Predicting Player Performance in Tactical Games with the Lightgbm Algorithm

Image-based rendering and reconstruction (IBR) approaches minimize time and costs to develop video-game assets, aiming to assist small game studios and indie game developers survive in the competitive video-game industry. To further investigate the interplay of IBR on developers’ efficiency, game performance, and players’ gaming experience we conducted two evaluation studies: a comparative, ecologically valid study with professional game developers who created games with and without an IBR-based game development pipeline, and a user study, based on eye-tracking and A/B testing, with gamers who played the developed games. The analysis of the results indicates that IBR tools provide a credible solution for creating low cost video game assets in short time, sacrificing game performance though. From a player’s perspective, we note that the IBR approach influenced players’ preference and gaming experience within contexts of varying levels of player’s visual intersections related to the IBR-created game assets.

Enhancing player engagement through game balancing in digitally augmented physical games

Boland, M, Delude, K, and Miele, EM. Relationship between physiological off-ice testing, on-ice skating, and game performance in Division I female ice hockey players. J Strength Cond Res 33(6): 1620-1629, 2019-The purpose of this study was to identify off-ice testing variables that correlate with skating and game performance in Division I collegiate female ice hockey players. Twenty female forward and defensive players (19.95 ± 1.35 years) were assessed for mass, height, percent fat mass (%FAT), bone mineral density, predicted 1 repetition maximum (RM) absolute and relative (REL%) bench press (BP) and hex bar deadlift (HDL), lower-body explosive power, anaerobic power, countermovement vertical jump, maximum inspiratory pressure (MIP), and on-ice repeated-skate sprint (RSS) performance. The on-ice RSS test included 6 timed 85.6-m sprints with participants wearing full hockey equipment; fastest time (FT), average time (AT), and fatigue index (FI) for the first length skate (FLS; 10 m) and total length skate (TLS; 85.6 m) were used for analysis. Game performance was evaluated with game statistics: goals, assists, points, plus-minus, and shots on goal (SOG). Correlation coefficients were used to determine relationships. Percent fat mass was positively correlated (p ≤ 0.05) with FLS-FI and TLS-AT; TLS-FT was negatively correlated with REL%HDL; BP-RM was negatively correlated with FLS-FT and FLS-AT; MIP positively correlated with assists, points, and SOG; FLS-AT negatively correlated with assists. Game performance in female ice hockey players may be enhanced by greater MIP, repeat acceleration ability, and mode-specific training. Faster skating times were associated with lower %FAT. Skating performance in female ice hockey players may be enhanced by improving body composition, anaerobic power, and both lower- and upper-body strength in off-ice training.

Association between post-game recovery protocols, physical and perceived recovery, and performance in elite Australian Football League players

This study examined the effects of individual characteristics and contextual factors on training load, pre-game recovery and game performance in adult male semi-professional basketball. Fourteen players were monitored, across a whole competitive season, with the session-RPE method to calculate weekly training load, and the Total Quality Recovery Scale to obtain pre-game recovery scores. Additionally, game-related statistics were gathered during official games to calculate the Performance Index Rating (PIR). Individual characteristics and contextual factors were grouped using k-means cluster analyses. Separate mixed linear models for repeated measures were performed to evaluate the single and combined (interaction) effects of individual characteristics (playing experience; playing position; playing time) and contextual factors (season phase; recovery cycle; previous game outcome; previous and upcoming opponent level) on weekly training load, pre-game recovery and PIR. Weekly load was higher in guards and medium minute-per-game (MPG) players, and lower for medium-experienced players, before facing high-level opponents, during later season phases and short recovery cycles (all p < 0.05). Pre-game recovery was lower in centers and high-experience players (p < 0.05). Game performance was better in high-MPG players (p < 0.05) and when facing low and medium-level opponents (p < 0.001). Interestingly, players performed better in games when the previous week’s training load was low (p = 0.042). This study suggests that several individual characteristics and contextual factors need to be considered when monitoring training load (playing experience, playing position, playing time, recovery cycle, upcoming opponent level), recovery (playing experience, playing position) and game performance (opponent level, weekly training load, pre-game recovery) in basketball players during the competitive season.

Multiplayer Dynamic Difficulty Adjustment (mDDA) is a method of reducing the difference in player performance and subsequent challenge in competitive multiplayer video games. As a balance of between player skill and challenge experienced is necessary for optimal player experience, this experimental study investigates the effects of mDDA and awareness of its presence on player performance and experience using subjective and biometric measures. Early analysis indicates that mDDA normalizes performance and challenge as expected, but awareness of its presence can reduce its effectiveness.

Higher resolution is one of the main directions and drivers in the development of virtual reality (VR) head-mounted displays (HMDs). However, given its associated higher cost, it is important to determine the benefits of having higher resolution on user experience. For non-VR games, higher resolution is often thought to lead to a better experience, but it is unexplored in VR games. This research aims to investigate the resolution tradeoff in gameplay experience, performance, and simulator sickness (SS) for VR games, particularly first-person shooter (FPS) games. To this end, we designed an experiment to collect gameplay experience, SS, and player performance data with a popular VR FPS game, Half-Life: Alyx. Our results indicate that 2K resolution is an important threshold for an enhanced gameplay experience without affecting performance and increasing SS levels. Moreover, the resolution from 1K to 4K has no significant difference in player performance. Our results can inform game developers and players in determining the type of HMD they want to use to balance the tradeoff between costs and benefits and achieve a more optimal experience.

The objectives are to explore the effect of a random forest algorithm on the state prediction and fault classification of smart meters, so that the smart meters can run more stably. Based on the principle of the random forest algorithm and Light Gradient Boosting Machine (LightGBM) algorithm, its theoretical basis and application are deeply analyzed and improved. An improved fault classification and state prediction model of smart meters is designed based on a random forest‐improved LightGBM algorithm. The built model algorithm is evaluated by utilizing public data sets. The results show that, by preprocessing the fault data set of smart meters, 8 fault feature types including fault type, working time, and fault month are obtained. When the improved LightGBM algorithm is trained based on random forest, the average accuracy of the algorithm is 67.65%, the average recall rate is 64.11%, and the average F1 value is 65.73%. Meanwhile, the difference between the algorithm and the random forest algorithm and the Correlation‐based Feature Selection (CFS) algorithm is studied. Therefore, the prediction accuracy and fault classification of the constructed model algorithm for smart meters are higher than those of the other two algorithms. It indicates that the algorithm has a good application effect and high practical application value and can provide a scientific and useful reference for the follow‐up research of smart meters.

Cricket is one of the team games played over 50 countries in different levels. Though the performance of each batsman in the team can be easily quantified, the prediction of player performance is arduous. This paper demonstrates a methodology to predict the performance of cricket batsman in test-match series. In this study, longitudinal test cricket data have been collected over five years of period. A model is developed to predict the player performance as a function of certain characteristics related to the player, the team and the match series. Due to the hierarchical nature of the collected cricket data, a three stage hierarchical linear model is proposed in this investigation. According to the outcome of the analysis, the handedness of the player (batsman) and the rank of the team significantly influence player performance. Finally, an accurate prediction of player performance is conducted using the proposed model.

With the emerging demand for video games as entertainment, it is important to examine human behavior when designing and developing games. Past research has shown support that gender stereotypes imposed on female gamers are prevalent within this space. Another study showed that when female players perceive being in an all male performative environment, they are prone to stereotype threat which hinders their own performance. To continue the efforts of exploring the detrimental effects of this problem, a pilot study was devised using a 2x2 factorial within subject design with team compositions (male vs female) and atmosphere (negative vs positive interactions) as the independent variables. Ten female participants played a first-person-shooter game called Valorant by Riot Games across two consecutive days where they would play in both team compositions for gender. At the end of each game, participants took a questionnaire to self-rate their game experience, anxiety levels and desirability to communicate during gameplay. Additionally, researchers observed game performance, frequency of in-game communication and confederate engagement. Researchers hypothesized that participants who played in the all high ranking male team with negative interactions would communicate less frequently and have lower desirability to communicate, have worse game performance, less engagement with the confederates, higher levels of anxiety and overall worse game experience than other conditions. Statistical analyses revealed the atmosphere factor (positive vs negative) significantly influenced performance outcomes (e.g., kills and ranking) and subjective experiences (e.g., game experience, team interactions, feelings of respect, safety, disrespected, discouraged, and supported). However, there was no significant main effect for team composition and no significant interaction effect between team composition and atmosphere. Continued research of this problem space is crucial to ensure video games are inclusive and enjoyable for all.

Can video game players embody the game world without embodying a concrete, bodily-shaped avatar? By exploring the players’ embodied experiences in real-time strategy games where no concrete avatar is present, we argue that the answer is yes. Specifically, we argue that the players can embody a game world consisting of various controlled game items by incorporating game mechanics with their bodily actions. Such an embodiment enables skilled players to develop a nuanced perception of possible actions within the game and to react more effectively than novices in a given situation. Focusing on Age of Empires 2, we analyze three key skills —micro-management, quick-walling, and map reading— that differentiate experts’ game performance from novices’, showing that the difference in their performance rests on the difference in their embodiment of the game rather than the difference in their reflective knowledge. Ultimately, the findings suggest that players’ capacity to perceive and execute actions is fundamentally shaped by their embodied experiences within the game’s context.

Digital game experience is not a one-dimensional concept. Great variety exists in game genres and players, and game experiences will differ accordingly. To date, game experience is studied in a differentiated way, meaning that most studies focus on one specifi c game experience dimension. The objective of our study was twofold. First, we wanted to obtain a comprehensive picture of fi rst-hand experiences of playing digital games. We conducted six focus group interviews including different types of gamers with the aim of eliciting a wide array of lay-conceptualizations of game experience. Second, we aimed to develop a categorization of game experience dimensions. This was established by discussing and integrating theoretical and empirical fi ndings. Our categorization revealed nine dimensions: enjoyment, fl ow, imaginative immersion, sensory immersion, suspense, competence, tension, control and social presence. This categorization has relevance for both game scholars and game developers wanting to get to the heart of digital game experience.

Article impact statement: Online games can improve the collection of data on human decision-making in situations relevant to conservation and stakeholder trade-offs. This article is protected by copyright. All rights reserved.

The online gaming industry increasingly incorporates virtual agents to enhance player experiences. Although prior literature has explored the provision of virtual agents in gaming, research on technological advancements remains limited. In this study, we investigate how introducing artificial intelligence (AI) powered agents as virtual opponents (versus rule-based opponents) influences human players’ engagement and performance. Leveraging a large-scale quasi-field experiment in a multiplayer online racing game, we employ difference-in-differences analyses with matching strategies and show that, on average, the introduction of AI opponents can have ‘discouragement effects’ on players, resulting in reduced player engagement and decreased performance. Our research also identifies differential long-term impacts of AI opponents on engagement and performance. In addition, our mechanism exploration reveals that introducing AI opponents increases competition intensity and immersion in the game, and these two factors exhibit opposing influences on players’ subsequent behavior. Specifically, heightened competition hinders players’ further engagement and performance progression, whereas a more immersive experience encourages more gaming participation and better performance. Further, we find the effects of AI opponents on player engagement and performance vary by player motivation and skill levels, such that competition-oriented and highly skilled players are more receptive to AI opponents. Moreover, our findings indicate that optimal engagement and performance outcomes occur when players compete against opponents with comparable and low competence levels, respectively. Lastly, we observe an inverted U-shaped relationship between the proportion of AI opponents and players’ engagement and performance, wherein both insufficient and excessive exposure to AI opponents lead to diminishing outcomes. Our study contributes to the literature on human-AI interactions by offering novel empirical evidence on the impact of AI opponents on human players’ experiential and instrumental outcomes and disentangling the underlying mechanisms. This work also offers practical implications for game designers and policymakers regarding the design of AI-integrated competitive environments.

The purpose of this study was to investigate flow experience and situational interest in a math learning game that included adaptive scaffolding. Fifty-two Finnish 5th graders played the game about fractions at home during COVID-19 enforced distance learning. The results showed that flow experience correlated positively with situational interest. Importantly, a deeper analysis of the Flow Short Scale (FSS) subscales revealed that only absorption by activity but not fluency of performance explained variance in situational interest. That is, at least in game-based adaptive learning, situational interest is mostly related to immersive aspects of flow. Results also revealed that students with better in-game performance had higher flow experiences, but their levels of prior knowledge were not related to flow levels. In contrast, students with lower prior fraction number knowledge showed higher situational interest, which might be partly attributed to the additional game elements provided to struggling students in the form of adaptive scaffolds. Moreover, the study demonstrated that the developed adaptive scaffolding approach and in-game self-reporting measures worked well. Finally, the implications of these findings for flow experience and situational interest research in game-based learning context are discussed.