The way we play computer games is so predictable

We're tracking the way people play online games - to help shape the games of the future


Our behaviour models are based on data from the two billion people worldwide who play computer games

Our researchers have developed behavioural models which cast new light on the different ways we play computer and mobile phone games.

The models have been developed over several years, and focus on breaking down how we play into sets of playstyles. The details of these playstyles can be fed into prediction models that tell us how long someone is likely to play a game, as well as how they will play it. They also identify the issues that might prevent a player from enjoying a game.

Prediction analysis

The research is already at work in the games industry where behavioural models and prediction analysis are used to improve or modify games to make them more suited to different player motivations and interests.

The work is led by Anders Drachen, Professor in Computer Science and Digital Games in our Digital Creativity Labs, and involves several international universities and companies.

Professor Drachen explained: “We have a tendency to consider ourselves unique and unpredictable, but in computer games, new research shows that this is far from the case. In fact, we can be categorised into groups of people showing the same behaviour, and what we do in the game in the future is predictable.

“For example, how you play a game will reveal how long you are going to stay interested for. This means that games can now begin to change to provide you with the best possible experience, catching any problems well ahead of time and attempting to eliminate them.”

Data models

The teams built data models based on games such as Destiny and Just Cause – vast, complex open-world games – as well as thousands of other games across PCs, consoles and mobile platforms.

In addition to the models used to define playstyles and predict behaviour, the research teams also identified some basic patterns in how we play games, for example:

  • Most PC and console games are played from two to thirty hours. The vast majority are played for less than eight hours, and very few games see more than 35 hours of playtime on average. This means that a lot of game content is never experienced
  • Only a portion of people who start a game complete it – this is especially the case for mobile games where most people have typically left after less than two minutes
  • Social connections formed in and around games are powerful motivators to keep playing. If a player’s friends play the same game, or you have in-game friends, the higher the rates of game retention. This is one of the reasons why multi-player games are so popular.

The researchers based their studies on data gathered from samples of the two billion people worldwide who play computer games – a number that is increasing rapidly.

Behaviour patterns

“There are distinct patterns in our behaviour, and this means that we can figure out your play style and what you are interested in from the game,” says Professor Drachen.

“It also means we can look at how you play a game to begin with, and then train algorithms to predict for how long you will keep playing before losing interest. We can also move beyond that, and predict for example whether you will purchase something in a game, or even what it is in the game that keeps you interested, and thus adapt the game to give you more of what you like.”

He added: “This research is extremely useful for the games industry because the algorithms allow companies to predict the behaviour of players, and take steps to improve games accordingly.

“Identifying user experience problems is also incredibly useful for the industry, especially for games played over long time periods.”

The text of this article is licensed under a Creative Commons Licence. You're free to republish it, as long as you link back to this page and credit us.

This research is extremely useful for the games industry because the algorithms allow companies to predict the behaviour of players”

Professor Anders Drachen
Featured researcher
Professor Anders Drachen

Professor Anders Drachen

Research interests in the analysis of user behaviour in – and around – games.

View profile

Discover the details

Explore the work of our Digital Creativity Labs

Rapid Prediction of Player Retention in Free-to-Play Mobile Games, was published in Proceedings of the Artificial Intelligence in Interactive Digital Entertainment Conference - available here

How playstyles evolve: Progression Analysis and Profiling in Just Cause 2, was published in Proceedings of the International Conference on Entertainment Computing

Predicting Player Churn in Destiny: A Hidden Markov Models Approach to Predicting Player Departure in a Major Online Game, was published in Proceedings of the Computational Intelligence in Games Conference