Data-driven gameplay experience evaluation: Leveraging multimodal data in Games User Research

近年来，数字游戏的开发和研究发生了许多变化，包括新的交互模式，扩大玩家人口统计和新的商业模式。这些都带来了机遇，但也带来了更多的不确定性。因此，在过去的几年里，对用户及其与数字游戏互动的研究需求不断增长。游戏用户研究（GUR）是一个相对较新的领域，使用HCI和心理学的评估方法，旨在通过为研究人员和设计师提供有关游戏玩法的数据驱动信息来改进游戏设计。虽然标准的人机交互评估方法在理解生产力应用程序的可用性方面取得了进展，但视频游戏的特定特征，例如只要是游戏设计的一部分，就允许挫折，这意味着许多既定的用户研究方法不能以同样的方式应用于GUR。此外，游戏用户研究人员通常要处理大量复杂的多模态数据集。因此，挑战之一是提高数据收集方法的效率，并在研究和设计决策中对这些数据进行有意义的解释。因此，优化适合GUR的评价方法已成为游戏和HCI研究人员的关键研究课题之一（如CHI 2019“游戏中的用户体验研究”课程，CHI 2016“独立和非营利组织的GUR”研讨会以及CHI 2020 Games & Play特别兴趣小组（SIG会议）等各种GUR相关活动所指出的那样：“塑造游戏和HCI研究的下一个十年”）。 我的研究计划侧重于这些关键的研究课题，通过开发工具和框架，使用系统和科学的方法分析玩家体验。该提案解决了应用GUR设置中的一个重要问题，即通过利用多模式游戏数据（例如玩家通过采访的评论、他们在游戏中的动作和通过遥感勘测的运动）来改善对玩家体验的评估。这项研究的主要成果将包括玩家体验评估的新方法，以及根据实际案例研究和玩家体验科学实验的结果产生的最佳实践和游戏设计指南。 这里提出的工作分为两个研究要素：首先，通过探索当前的方法，我计划开发混合方法，提高定性和定量数据收集，分析和可视化的有效性和效率。第二，应用数据驱动的设计和决策来推进混合现实（增强现实和虚拟现实）和流媒体（例如电子竞技）内容的游戏开发框架和设计指南。因此，这项研究的成果将包括适用的工具和方法，以及优化游戏行业人机交互（HCI）方法的最佳实践。

In recent years, there have been many changes in digital games development and research, including new interaction modes, widening player demographics and new business models. These present opportunities, but also additional uncertainties. As a result, the demand for research studies dealing with users and their interactions with digital games has grown in the past few years. Games user research (GUR) is a relatively new field using evaluation methods from HCI and psychology aiming to improve game design by providing data-driven information about gameplay for researchers and designers. Although standard HCI evaluation methods have made progress in understanding the usability of productivity applications, the specific characteristics of video games, such as frustration being allowed as long as it is part of the game design, mean that many established methods of user research cannot be applied in the same way for GUR. Moreover, games user researchers are usually dealing with massive and complex multimodal datasets. Thus, one of the challenges is in improving the efficiency of data gathering approaches and in interpreting these data meaningfully in research and in making design decisions. For these reasons, optimisation of evaluation methods suitable for GUR has become one of the key research topics for games and HCI researchers (as noted in various GUR related events such as the CHI 2019 Course on "UX Research in Games", the CHI 2016 Workshop on "GUR for Indie and non-profit organisations", and CHI 2020 Games & Play Special Interest Group (SIG meeting): "Shaping the Next Decade of Games and HCI Research"). My research program focuses on these key research topics by developing tools and frameworks for analysing player experience using systematic and scientific approaches. This proposal tackles an important problem in applied GUR setting, which is to improve evaluation of player experience by leveraging multimodal gameplay data (such players comments via interview, their in-game actions and movement via telemetry). Major outcomes of this research will include novel approaches for player experience evaluation as well as best practices and game design guidelines generated based on results from practical case studies and scientific experiments of player experience. The work proposed here is divided into two research elements: First, by exploring current approaches I plan to develop mixed methods that improve the effectiveness and efficiency of qualitative and quantitative data collection, analysis and visualization. Second, applying data-driven design and decision making to advance game development frameworks and design guidelines in mixed-reality (augmented and virtual reality) and streaming (e.g. eSport) content. Hence the outcome of this research will include applicable tools and methods as well as best practices on optimisation of human-computer interaction (HCI) methods for the game industry.