New Study Finds Taking Breaks Boosts Team Performance

| June 27, 2018 

USC computer scientists leverage artificial intelligence to analyze the performance of individuals in online team-based games, producing unique insights into the psychology of performance

ISI researchers used artificial intelligence to measure individual performance and overall team performance in 242,000 matches of the “League of Legends” online game. Photo/Pexels.

Want to be a good team player? Take a break. It may improve not only your own performance but the chances of your team winning overall, says a new study by a team of USC computer scientists.

Researchers from USC Viterbi’s Information Sciences Institute (ISI) crunched data from thousands of players in a popular online video game to analyze individual performance in teams over time. They also examined the impact of expertise on performance and other factors influencing player behavior, such as continuing to play versus ending the session.

“We’re trying to understand what drives certain patterns of behavior in real social environments and online gaming allows us to study this question from different perspectives in a natural setting,” said the study’s lead author Anna Sapienza, an ISI postdoctoral research associate.

“We’re trying to understand what drives certain patterns of behavior in real social environments.” Anna Sapienza

“We’re looking not just at performance but what drives engagement in people. Finding a way to understand what drives people could be useful not just for gaming, but any type of team-based task in everyday life.”

Sapienza co-authored the paper with ISI research team leaders Emilio Ferrara, a research assistant professor in computer science, Kristina Lerman, a research associate professor in computer science, ISI graduate research assistant Yilei Zeng, and former visiting researcher Alessandro Bessi.

Maximizing performance 

For this study, the researchers analyzed online gaming data collected from 242,000 matches played by 16,665 players of League of Legends, a popular multiplayer online battle arena game, between May 2014 and January 2016.

Study co-authors ISI postdoctoral research associate Anna Sapienza and ISI research team leader Emilio Ferrara, a USC Viterbi assistant professor in computer science. Photo/Caitlin Dawson.

Specifically, they used machine learning algorithms to look for patterns in player performance over the course of a single gaming session, measuring both individual performance and overall team performance. A session is defined as a period of game play activity without an extended break.

“We’re analyzing data from thousands of people and tens of thousands of observations, so the complexity grows exponentially: we have to handle a huge amount of raw data,” said Ferrara.

“This is challenging, but it also gives us a lot of data to scrutinize and allows us to look at the variations between expert versus novice players. Compared to a lab study or survey, this gives us a much larger world of options to look at to try to understand the link between motivation and performance.”

They found that players who took regular breaks tended to do better than those who did not—and so did their teams. Player performance decreased on average 8 to 10 per cent between the beginning and end of a game session. Further, they observed that the team to which a player is assigned saw lower win rates on average if that individual had already played other matches without a break.

Put your mind at rest

This performance deterioration pattern could point to a phenomenon called cognitive depletion, which occurs when we work at taxing, attention-demanding tasks continuously for long periods of time. Extended attention to a single task actually hinders performance—like muscles during physical exercise, after prolonged flexing the mind needs to take a rest.

“As you get tired, this cognitive mechanism is triggered and output quality is diminished,” said Ferrara. “In fact, after a long gaming session, your performance is actually worse than that of a less skilled but fresh player.”

So, does this fly in the face of the old adage that practice makes perfect? Not necessarily. The researchers found that in the long term, more experienced players were less likely to suffer from cognitive depletion.

“The effect of cognitive depletion is much more drastic for novices,” said Ferrara. “In the long run, it seems you build up a kind of mental stamina, so you’re not as effected by this cognitive depletion.”

“The effect of cognitive depletion is much more drastic for novices.” Emilio Ferrara

The researchers also used machine learning to accurately predict whether a player would choose to continue to play or end the session. They found players tended to quit the game session after a certain number of matches in which their performance had declined.

In addition to providing insights into individual performance in teams, the analysis could improve understanding of gaming addiction disorder (recently classified as a disease by the World Health Organization) and the cognitive impact of extended gameplay.

Greater than the sum of its parts

While similar short-term performance deterioration has been observed in the context of different online activities, such as commenting on Reddit or Twitter, this is the first time the depletion effect was observed in the context of team work and, in particular, online games.

“Studies tend to focus on the overall quality of teams, rather than scrutinizing the contributions of individuals,” said Ferrara. “They say the team as a whole is greater than the sum of its parts, but that’s not necessarily always the case; our research shows that an individual’s behavior can make or break a team.”

The study, titled “Individual Performance in Team-Based Online Games,” was published June 20 in Royal Society Open Science.

 

Published on June 27th, 2018

Last updated on May 16th, 2024

Share this Story