Esports research ideas: Tactical insights – Esports
In her ongoing exploration of esports research, Dr. Darina Goldin from Bayes Esports presents an innovative idea that could revolutionize the esports ecosystem. She emphasizes the potential to leverage the vast data available in numerous games for detailed tactical analysis—an opportunity largely unexplored in traditional sports.
A decade ago, my competitive journey in Team Fortress 2 as a member of a top-20 European team showcased this game’s immense potential. Renowned as a premier first-person shooter, TF2 remains a high-stakes favorite among players.
During that era, competitive gameplay primarily revolved around 6v6 matches, emphasizing map control. Gaining early access to central locations often decided the outcome of rounds, highlighting the significance of precise timing and strategic positioning.
As a cohesive unit, our practices focused on meticulously crafted “rollouts” for each map—each team member trained to understand their specific path and role in reaching the center first. We developed actionable strategies tailored to our opponents, meticulously documented in my personal playbook, a cherished reminder of those strategic formations.
Transformative Strategies
In team sports, success hinges on robust strategies, a principle that equally applies to esports. At elite levels, skills such as mechanics and aiming tend to be comparable across competitors; thus, intangible elements—decision-making, synergy, and tactical acumen—prove decisive. Selecting points of attack in Counter-Strike or determining the optimal moment to engage neutral monsters in League of Legends can significantly influence match trajectories. Likewise, effective response strategies following player losses can avert disastrous outcomes. The pivotal role of opening tactics, particularly champion selection, is familiar to many in the industry.
The endeavor to quantify and assess these tactics isn’t novel. Professional teams increasingly engage dedicated analysts, while innovative companies like Shadow offer support services. However, the industry is still in its infancy; the integration of machine learning holds potential to deepen our comprehension of gameplay dynamics.
Pioneering Analysis
Esports are poised to be at the forefront of AI-driven analytical advancements. The abundant high-quality data available allows for insights far surpassing traditional commentary and analysis methods.
Consider soccer: generating automated insights necessitates transforming match data into machine-readable formats, a labor-intensive process involving meticulous record-keeping of ball possession, passes, shots, and player movements. This sparse data often requires extensive manual labor to enhance. Even with player tracking technology, assembling a comprehensive dataset for analysis demands significant resources.
In contrast, esports inherently exist as rich, data-driven environments. With the right data provider, every action during a match can be meticulously tracked—player movements, actions taken, and even button presses are all recorded. At Bayes Esports, our access to official publisher data enables us to curate expansive datasets. Should any crucial information be absent, we can swiftly amend it through code adjustments, re-executing our data queries with ease.
New Horizons Ahead
Pioneering work has already initiated this journey. In 2021, Bayes Esports data scientist Gustav Geißler successfully employed machine learning to categorize various Counter-Strike strategies, producing comprehensive statistics that unveiled prevalent trends worthy of further inquiry.
We envision establishing a systematic catalog of CS:GO openings, akin to chess moves, evaluating their effectiveness while accounting for various situational factors. This analysis may extend to exploring the relationship between strategic choices and in-game economic conditions or team member dependencies. The scope of this research is virtually limitless.
Similar methodologies can be applied to Multiplayer Online Battle Arena (MOBA) games, where focus shifts from initial movement to goals accomplished and items purchased. Understanding tactical abstractions facilitates investigation into team dynamics, revealing shifts in strategies across different lineups or following significant game patches.
Yet, this is merely the initial phase—the scholarly exploration of these concepts remains nascent. Where are the aspiring researchers eager to delve into esports analytics? We must cultivate a thriving academic discourse around esports tactics in the upcoming years.
Collaborations between machine learning specialists and sports scientists are emerging, yet numerous research inquiries remain unanswered. Many aspects of esports analysis resonate beyond gaming; these extensive datasets can inform studies on team dynamics and collaboration, offering insights applicable to traditional sports as well.
