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].”
Method
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Human-Like Playtesting with Deep Learning. https://app.cochat.ai/papers/10.1109/cig.2018.8490442
- The paper
- Gudmundsson, S. F., Eisen, P., Poromaa, E., Nodet, A., Purmonen, S., Kozakowski, B., Meurling, R., & Cao, L. (2018). Human-Like Playtesting with Deep Learning. 2018 IEEE Conference on Computational Intelligence and Games (CIG). https://doi.org/10.1109/cig.2018.8490442
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.