This startup by BITS Pilani alumnus is betting on AI and analytics to predict match outcomes
Startup Predict22's founder Shubham Bindlish has created algorithms to provide sports fans with interesting insights and predictions on matches.
Today, online gaming has become one of the most profitable entertainment industries across the world. The rise of mobile technology has also made it easier and more convenient for sports fans to consume their favourite sport anytime, anywhere. And the love for sports has led to the rise of many online fantasy leagues in the country.
To help people test their betting strategies in sports, 28-year-old Shubham Bindlish founded Predict22 in 2016 in Bengaluru.
The startup makes use of artificial intelligence (AI) and analytics to provide sports fans with interesting insights and predictions.
“I have always been fascinated with algorithms and how they can be deployed when there is data available. Sport has so much data that it is never available to the public to crunch, and is owned by broadcasters, betting companies, and strategy game companies,” says Shubham.
The startup claims its almost 100 percent accurate with its predictions.
Watch Shubham in a conversation with YourStory here:
How it started?
A computer science graduate from BITS Pilani, Shubham joined Goldman Sachs as a developer. But he was interested in sports and how sports data was being used. After three years of work experience, he put his savings of Rs 15 lakh to start the company in 2016.
Initially, Shubham started exploring his models in cricket, and their first bet was the 2016 T20 World Cup, followed by the IPL and Champions Trophy. Later, he tried the model with other sports, namely basketball and football. The algorithms, claims the founder, have been accurate 95 percent of the time.
Shubham also spent more time training his algorithm to understand player consistency, performance over time, and relationship while playing against certain bowlers and teams.
“I wanted to equip people with the ability to make good decisions and not just go blindly put their money," says the founder.
After two years of extensive training of the algorithm, he began testing it in a closed beta. He says, his algorithm began to beat every NBA prediction, which gave Shubham the confidence to open up the website for individuals to come and test their strategies.
How it works?
The website has the schedule of all games. Individuals can go and register on the website and then start selecting their favourite players. They can predict game results based on the team they pick.
Since algorithms are the key here, Shubham uses machine learning algorithms like logistic regression, Markov Chains, neural networks, and random forest. These techniques not just look at the time series of data, but also the classification of the data, which ultimately leads to someone winning or losing the game.
One can login and start predicting the outcome of games. Once you get the predictions right, users can then apply the same strategies on betting sites as well. At the moment, the selection is for free, and the website does not charge its users. Eventually, the company hopes to start asking people to pay for testing their strategies.
“You learn to get better at gaming sports platforms,” says Shubham.
For Predict22, scaling up by signing up with all the tournaments as strategic analytics partners, and making people use the site at the lowest customer acquisition cost was a challenge.
Despite being a good software expert and data scientist, Shubham had to come up with a business model that could help him scale, since it is a B2C business.
When asked about their business model, the founder said he does not want to reveal the pricing of the business as he is currently acquiring customers. The company also did not reveal details about the total number of users and its revenues.
The market and future
According to the new market research report Sports Technology Market by Technology, Sports and Geography - Global Forecast to 2024, published by MarketsandMarkets, the Sports Technology Market was valued at $8.9 billion in 2018, and is projected to reach $31.1 billion by 2024. It is expected to grow at a CAGR of 20.63 percent during the forecast period.
While the market opportunity is huge, Predict22 just has to predict its way to success by getting consumers to pay for the product, and ensuring consumers will money on betting platforms.
The spending on sports-based AI is also big in the US. Hudl, Synergy, Duul are some of the companies going after this market. In India, there is Dream 11, which is the poster child of India’s fantasy gaming ecosystem. It allows users to create imaginary teams and earn cash based on selected players’ actual match performance.
Each good move gets gamers cash rewards as well. And then, there is also Edisn.AI, which is in the business of analytics and visualisation.
In the next 18 months, Predict22 wants to increase its user base to a million users, and also plans to raise money from angel investors.
(Edited by Megha Reddy)