Streamlining Esports Analysis: How GPT Transforms Data Interaction with Natural Language

Vincent Engel

5 min read

·

Aug 2, 2023

Artificial intelligence (AI) has the potential to revolutionize a multitude of industries. This article aims to start a dialogue on leveraging large language models (LLMs) to understand esports data better. This is just the beginning of a new frontier for data interaction and analysis. Join us as we delve into the fascinating world of AI-enhanced esports analysis.

I’ve been using Siri on my phone for years. What is all this fuss about?

Indeed, AI is not a novelty. Its roots stretch back many years, with AI-powered characters challenging us in video games like Mario Kart and sophisticated algorithms powering Google’s image classification capabilities. This can already be used to accomplish things beyond human ability. GRID has also used traditional forms of machine learning / AI to predict the outcomes of esports series successfully. We’re exploring a new frontier of AI — the generative capabilities offered by Large Language Models from industry leaders like OpenAI and Meta. These are not just any AI; they’re game-changers.

The Power of Large Language Models

Your first interaction with an LLM probably looked something like this:

User: Hello

AI: Hello, how can I assist you today?

User: What are the three best CSGO esports teams?

AI: The three best CSGO teams are:

1. Astralis: Known for its dominant reign in the Counter-Strike scene from 2018 to 2019, Astralis is considered one of the greatest teams in the history of CSGO. They have won multiple major championships in CSGO.

2. Natus Vincere (Na’Vi): Featuring superstar player Oleksandr “s1mple” Kostyliev, Na’Vi has been a top contender in the CSGO scene for many years. They have won various premier tournaments, showing consistent high-level performance.

3. Team Vitality: A top French team that has risen to prominence recently thanks to their star player Mathieu “ZywOo” Herbaut. They have consistently placed high in multiple premier CSGO tournaments.

Simply amazing, a human-readable answer quickly and concisely. It even gave us some insights as to why it considers these teams to be among the best. But soon, you will likely see the limitations of the model.

User: What was the last tournament Astralis won?

AI: I’m sorry, but as an AI model developed by OpenAI, I was last trained on new data in October 2021, and I can’t provide real-time updates.

LLMs and GRID: An Unstoppable Duo

Imagine a world where AI doesn’t falter when asked about the latest tournament a specific CSGO team won. This is precisely what can be achieved by integrating the GRID Data Platform with LLMs. Recently, LLMs like GPT 4 have gained the ability to retrieve data and interact with external services. This means they can easily access real-time data to answer user queries. This advancement can be used to enhance the accessibility of esports data significantly.

• Need a list of all the times a team lost on Sundays? No problem.
• Want to determine how many hours a player participated in tournaments last year? Easy.
• Curious about how often the new map was used during the latest tournament and which team picked it most? Or the win percentage of the teams that picked it? Here you go.

LLMs can enhance the accessibility of esports data for everyone involved in the industry. Whether you’re a team coach seeking to improve strategies, a fan curious about their favorite player’s stats, or a game developer wanting to understand the best players’ strategies for game balancing, LLMs can be beneficial.

AI on the GRID

Let’s take a look at an example of GRID + AI in action.

As you can see, you can interact with GRID’s data as if it was another person aware of all the developments in esports and can make accurate assumptions based on that. This is a rudimentary implementation that has actually been written in collaboration with GPT 4.

This is definitely not the last time you’ll hear us talk about LLMs and AI. We’re just as stoked as you are about the future use cases this new technology opens up. Has this sparked your imagination? I would love to see what you can do with GRID’s data and AI. We’re currently running a GRID Esports DataJam in partnership with OperaGX— sign up, build and submit your project, and get a chance to tap into a prize pool of over $20,000.

Also, if you found all of this captivating and you could see yourself working on similar things showcased by the proof of concept, please look at GRID’s current open positions and consider applying. It would be great to have you as a part of the team!

Vincent Engel is a Senior Software Engineer at GRID, where he works at the core of the GRID Data Platform. Living and breathing the dynamics of Go, Rust, Kafka and AWS in his software engineering journey, he is striving not just to innovate but also to inspire. Always up for sharing some tech wisdom and jumping on a new wave of the ever-evolving tech landscape. Make sure to buckle up for some thrilling insights coming from Vincent’s posts!