Human-Like Playtesting with Deep Learning
The DOI matches its Crossref record; 50 of the 53 references it cites were found in a registry or an index.
Abstract
We present an approach to learn and deploy human-like playtesting in computer games based on deep learning from player data. We are able to learn and predict the most "human" action in a given position through supervised learning on a convolutional neural network. Furthermore, we show how we can use the learned network to predict key metrics of new content - most notably the difficulty of levels. Our player data and empirical data come from Candy Crush Saga (CCS) and Candy Crush Soda Saga (CCSS). However, the method is general and well suited for many games, in particular where content creation is sequential. CCS and CCSS are non-deterministic match-3 puzzle games with multiple game modes spread over a few thousand levels, providing a diverse testbed for this technique. Compared to Monte Carlo Tree Search (MCTS) we show that this approach increases correlation with average level difficulty, giving more accurate predictions as well as requiring only a fraction of the computation time.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Researchers and developers have seen much success in bringing AI to bear on major problems in the industry, such as cheat detection in competitive games [26], as well as to improve common development processes such as playtesting, bug reporting, and other labor intensive aspects of game production [17,42].”
From Is Resistance Futile?: Early Career Game Developers, Generative AI, and Ethical Skepticism (2024) · cites this paper as [17]
In the citing paper
“Current narratives of these tools as productivity-enhancers, timesavers, and replacement labor [17,42,44] run counter to the values held by this generation of artists, designers, and (with some exception) programmers.”
From Is Resistance Futile?: Early Career Game Developers, Generative AI, and Ethical Skepticism (2024) · cites this paper as [17]
In the citing paper
“Therefore, researchers proposed methods to automate the playtesting process [1] [2] [3].”
From Playtesting: What is Beyond Personas (2023) · cites this paper as [2]
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“In addition, we also experimented with batch normalization [44] and residual network [45] but neither managed to provide better generalization capability.”
MethodVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 2×
“It is also worth mentioning that adding the classic Fully-Connected (FC) layers with dropout regularization [38], [42] performed inferior to GAP.”
BackgroundMethodImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818 · cited 1×
“In addition, we also experimented with batch normalization [44] and residual network [45] but neither managed to provide better generalization capability.”
MethodMastering the game of Go with deep neural networks and tree search
Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., van den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., et al. (2016) · Nature · doi:10.1038/nature16961 · Cited by 15,612 · cited 4×
“[11] (AlphaGo) managed to improve the prediction accuracy to 55.”
BackgroundMethodInfluentialBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S. (2024) · arXiv (Cornell University) · doi:10.57702/o9raffed · Cited by 14,214
Mastering the game of Go without human knowledge
Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., Hubert, T., Baker, L., Lai, M., Bolton, A., et al. (2017) · Nature · doi:10.1038/nature24270 · Cited by 9,084 · cited 1×
“[32] managed to create an agent that could outperform any previously best artificial and human Go player in the game of Go.”
BackgroundFast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
Clevert, D.-A., Unterthiner, T., & Hochreiter, S. (2015) · arXiv · doi:10.48550/arxiv.1511.07289 · Cited by 4,513 · cited 1×
“[10]–[12], [38], [42]), we opted to use the Exponential Linear Unit (ELU) function [43] instead, because it improved the validation accuracy by about 2.”
MethodA Survey of Monte Carlo Tree Search Methods
Browne, C. B., Powley, E., Whitehouse, D., Lucas, S. M., Cowling, P. I., Rohlfshagen, P., Tavener, S., Perez, D., Samothrakis, S., & Colton, S. (2012) · IEEE Transactions on Computational Intelligence and AI in Games · doi:10.1109/tciaig.2012.2186810 · Cited by 2,970 · cited 1×
“We benchmark this approach against an approach using MCTS [34].”
Method
Research works printed but not linked: 1
- silva, 2016, AI as Evaluator: Search Driven Playtesting in Game Design, Proceedings of AAAI
Cited by
Deep Learning for Video Game Playing
Justesen, N., Bontrager, P., Togelius, J., & Risi, S. (2020) · IEEE Transactions on Games · doi:10.1109/tg.2019.2896986 · Cited by 257
“Recently, the game company King published results using imitation learning to learn policies for playtesting of Candy Crush levels, showing a promising direction for new design tools [41].”
BackgroundDeep learning for procedural content generation
Liu, J., Snodgrass, S., Khalifa, A., Risi, S., Yannakakis, G. N., & Togelius, J. (2021) · Neural Computing and Applications · doi:10.1007/s00521-020-05383-8 · Cited by 143
“A CNN was trained on human player data, and then used to predict the action that human players most likely to select when playing levels that were unseen during training [35].”
MethodAutomated video game testing using synthetic and humanlike agents
Ariyurek, S., Betin-Can, A., & Surer, E. (2021) · IEEE Transactions on Games · doi:10.1109/tg.2019.2947597 · Cited by 66
“Humanlike agents are better suited to analyze the difficulty of the game [10], can become genuine opponents [11],”
BackgroundMethodDeep Learning Application Pros And Cons Over Algorithm
Moshayedi, A. J., Roy, A. S., Kolahdooz, A., & Shuxin, Y. (2022) · EAI Endorsed Transactions on AI and Robotics · doi:10.4108/airo.v1i.19 · Cited by 135
“(2018) [100] showed that it is possible to predict the next most “human” action, and from that, they can update future contents of the game.”
BackgroundAugmenting Automated Game Testing with Deep Reinforcement Learning
Bergdahl, J., Gordillo, C., Tollmar, K., & Gisslen, L. (2020) · 2020 IEEE Conference on Games (CoG) · doi:10.1109/cog47356.2020.9231552 · Cited by 91
“vised learning using human game play data has successfully been applied to testing in 2D mobile puzzle games by predicting human-like actions to evaluate the playability of game levels [6].”
MethodFrom Chess and Atari to StarCraft and Beyond: How Game AI is Driving the World of AI
Risi, S., & Preuss, M. (2020) · Künstliche Intell. · doi:10.1007/s13218-020-00647-w · Cited by 94
“Other companies, such as King, are investing heavily in deep learning methods to automatically learn models of players that can then be used for playtesting new levels quickly [25].”
BackgroundPredicting Mid-Air Interaction Movements and Fatigue Using Deep Reinforcement Learning
Cheema, N., Frey-Law, L. A., Naderi, K., Lehtinen, J., Slusallek, P., & Hämäläinen, P. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376701 · Cited by 66
“Computational user models have been successfully applied in, e.g., game playing [28, 70, 72] and typing [54].”
MethodVideo Game Automated Testing Approaches: An Assessment Framework
Albaghajati, A., & Ahmed, M. (2020) · IEEE Transactions on Games · doi:10.1109/tg.2020.3032796 · Cited by 70
“…human data and learning algorithms, or by using specialized and restricted heuristics, or mixing between different algorithms. a) Machine learning [19], [22], [53]–[56]: A semiauto-mated gameplay analysis by machine learning using an active learning approach was presented by Southey et al. [22],…”
BackgroundMethodInfluentialDungeons & Replicants: Automated Game Balancing via Deep Player Behavior Modeling
Pfau, J., Liapis, A., Volkmar, G., Yannakakis, G. N., & Malaka, R. (2020) · 2020 IEEE Conference on Games (CoG) · doi:10.1109/cog47356.2020.9231958 · Cited by 48
“[41] utilized atomic choices in order to predict the difficulty of various levels of a Match-3-Puzzle game.”
MethodTowards automated video game testing: still a long way to go
Politowski, C., Guéhéneuc, Y., & Petrillo, F. D. S. (2022) · doi:10.1145/3524494.3527627 · Cited by 64
“[10] uses autonomous agents predict the difficulty of a newgame level automatically.”
BackgroundInfluentialUsing reinforcement learning for load testing of video games
Tufano, R., Scalabrino, S., Pascarella, L., Aghajani, E., Oliveto, R., & Bavota, G. (2022) · ACM/IEEE International Conference on Software Engineering: Companion Proceedings (ICSE-Companion) · doi:10.1145/3510003.3510625 · Cited by 61
“, power-up item effects) [37, 49, 51] and in testing the game difficulty [25, 43].”
BackgroundMethodInternational Conference on the Foundations of Digital Games
(2020) · doi:10.1145/3402942 · Cited by 49
Predicting Game Difficulty and Churn Without Players
Roohi, S., Relas, A., Takatalo, J., Heiskanen, H., & Hämäläinen, P. (2020) · Proceedings of the Annual Symposium on Computer-Human Interaction in Play · doi:10.1145/3410404.3414235 · Cited by 34
Playtesting in Match 3 Game Using Strategic Plays via Reinforcement Learning
Shin, Y., Kim, J.-W., Jin, K., & Kim, Y. (2020) · IEEE Access · doi:10.1109/access.2020.2980380 · Cited by 31
“The previously proposed algorithms include methods such as MCTS [26], deep neural network (DNN) [27], CNN [28], [29], and reinforcement learning [30].”
MethodIs Resistance Futile?: Early Career Game Developers, Generative AI, and Ethical Skepticism
Boucher, J. D., Smith, G., & Telliel, Y. D. (2024) · Proceedings of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3641889 · Cited by 30
“Current narratives of these tools as productivity-enhancers, time-savers, and replacement labor [17, 42, 44] run counter to the values held by this generation of artists, designers, and (with some exception) programmers.”
BackgroundResultRiverGame - a game testing tool using artificial intelligence
Păduraru, C., Paduraru, M. G., & Ştefănescu, A. (2022) · IEEE International Conference on Software Testing, Verification and Validation (ICST) · doi:10.1109/icst53961.2022.00048 · Cited by 33
Predicting Game Difficulty and Engagement Using AI Players
Roohi, S., Guckelsberger, C., Relas, A., Heiskanen, H., Takatalo, J., & Hämäläinen, P. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3474658 · Cited by 35
“[22] have exploited demonstrations from human players for the supervised training of a deep learning agent to predict level difficulty in match-3 games.”
BackgroundPfau, J., Liapis, A., Yannakakis, G. N., & Malaka, R. (2022) · IEEE Transactions on Games · doi:10.1109/tg.2022.3167728 · Cited by 42
“…in this work, Holmgård et al. [15] conflated atomic player behavior into procedural personas to simulate and test different play styles in a Dungeon Crawler game and Gudmundsson et al. [47] utilized atomic choices in order to predict the difficulty of various levels of a Match-3 -Puzzle game.”
BackgroundDeep Reinforcement Learning in Match-3 Game
Kamaldinov, I., & Makarov, I. (2019) · IEEE Conference on Games (CoG) · doi:10.1109/cig.2019.8848003 · Cited by 25
“Key research in Match-3 agents are King’s projects [1], [10], [11], [12].”
Navigates Like Me: Understanding How People Evaluate Human-Like AI in Video Games
Milani, S., Juliani, A., Momennejad, I., Georgescu, R., Rzepecki, J., Shaw, A., Costello, G., Fang, F., Devlin, S., & Hofmann, K. (2023) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3581348 · Cited by 36
“In automated playtesting of games [23, 68], AI agents that act like real users would enable video game designers to expedite the iterative development process while also alleviating the burden of game players to extensively evaluate new content.”
BackgroundAutomated game testing using computer vision methods
Păduraru, C., Paduraru, M. G., & Ştefănescu, A. (2021) · IEEE/ACM International Conference on Automated Software Engineering Workshops (ASEW) · doi:10.1109/asew52652.2021.00024 · Cited by 27
Blood Bowl: A New Board Game Challenge and Competition for AI
Justesen, N., Uth, L. M., Jakobsen, C., Moore, P. D., Togelius, J., & Risi, S. (2019) · IEEE Conference on Games (CoG) · doi:10.1109/cig.2019.8848063 · Cited by 25
“This technique has been applied to several games, including StarCraft [5], and Candy Crush Saga [3].”
MethodTowards Interactive Training of Non-Player Characters in Video GamesPreprint
Borovikov, I., Harder, J., Sadovsky, M. G., & Beirami, A. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1906.00535 · Cited by 29
“Using human demonstrations helps training artificial agents in many applications and in particular in video games (Gudmundsson et al., 2018), (Harmer et al.”
BackgroundStrategies for Using Proximal Policy Optimization in Mobile Puzzle Games
Kristensen, J. T., & Burelli, P. (2020) · International Conference on the Foundations of Digital Games (FDG) · doi:10.1145/3402942.3402944 · Cited by 22
“color or gravity), one-hot encoding each board piece by the unique combination of attributes may lead to a very large and sparse representation, as seen in [9].”
BackgroundEstimating Player Completion Rate in Mobile Puzzle Games Using Reinforcement Learning
Kristensen, J. T., Valdivia, A., & Burelli, P. (2020) · IEEE Conference on Games (CoG) · doi:10.1109/cog47356.2020.9231581 · Cited by 21
“, [2]) using the rank correlation can instead be used to give an estimate on how a level is compared to other levels.”
BackgroundMethodAutomated Playtesting with a Cognitive Model of Sensorimotor Coordination
Lee, I., Kim, H., & Lee, B. (2021) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3474085.3475429 · Cited by 16
“In other words, data collected from playtest-participants are useful in revealing potential problems of the game but cannot be generalized to players of a population level [9, 17].”
BackgroundMethodInfluentialRaidEnv: Exploring New Challenges in Automated Content Balancing for Boss Raid Games
Jeon, H., Baek, I.-C., Bae, C.-M., Park, T., You, W., Ha, T.-G., Jung, H., Noh, J.-H., Oh, S., & Kim, K.-J. (2023) · IEEE Transactions on Games · doi:10.1109/tg.2023.3335399 · Cited by 20
“Weadoptthewin ratemetrictoevaluateanagent’sperformanceinanenvironment, andthewinrateisawidelyusedmetricforagent-basedplay testing[2],[47];thewinrateiscalculatedbydividingthenumber ofwinsbytotalnumberofepisodes.”
BackgroundReleasing the Power of In-Network Aggregation With Aggregator-Aware Routing Optimization
Luo, S., Yu, X., Li, K., & Xing, H. (2024) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2024.3423380 · Cited by 14
“Nowadays, machine learning (ML), especially deep learning, has demonstrated great capabilities and achieved great success in abundant fields like machine vision [1], natural language processing [2], weather prediction [3], content generation [4], and game playing [5].”
Advances in AI-assisted Game Testing
Roohi, S. (2022) · Aaltodoc (Aalto University) · Cited by 16
Monster Carlo 2: Integrating Learning and Tree Search for Machine Playtesting
Keehl, O. G., & Smith, A. M. (2019) · IEEE Conference on Games (CoG) · doi:10.1109/cig.2019.8847989 · Cited by 13
“1) Input Representation: Our game state representation (input to a neural network) was inspired by King’s experiments used for testing Candy Crush [2], namely, having a binary (0/1) channel for each relevant feature of a cell in the game’s grid.”
BackgroundMethodInfluentialStatistical Modelling of Level Difficulty in Puzzle Games
Kristensen, J. T., Valdivia, A., & Burelli, P. (2021) · IEEE Conference on Games (CoG) · doi:10.1109/cog52621.2021.9619050 · Cited by 12
“[6]), in scenarios where it is necessary to operationalise difficulty, such as for dynamic difficulty adjustment or automated playtesting, it is common to use the probability of task success as an objective measure of difficulty [5], [7], [9], [13], [15], [17].”
BackgroundAugmenting Automated Game Testing with Deep Reinforcement LearningPreprint
Bergdahl, J., Gordillo, C., Tollmar, K., & Gisslén, L. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.15819 · Cited by 12
“vised learning using human game play data has successfully been applied to testing in 2D mobile puzzle games by predicting human-like actions to evaluate the playability of game levels [6].”
MethodToward Automated Assessment of User Experience in Extended Reality
Ansari, S. G. (2020) · IEEE International Conference on Software Testing, Verification and Validation (ICST) · doi:10.1109/icst46399.2020.00056 · Cited by 9
“in [5] a framework is proposed for estimating level difficulty of match-3 games in which the CNN-based agent is trained on human-player data.”
BackgroundAstroBug: Automatic Game Bug Detection Using Deep Learning
Azizi, E., & Zaman, L. (2024) · IEEE Transactions on Games · doi:10.1109/tg.2024.3402626 · Cited by 16
“Gudmundsson et al. [7] presented a novel framework for estimating the difficulty levels of Match-3-Puzzle games, utilizing a CNN trained on human-played data.”
Assessing Video Game Balance using Autonomous Agents
Politowski, C., Petrillo, F. D. S., El-Boussaidi, G., Ullmann, G. C., & Guéhéneuc, Y. (2023) · IEEE/ACM International Workshop on Games and Software Engineering (GAS) · doi:10.1109/gas59301.2023.00011 · Cited by 12
“[6] tested the difficulty of a game level using autonomous agents, powered by a CNN, to simulate gameplay and the “success rate” against human players.”
MethodKristensen, J. T., & Burelli, P. (2024) · International Journal of Computer Games Technology · doi:10.1155/2024/5592373 · Cited by 11
“In the specific case of puzzle games, (deep) reinforcement learning (RL) and Monte-Carlo Tree Search (MCTS) based approaches are among the most commonly used to estimate the difficulty of levels [17, 38–42].”
BackgroundMethodResultInfluentialA Literature Review of Software Testing Practices and Frameworks in the Video Gaming Industry
Roque, A., Sotomayor, J. P., Santiago, D., & Clarke, P. (2025) · Software Testing Verification and Reliability · doi:10.1002/stvr.70001 · Cited by 12
Analyzing cross-platform gaming experiences using topic modeling
Sim, Y., Chung, T., & Park, I. (2025) · Entertainment Computing · doi:10.1016/j.entcom.2025.100946 · Cited by 11
A Lightweight Approach of Human-Like Playtest for Android Apps
Zhao, Y. T., Tang, E., Cai, H., Guo, X. W., Wang, X., & Meng, N. (2022) · IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER) · doi:10.1109/saner53432.2022.00047 · Cited by 8
“Some approaches use machine learning (ML) to test games by training models [7], [8].”
BackgroundMethodInfluentialPlaytesting: What is Beyond Personas
Ariyurek, S., Surer, E., & Betin-Can, A. (2023) · IEEE Transactions on Games · doi:10.1109/tg.2022.3165882 · Cited by 7
“Therefore, researchers proposed methods to automate the playtesting process [1] [2] [3].”
MethodPersonalized Game Difficulty Prediction Using Factorization Machines
Kristensen, J. T., Guckelsberger, C., Burelli, P., & Hämäläinen, P. (2022) · ACM Symposium on User Interface Software and Technology (UIST) · doi:10.1145/3526113.3545624 · Cited by 7
“[18] where an AI game-playing agent was used to extract a preliminary estimate of the level pass rate.”
MethodPiergigli, D., Ripamonti, L. A., Maggiorini, D., & Gadia, D. (2019) · IEEE Conference on Games (CoG) · doi:10.1109/cig.2019.8848061 · Cited by 10
“, [11,12,13,14]).”
World of Bugs: A Platform for Automated Bug Detection in 3D Video Games
Wilkins, B., & Stathis, K. (2022) · IEEE Conference on Games (CoG) · doi:10.1109/cog51982.2022.9893616 · Cited by 12
“Automated bug detection (ABD) is an emerging field of study that is receiving increasing attention in the video games research community [1, 4, 5, 12, 13, 10, 14].”
BackgroundLearning a Behavioral Repertoire from Demonstrations
Justesen, N., González-Duque, M., Cabarcas, D., Mouret, J.-B., & Risi, S. (2020) · IEEE Conference on Games (CoG) · doi:10.1109/cog47356.2020.9231897 · Cited by 7
“Methods based on IL, also known as Learning from Demonstration (LfD), have shown promise in the field of robotics [15; 16; 17; 18; 19] and games [6; 10; 20; 21; 22; 23; 24].”
MethodAutomatic Bug Detection in Games using LSTM Networks
Azizi, E., & Zaman, L. (2023) · IEEE Conference on Games (CoG) · doi:10.1109/cog57401.2023.10333253 · Cited by 9
“With the rise of Artificial Intelligence (AI), game developers now have a new tool to detect bugs in their games [5] [8] [9].”
BackgroundA concise review of intelligent game agent
Li, H., Pang, X., Sun, B., & Liu, K. (2024) · Entertainment Computing · doi:10.1016/j.entcom.2024.100894 · Cited by 9
A Lightweight Approach of Human-Like PlaytestingPreprint
Zhao, Y. T., Zhang, W., Tang, E., Cai, H., Guo, X. W., & Meng, N. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2102.13026 · Cited by 6
“Several approaches were introduced to automate game testing [24, 30, 39, 40, 41, 45].”
BackgroundMethodExploring the Dynamic Properties of Interaction in Mixed-Initiative Procedural Content Generation
Alvarez, A. (2020) · doi:10.24834/isbn.9789178771400 · Cited by 6
Generating Personas for Games with Multimodal Adversarial Imitation Learning
Ahlberg, W., Sestini, A., Tollmar, K., & Gisslén, L. (2023) · 2023 IEEE Conference on Games (CoG) · doi:10.1109/cog57401.2023.10333167 · Cited by 5
“Supervised deep learning has been used to simulate humanlike playtesting to predict level difficulty [14]; however, the”
MethodXu, B., Zhao, S., Wu, R., Huang, Z., Wang, J., Hu, Z., Wang, K., Liu, H., Lv, T., Li, L., et al. (2025) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3711896.3736929 · Cited by 7
“It emphasizes fundamental agents and encompasses techniques ranging from rule-based [12, 25, 35, 36, 70] to Reinforcement Learning (RL)-based approaches [20, 52, 67].”
Zero-Shot Reasoning: Personalized Content Generation Without the Cold Start Problem
Hafnar, D., & Demšar, J. (2024) · IEEE Transactions on Games · doi:10.1109/tg.2024.3421590 · Cited by 7
“. board_width”
ARAPID: Towards Integrating Crowdsourced Playtesting into the Game Development Environment
Paranthaman, P. K., & Cooper, S. (2019) · Annual Symposium on Computer-Human Interaction in Play · doi:10.1145/3311350.3347163 · Cited by 5
“Uncovering design issues early on through frequent playtest-ing is a crucial aspect of the iterative design process in game development [19, 46, 26, 24, 50, 51, 61].”
BackgroundTesting match-3 video games with Deep Reinforcement LearningPreprint
Napolitano, N. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2007.01137 · Cited by 5
“it it would be therefore desirable to have a synthetic system to simulate the behavior of the human players [6].”
BackgroundA Language-guided Acceleration Method for Smoke Testing of Game Quests
Hu, J., Zhang, M., Liu, B., Wu, Y., & Chen, Y. (2024) · IEEE International Symposium on Software Reliability Engineering Workshops (ISSREW) · doi:10.1109/issrew63542.2024.00039 · Cited by 4
“Some works trained human-like agents that imitate human testers [5] or human players [6] to produce test sequences.”
Emergence of Player Tactics by expert-guided Machine Learning: An industry tower defence case study
Anderson, S., & Falconer, R. E. (2025) · Entertainment Computing · doi:10.1016/j.entcom.2025.100963 · Cited by 4
Kalafatis, E., Mitsis, K., Zarkogianni, K., Αθανασίου, Μ., & Nikita, K. S. (2025) · IEEE Transactions on Games · doi:10.1109/tg.2025.3589439 · Cited by 5
“The employment of deep learning (DL) techniques in games has recently been investigated, with applications in many settings, from PCG [12] to GT [13].”
Interactive Design Exploration of Game StagesUsing Adjustable Synthetic Testers
Suetake, H., Fukusato, T., Cruz, C. A., Nealen, A., & Igarashi, T. (2020) · International Conference on the Foundations of Digital Games (FDG) · doi:10.1145/3402942.3402982 · Cited by 4
“The use of bots as synthetic testers has been proposed in the past [Gaina et al. 2017; Gudmundsson et al. 2018; Holmgård et al. 2018; Isaksen et al. 2017; Khalifa et al. 2016; Nelson 2011].”
MethodAutomated Game Testing With Online Search Agent and Model Construction, a Study
Shirzadehhajimahmood, S., Prasetya, I. S. W. B., Prandi, D., Dignum, F. P. M., Dastani, M. M., & Keller, G. (2025) · Software Testing Verification and Reliability · doi:10.1002/stvr.70002 · Cited by 4
Lineup Mining and Balance Analysis of Auto Battler
Xu, J., Chen, S., Zhang, L., Wang, J., Zhang, C., Jing, Y., Wang, Z., & Zhu, X. (2020) · International Conference on Aviation Safety and Information Technology (ICASIT) · doi:10.1145/3434581.3434611 · Cited by 3
CandyRL: A Hybrid Reinforcement Learning Model for Gameplay
Karimi, S., Asadi, S., Lorenzo, F., & Payberah, A. H. (2022) · International Conference on Machine Learning and Applications · doi:10.1109/icmla55696.2022.00098 · Cited by 3
“Figure 1 depicts an example of the three-dimensional representation of the board [7].”
BackgroundInfluentialDeep ensemble learning of tactics to control the main force in a real-time strategy game
Han, I., & Kim, K.-J. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-15742-x · Cited by 3
Lineup Mining and Balance Analysis of Auto Battler
Xu, J., Chen, S., Zhang, L., & Wang, J. (2020) · IEEE International Conference on Computer and Communications (ICCC) · doi:10.1109/iccc51575.2020.9345074 · Cited by 3
“Linear regression is a commonly applied regression algorithm with an overall relatively fast modeling speed in the case of a large amount of data and no complicated calculations [17, 18].”
MethodA human centered approach to logic puzzles.
Lynch, A. M. (2021) · Annual Symposium on Computer-Human Interaction in Play · doi:10.1145/3450337.3483515 · Cited by 3
Automated playtesting of platformer games using reinforcement learning
Sriram, V. (2019) · Digital Repository Service (Northeastern University) · doi:10.17760/d20335186 · Cited by 9
A Mapping of Recording-based Game Test Automation Tools
Mioto, V., & Petrillo, F. D. S. (2025) · IEEE/ACM International Workshop on Games and Software Engineering (GAS) · doi:10.1109/gas66647.2025.00006 · Cited by 3
“In fact, game companies like Electronic Arts (EA), UBISOFT, KING, and Activision in-vested in this approach, their results show how the models can adapt to different scenarios during the gameplay and enhance the test coverage [36]–[38].”
Towards LLM-Based Automatic Playtest
Zhao, Y., & Tang, C. (2025) · ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE) · doi:10.1145/3696630.3728700 · Cited by 3
“However, DL-based [21],[6],[14], [25] or RL-based tools [17] [32] cannot be directly applied to new games without significant retraining, requiring extensive data and computational resources.”
Real-Time Error Detection In Digital Games Based on the YOLO V10 Model
Savran, M., & Bulut, H. (2024) · International Artificial Intelligence and Data Processing Symposium (IDAP) · doi:10.1109/idap64064.2024.10710659 · Cited by 3
“Some studies use deep reinforcement learning or deep learning techniques to create human-like agents and detect bugs based on game state data [14,15].”
BackgroundUse All Your Skills, Not Only The Most Popular Ones
Lorenzo, F., Asadi, S., Karnsund, A., Cao, L., Wang, T., & Payberah, A. H. (2020) · IEEE Conference on Games (CoG) · doi:10.1109/cog47356.2020.9231638 · Cited by 2
“The state space of the environment consists of a three-dimensional representation of the board (i.e., 9 × 9 × 32 ), where the third dimension is a one-hot encoding of any of the 32 different elements available in the levels we use here, each associated to a binary layer [11].”
MethodSelf-citationVideo Game's Mechanics Generation and Balancing
Sahibgareeva, G. F., Kugurakova, V. V., & Bolshakov, E. S. (2022) · Naučnyj servis v seti internet · doi:10.20948/abrau-2022-6 · Cited by 4
Exploring Game Design through Human-AI Collaboration
Alvarez, A. (2022) · doi:10.24834/isbn.9789178773084 · Cited by 8
Comparative Analysis of GAT and BERT for Human-Like Playtesting
Fragkedaki, K., Panagiotakopoulos, T., Biasielli, M., & Wang, H. (2025) · IEEE Conference on Games (CoG) · doi:10.1109/cog64752.2025.11114390 · Cited by 1
“Specifically, we explore two alternative deep learning approaches for automated playtesting, BERT and Graph Attention Networks (GAT), and compare them to the CNN-based agent proposed by [7], which serves as our benchmark.”
InfluentialMoving towards automated game play-testing
Malosto, C. G. D. A., Campos, L. C. D., & Knop, I. D. O. (2025) · Simpósio Brasileiro de Jogos e Entretenimento Digital (SBGames) · doi:10.5753/sbgames_estendido.2025.10236 · Cited by 1
“In the works of [Gudmundsson et al. 2018] and [Zook et al. 2019], MCTS is used alongside CNNs.”
BALANCING INTRANSITIVE RELATIONSHIPS IN MOBA GAMES USING DEEP REINFORCEMENT LEARNING
Exton, C. S. A. C. (2020) · IADIS International Conferences Interfaces and Human Computer Interaction; Game and Entertainment Technologies; and Computer Graphics, Visualization, Computer Vision and Image Processing (IHCI/GET/CGVCVIP) · doi:10.33965/ihci_get2020_202010l016 · Cited by 1
“The most relevant research in this area was carried out by King and evaluated the possibility of using Deep Learning to achieve Human performance playtesting on their match 4 "Crush Saga" games (Stefan Freyr Gudmundsson, 2018).”
MethodAdvances in Measuring Inflation Within Virtual Economies Using Deep Reinforcement Learning
Stephens, C., & Exton, C. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-10161-8_16 · Cited by 1
Cautious Curiosity: A Novel Approach to a Human-Like Gameplay Agent
Zhou, C., Machado, T., & Harteveld, C. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment · doi:10.1609/aiide.v19i1.27533 · Cited by 3
“(Gudmundsson et al. 2018) trained a convoluted neural network that takes player data from Candy Crush Saga (CCS) and Candy Crush Soda Saga (CCSS) as input to predict human actions in a given position.”
BackgroundBayesian Optimization-based Search for Agent Control in Automated Game Testing
Celemin, C. E. (2024) · IEEE Conference on Games (CoG) · doi:10.1109/cog60054.2024.10645653 · Cited by 3
Assessing Multiplayer Level Design Using Deep Learning Techniques
Stephens, C., & Exton, C. (2020) · International Conference on the Foundations of Digital Games (FDG) · doi:10.1145/3402942.3409789 · Cited by 1
Boosting Exploration of Low-Dimensional Game Spaces with Stale Human Demonstrations
Chang, K., & Smith, A. M. (2021) · IEEE Conference on Games (CoG) · doi:10.1109/cog52621.2021.9619154 · Cited by 1
“pressing pre-defined button sequences to traverse menus [1]) towards robust, search-based testing using agents trained on human demonstration data [15].”
BackgroundRinascimento: searching the behaviour space of SplendorPreprint
Bravi, I., & Lucas, S. M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.08371 · Cited by 1
“Another set of games where AI playtesting was adopted is Candy Crush Saga and Candy Crush Soda Saga (Gudmundsson et al. 2018), the authors trained a Convolutional Neural Network (CNN) to predict the most humanlike actions based on human data and used the CNN to playtest levels and predict their…”
BackgroundAleksić, V., & Politis, D. (2021) · International Scientific Conference Sinteza · doi:10.15308/sinteza-2021-16-20
“Efficient fast GPUs and affordable video memory enabled DL methods to surpass image and speech recognition applications towards even more complex cases, such as automatically learning player models that are used for fast artificial playtesting of new levels during game development [11] (Fig.”
BackgroundUlyanov, D., & Savelyev, D. (2021) · International Conference on Information Technology and Nanotechnology (ITNT) · doi:10.1109/itnt52450.2021.9649189
Image to level? analysing the playability of image-based, automatically-generated platformer levels
Konda, S. T. (2021) · Digital Repository Service (Northeastern University) · doi:10.17760/d20416818
One Game Fits All: Personalized Content Generation in Mobile Games
Hafnar, D., & Demšar, J. (2022) · Uporabna informatika · doi:10.31449/upinf.179
“Gudmundsson et al. [6], researchers from game developer King, tested human-like playtesting in computer games based on ML trained on player data.”
BackgroundAnokhin, A., Сергеевич, П. Д., Садовникова, Н. П., Финогеев, А. А., & Gurtyakov, A. (2023) · Международный журнал Программные продукты и системы · doi:10.15827/0236-235x.141.046-059 · Cited by 5
Automated Game Design Testing Using Machine Learning
Stephens, C., & Exton, C. (2023) · Springer eBooks · doi:10.1007/978-3-319-08234-9_385-1
Research on Wordle based on time series ARIMA model and XGBoost model
Li, F., Guo, P., & Liao, Z. (2023) · International Conference on Computers, Information Processing and Advanced Education (CIPAE) · doi:10.1109/cipae60493.2023.00046
“After inputting the word, the XGBoost (eXtreme Gradient Boosting) Model is established to obtain the percentage distribution of the number of tries players has used to know the answer [8,9] .”
MethodAutomated Game Design Testing Using Machine Learning
Stephens, C., & Exton, C. (2024) · Springer eBooks · doi:10.1007/978-3-031-23161-2_385
Maximize Score in stochastic match-3 games using reinforcement learning
Roayaei, M., & Afroughrh, A. (2024) · Archives de l'Institut d'Hessarek. · doi:10.61186/jsdp.20.4.129
Modeling Player Types with LLMs: A Framework for Belief- and Motivation-Driven NPC Behavior
Starace, J., & Soule, T. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-10518-9_21
Android Game Testing using Reinforcement Learning
Khurana, S. (2023) · VTechWorks (Virginia Tech)
Tanskanen, V., Rajani, C., Hämäläinen, P., Guckelsberger, C., & Klami, A. (2024) · Aaltodoc (Aalto University)
Kvantitatiivinen tutkimus vaikutuskykyä maksimoivan ja yleisen peliagentin vuorovaikutuksesta
Mäkipää, A. (2022) · Aaltodoc (Aalto University)
What makes an action sequence enjoyable to watch?Preprint
Chou, J.-P., Zheng, K., Chu, J., Agrawala, M., & Fan, J. E. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.30864
“We operationalized difficulty as the probability of failing to complete an obstacle course (Gudmundsson et al., 2018; Isaksen et al., 2017).”
Metadata
- Authors
- Gudmundsson, S. F., Eisen, P., Poromaa, E., Nodet, A., Purmonen, S., Kozakowski, B., Meurling, R., & Cao, L.
- ORCID
- Cao, Lele
- Venue
- 2018 IEEE Conference on Computational Intelligence and Games (CIG)
- Publisher
- IEEE
- Published
- 2018-08
- Pages
- 1-8
- OpenAlex
- W2897406577
- Cited by
- 90
- References
- 53
- Citations by year
- 2025: 10 · 2024: 12 · 2023: 10 · 2022: 15 · 2021: 12 · 2020: 21 · 2019: 10
- FWCI
- 4.39
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- IEEE Conference on Computational Intelligence and Games (CIG) (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence in Games
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Social Sciences › Sociology and Political Science › Digital Games and Media
- Keywords
- game testing, video games, player data, convolutional neural networks, game content creation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.