfau, Johannes, & El-Nasr, M. S. (2023, September). In 2023 11th International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW) (pp. 1-7). IEEE.

27

Empathetic AI for Empowering Resilience in Games.
Habibi, R., Pfau, Johannes, Holmes, J., & El-Nasr, M. S. (2023). In EXAG Workshop, 2023 AAAI Conference on AI for Interactive Digital Entertainment

28

Low Latency Feedback: Contrasting Training Outcomes of Immediate versus Delayed Feedback in Speedrunning Super Mario Bros.
Bikas, I., Pfau, Johannes, Muender, T., & Malaka, R. (2024, October). In Proceedings of the ACM on Human-Computer Interaction, (CHI PLAY)

29

Mental Wear and Tear: An Exploratory Study on Mental Fatigue in Video Games Using the Example of League of Legends.
Bikas, I., Pfau, Johannes, Dänekas, B., & Malaka, R. (2022, October). In International Conference on Entertainment Computing (pp. 125-139).

30

Grinding to a Halt: The Effects of Long Play Sessions on Player Performance in Video Games..
Bikas, I., Pfau, Johannes, Muender, T., Alexandrovsky, D., & Malaka, R. (2023, October). In Companion Proceedings of the Annual Symposium on Computer-Human Interaction in Play (pp. 36-42).

31

Space Out Gaming: Comparing Distributed Practice Sessions with Massed Play.
Bikas, I., Pfau, Johannes, Muender, T., & Malaka, R. (2024, May). In Proceedings of the 19th International Conference on the Foundations of Digital Games (pp. 1-4).

32

Player Types and Achievements – Using Adaptive Game Design to Foster Intrinsic Motivation.
Volkmar, G., Pfau, Johannes, Teise, R., & Malaka, R. (2019, October). In Extended Abstracts of the Annual Symposium on Computer-Human Interaction in Play Companion Extended Abstracts (pp. 747-754).

Applied Games

Apart from entertainment games, I also developed games with a purpose, as well as tools, AI, and analytics for those. As outlined above, these aimed at detecting and quantifying resilience 23-27, sustaining intrinsic motivation for long-term physical therapy 38, or to accumulate human computation input about movement trajectories, tool preferences and contextual action selection for robotic knowledge databases 33-37. Apart from that, I assisted developing applied games for usable security 40-41, investigated repair strategies in voice control games or systems 39, and studied the intersection of religion and video games with my PhD student Sai Siddartha Maram: what topics are being in discussed in either community 19, how to draw affordances from mythological backgrounds 43, and how to do this in a culturally sensitive way 42.


 
33

Give MEANinGS to Robots with Kitchen Clash: A VR Human Computation Serious Game for World Knowledge Accumulation.
Pfau, Johannes, Porzel, R., Pomarlan, M., Cangalovic, V. S., Grudpan, S., Höffner, S., Bateman, J., & Malaka, R. (2019). In Entertainment Computing and Serious Games: First IFIP TC 14 Joint International Conference, ICEC-JCSG 2019.

34

Can You Rely on Human Computation? A Large-scale Analysis of Disruptive Behavior in Games With A Purpose.
Pfau, Johannes, & Malaka, R. (2019, October). In Extended Abstracts of the Annual Symposium on Computer-Human Interaction in Play Companion Extended Abstracts (pp. 605-610).

35

We Asked 100 People: How Would You Train Our Robot?.
Pfau, Johannes, & Malaka, R. (2020, November). In Extended Abstracts of the 2020 Annual Symposium on Computer-Human Interaction in Play (pp. 335-339).

36

Ontology-based Understanding of Everyday Activity Instructions.
Höffner, S., Porzel, R., Hedblom, M. M., Pomarlan, M., Cangalovic, V. S., Pfau, Johannes, Bateman, J., & Malaka, R. (2021). In Semantic Web.

37

Deep Understanding of Everyday Activity Commands for Household Robots.
Höffner, S., Porzel, R., Hedblom, M. M., Pomarlan, M., Cangalovic, V. S., Pfau, Johannes, Bateman, J., & Malaka, R. (2022). In Semantic Web.

38

Do You Think This is a Game? Contrasting a Serious Game with a Gamified Application for Health.
Pfau, Johannes, Smeddinck, J. D., Volkmar, G., Wenig, N., & Malaka, R. (2018, April). In Extended Abstracts of the 2018 CHI Conference on Human Factors in Computing Systems (pp. 1-6).

39

“I Didn’t Catch That, But I’ll Try My Best”: Anticipatory Error Handling in a Voice Controlled Game.
Zargham, N., Pfau, Johannes, Schnackenberg, T., & Malaka, R. (2022, April). In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems (pp. 1-13).

40

Enhancing Game-Based Learning Through Infographics in the Context of Smart Home Security.
Bahrini, M., Zargham, N., Pfau, Johannes, Lemke, S., Sohr, K., & Malaka, R. (2020). In Entertainment Computing–ICEC 2020.

41

Good vs. Evil: Investigating the Effect of Game Premise in a Smart Home Security Educational Game.
Bahrini, M., Zargham, N., Pfau, Johannes, Lemke, S., Sohr, K., & Malaka, R. (2020, November). In Extended Abstracts of the 2020 Annual Symposium on Computer-Human Interaction in Play (pp. 182-187).

42

AstraVerse: Establishing a Culturally Sensitive Framework for Integrating Elements from Mythological Backgrounds.
Maram, S. S., Pfau, Johannes, Habibi, R., & Seif El-Nasr, M. (2022, October). In International Conference on Entertainment Computing (pp. 3-17).

43

A Visual Ethnographic Study at Cultural Spaces to Identify Character Creation Opportunities.
Maram, S. S., Pfau, Johannes, Dodechani, J. B., & Seif El-Nasr, M. (2023, April). In Proceedings of the 18th International Conference on the Foundations of Digital Games.

Teaching

In parallel to my research activities, I also contributed to a number of academic courses in the roles of tutor, guest lecturer, or seminar coordinator. From 2024 onwards, I started to develop and teach the full course AI-Driven Content Generation at Utrecht University 44. In it's design, I ensured coverage of theoretical background knowledge for students from diverse Master's programs, up-to-date advances of this rapidly evolving field, and practical application through weekly hands-on sessions using local models. Over the course of 8 weeks, students learned to understand and create multifaceted pieces of media, games or systems through generative AI, as can be seen best in the AICG 2024 Showreel Video below.

44

AI-Driven Content Generation.
Pfau, Johannes. At Utrecht University (2024-25).

45

Gamification and Applied Games (Seminars).
Bakkes, S.C.J., Pfau, Johannes. At Utrecht University (2024-25).

46

Scientific methods for Computing Science (Guest Lecture).
Bodlaender, H.L., Pfau, Johannes. At Utrecht University (2024-25).

47

Game Data Science (Guest Lectures).
Seif El Nasr, M., Pfau, Johannes. At University of California, Santa Cruz (2021).

48

Entertainment Computing (Guest Lecture, Tutorials).
Malaka, R., Pfau, Johannes. At University of Bremen (2021).

Collaborations

Companies and research instutions I worked with and for.