The 2019 Cricket World Cup finished five months ago, but there is an analytics startup that grew in popularity just before it began and is still going strong to this day. This is known as Predict 22 and they spend their time trying to give people accurate predictions about today’s match so that they can increase their chances of winning money.
What Did They Predict for the World Cup?
Before the World Cup started, they predicted that India, England, West Indies and Australia would qualify for the semi-finals. They went on to predict that England would lift the trophy and that the West Indies would be dark horses for the competition.
They got three of the four semi-finalists correct as England, India, and Australia all made it past the group stage. However, the West Indies had a poor tournament and were eliminated early on. They also correctly predicted that England would win the tournament.
So, how did they come up with such predictions? Well, the founder and CEO, Shubham Bindlish, says that they make use of databases and artificial intelligence to make extremely accurate predictions. He claims that they are the only analytics company in India that are using such technology to give predictions for sporting events. He also claims that there is nobody else on the market who can get their predictions as accurate as they can.
So Why Start with Cricket?
Well, as everybody is well aware of, cricket is extremely popular in India. So, it made sense to Bindlish to create a program that could predict the outcome of cricket matches. He is also a huge cricket fan himself, and it always nice to work on something related to an activity that you enjoy. Another reason why he chose cricket and not another sport such as football was that stats are easily available in cricket as there are many different databases available with all the information that you will ever need.
One fascinating thing about Bindlish is that he has his own database for Indian players. Ever since he was a child, he has been following Indian players and adding their personal stats to his own database (notebooks).
When he graduated with a degree in computer science, he got a job with Goldman Sachs and started to experience how data can be used for many things. He became more intrigued by data and its potential to answer questions such as who will win this match or who will win the tournament.
He called the program Predict22 due to the fact that cricket pitch is 22 yards long and that 22 players play in each match.
He Started with T20 Cricket
T20 cricket was a great place to start as it was the newest cricket format and there were many different databases available. He decided that the 2016 T20 World Cup would be the first tournament that Predict22 would try and predict. Him and his team received promising results from this tournament, so they kept on expanding their database and building algorithms to be able to predict the Indian Premier League and other T20 competitions.
Intrigued by his own successful predictions, he started to investigate whether it would work for other sports. As an ardent Chelsea fan, he decided to test it on football, with the World Cup in 2018 being the guinea pig. After this, they expanded to English Premier League, Serie A, La Liga, the Bundesliga, and the Indian Super League.
A lot of the predictions came true, but he wanted to see whether their algorithm would work on a sport that he does not follow. The sport that he chose was basketball, and they got strong results once again.
Making Use of Data
In order to make their accurate predictions, it stands to reason that they need a lot of data. His team have gathered a lot of historical data from free databases. The data that they collect is cleaned and stored in their in-house database.
They then use complex technology to process all of the data with ML algorithms, helping them learn more about it. This includes information such as match performances and player ratings. Then, the enriched data, alongside other algorithms, is put into a neural network which then delivers the predictions for matches and tournaments.
For their algorithms, they take the context of the game into account. For example, a century that is scored in an innings that only had 250 runs scored is more valuable than a century that is scored in an innings where 550 runs were scored. Also, getting a top batsman out, such as Virat Kohli, Joe Root, Kane Williamson, or Steve Smith is more valuable than getting a tailender out.
A Look at How They Have Grown
At the time of writing, Predict22 has 850,000 users. For the 2019 Indian Premier League, around 500,000 people used the program. This helps to highlight that an average cricket fan in India wants to add an analytical element to their cricket experience. When you consider just how much Indians love to bet on cricket, this will come as no surprise. After all, who would be stupid enough not to take advantage of something that can drastically improve your chances of winning?
In the coming year, they are planning to add a lot more features such as prediction analysis, player performance, and stats that are easy to scroll through. Predict22 is a tool that is definitely going to be around for many more years to come.