New research shows that pairing players strategically rather than simply matching equal opponents boosts engagement, reduces reliance on bots and challenges one of gaming’s most entrenched design principles.
BALTIMORE, June 1, 2026 — For years, the video game industry has operated on a simple assumption: the fairest match is the best match.
New research suggests that assumption may be costing gaming platforms millions of player-hours.
A study published in Management Science finds that the industry’s standard approach of matching players against opponents with similar skill levels is not necessarily the most effective way to keep people engaged. Instead, researchers show that more sophisticated matchmaking systems—ones that account for how players respond to recent wins, losses and competitive experiences over time—can significantly increase player retention.
The findings arrive as game publishers face mounting pressure to keep players engaged in an increasingly crowded market. With the global gaming industry projected to generate nearly $188 billion annually, even small improvements in retention can translate into substantial business impact.
“Skill-based matchmaking is intuitive and widely used because it feels fair,” said Xiao Lei, professor at the University of Hong Kong and co-author of the study. “But fairness and engagement are not always the same thing. Our research shows that carefully designed matchmaking policies can meaningfully extend player participation without sacrificing overall balance.”
The insight challenges one of the foundational ideas behind modern competitive gaming.
Traditional skill-based matchmaking, often called SBMM, is designed to create evenly balanced contests by pairing players of similar ability. While that approach promotes fairness, it also guarantees that many players lose as often as they win. Over time, repeated losses—or even short losing streaks—can drive frustration and push players away.
The researchers argue that matchmaking should not be viewed as a series of isolated decisions. Instead, it should be treated as a dynamic system in which every match influences what players do next.
To test that idea, the team developed a model that incorporates both player skill and the psychological effects of recent outcomes. The system then optimizes matchups over time with the goal of maximizing long-term engagement across the entire player ecosystem.
The gains were substantial.
In a large-scale analysis of 5.4 million matches from the online chess platform Lichess, optimized matchmaking increased engagement by 4% to 6% compared with conventional skill-based approaches. Under certain theoretical conditions, the model produced engagement gains of up to 50%.
“These are economically meaningful improvements at scale,” said Mingliu Chen, professor at the University of Texas at Dallas and co-author of the study. “For platforms with millions of active users, even a few percentage points of improved engagement can translate into substantial long-term value.”
The research also offers a fresh perspective on two of gaming’s most controversial practices: pay-to-win systems and AI-powered bots.
Pay-to-win mechanics are often criticized for giving paying players an unfair advantage. Yet the study finds that under specific conditions, these systems can increase overall engagement by altering the distribution of player skill in ways that stabilize the competitive environment.
The findings do not suggest that pay-to-win systems are universally beneficial. Rather, they reveal how deeply player behavior, incentives and matchmaking are interconnected—and how platform designers may be overlooking those relationships.
The study reached a similarly surprising conclusion about bots.
Many gaming platforms deploy AI-controlled opponents to influence player experiences and help maintain engagement. But researchers found that smarter matchmaking can achieve comparable or better results while relying on fewer bots.
“Engagement isn’t just about giving players easier wins or adding artificial opponents,” said Adam N. Elmachtoub, professor at Columbia University and co-author of the study. “It’s about dynamically managing the ecosystem of players over time.”
The implications extend well beyond gaming.
As digital platforms increasingly compete for attention, retention has become a defining challenge across industries—from online learning and gig-economy marketplaces to social platforms and other competitive digital environments. The study introduces a broader framework for understanding how repeated interactions and human memory shape long-term participation.
In other words, the future of engagement may depend less on optimizing individual experiences and more on managing entire ecosystems intelligently.
For game developers, that could mean rethinking one of the industry's most accepted rules. For digital platforms more broadly, it points to a larger lesson: keeping people engaged is not simply about creating better moments. It is about designing better sequences of moments—and the difference can be worth millions.
About INFORMS
INFORMS is the world's largest association for professionals and students in operations research (O.R.), AI, analytics, data science and related disciplines. Through its journals, conferences and professional resources, INFORMS advances data-driven decision-making and scientific innovation across industries and society.
About Management Science
Management Science is a leading peer-reviewed journal published by INFORMS that features rigorous research on management, strategy, operations, analytics and decision-making across organizations.
###
Media Contact
Rebecca Seel
Public Affairs Specialist, INFORMS
(443) 757-3578
Media Contact
Jeff Cohen
Chief Strategy Officer
INFORMS
Catonsville, MD
[email protected]
443-757-3565