A short time ago, a number of companies teamed together to launch the CS:GO-based “AI Hackathon”. The challenge for those that entered was to develop a real-time piece of software that could be used to predict the outcome of CS:GO matches as accurately as possible.
There’s nothing the coding community likes more than being handed a challenge on a plate and as expected, interest in the AI Hackathon was huge with 373 entries into the contest from 20 different countries around the world.
Three of those entries were selected by a trio of big companies involved in CS:GO esports, BLAST, Skybox Technologies, and Copenhagen Capacity to be showcased at the recent BLAST Premier Fall Finals event. After all, to test the veracity of a prediction tool, you would need to see it in action in a real CS:GO event to judge how effective it is.
How did the Algorithm Perform?
Any prediction tool lives and dies by its accuracy and the good news for the developers is that despite hitting 73% accuracy on average initially in the tournament, as the tournament progressed and more games were played (and more data collated by the program), its accuracy increased throughout the tournament.
That is a pretty good hit-rate for any application and it was not surprising to learn that Skybox Technologies quickly announced that they would roll out the AI Hackathon tool to the CS:GO community in 2021.
The company states that the release of the algorithm will “help teams, players, and coaches in training and preparation ahead of tournaments.”
“Vast Applications Within and Outside of Esports”
Speaking about the imminent release of the utility, the CEO of Skybox Technologies Ethan Cooper commented:
“We’re really happy with the engagement of the AI teams and the possibility to provide a platform to showcase the talent within the fanbase.”
“Such skills have vast applications within and outside of esports. The opportunity of working with some of the participants further is an added bonus’ and along with this, we are really excited about making the technology available to the broader CS:GO community next year.”
The three teams that landed the top three prizes in the AI Hackathon were:
- Data Pirates (India)
- Stacc Attacc (United States)
- Jin (Slovakia)
Skybox will now work with these programming teams to hone the accuracy of the CS:GO prediction tool still further ahead of its release to the wider community in 2021.
Are there other applications the CS:GO prediction utility could have?
This is a very exciting development for the CS:GO community as the application could have further uses within both esports and esports betting. For example, if the utility suggests that in one game a team has a 90% chance of winning, but that team is then defeated heavily by a much inferior team, could the app be used to ‘diagnose’ when match-fixing is taking place?
Similarly, esports betting companies could use the utility to give them a more accurate prediction of how the match will turn out and as such, they will be able to offer even more realistically priced markets on those games.
Teams can use the data of their team and opponents and feed it into the utility to try different strategies and team selections to see what offers them a better chance of success. It is certainly going to be interesting to see how its use affects the performance of the top-level teams across the world over 2021 and beyond.