Skip to main content
DA / EN

Esports analytics just got a major boost with the Time to Die 2 framework, predicting player deaths in Dota 2.

The world of esports is a rapidly growing industry, attracting millions of viewers from all over the world. The popular game Dota 2 is no exception, with its complex and fast-paced gameplay keeping viewers on the edge of their seats. However, broadcasting such a game comes with its own set of challenges, especially when it comes to camera positioning and creating engaging commentary.

In Dota 2, death events are a crucial aspect of the game that can lead to a dramatic shift in the balance of power. Unfortunately, with just one camera covering the entire playfield, it's easy for these events to go unnoticed during broadcasts. That's where the Time to Die 2 framework comes in - a highly advanced machine-learning model that predicts when a player is likely to die within five seconds.

Being able to predict a player's death within a game of Dota 2 is a game-changer in the field of esports analytics. This technology can inform commentators when a player is in danger, improving their ability to keep viewers engaged and informed. The machine learning model can also help coaches identify dangerous moments and train their players to avoid them.

However, predicting death events in Dota 2 is no easy feat. With over 120 heroes and more than 200 items to choose from, along with each player having a unique role to play, the game is highly complex. Plus, the "fog of war" mechanic means that players have limited visibility of the battlefield, further complicating the prediction process.

To create this highly accurate model, a team of data scientists spent three years analysing data from nearly 10,000 Dota 2 matches. Thanks to the open nature of the replay system, they were able to curate massive datasets of high-quality, high-volume data that allowed them to break down the complexity of Dota 2's data space. The result is an incredibly complex deep learning model that can accurately predict death events, providing enough time for broadcasters and commentators to alert viewers and create a more engaging viewing experience.

In short, the Time to Die 2 framework is a revolutionary development in the world of Dota 2 and esports broadcasting. By predicting death events with a high degree of accuracy, this technology helps commentators stay on top of the action and keeps viewers engaged. With further development, the machine learning framework will also become a powerful analysis tool for players and teams.

Full research report: https://doi.org/10.1016/j.mlwa.2023.100466

Whitepaper "Predicting Death in Dota 2"


Editing was completed: 26.05.2023