Our original goal
Our goal for this project was to create an application where our Codit colleagues could predict the outcome of the FIFA World Cup 2022 tournament games, competing against each other and AI bots through a leaderboard. This year, we also wanted to open up the application to our customers!
Additionally, we wanted to implement and use the newest technologies, make sure the application was easy, fast and safe to use, and have some fun during the creative process!
How did we fare in the creation of the application?
Now that the tournament has ended and Argentina has emerged victorious, we can look back at what we originally wanted to achieve, and how we did.
First of all, we can safely say we succeeded in creating a great application that was easy to use. This was down to the great user interface, which was very fast because of the scaling options in Microsoft Azure. It was also safe for our customers to use, thanks to our use of Azure AD B2C. Onboarding new customers wanting to join in was relatively easy to do, and they could easily log in to our application’s website using their own Azure Active Directory user account. We received some really positive feedback from customers who joined in, as well as some friendly competition!
We paid extra attention to making sure the API’s and the central database where we stored all the predictions and results could not be tampered with. This was through OAuth, managed identities, firewalls and other cool features such as the rate-limiting policies in Azure API Management.
During the duration of the tournament, we had zero incidents or downtime. However, we were also able to deploy new features such as being able to view the results of a colleague’s predictions! This is where the implementation of CI/CD pipelines in Azure DevOps came in very handy.
While creating a working application was obviously the main goal, we were very happy to see that we have had some great collaboration between colleagues and the different technology practices within Codit, all while learning about new technologies and applying our latest standards!
So how did we actually do in predicting the outcome of the games, and how did the Artificial Intelligence bots compare to us humans?
Within Codit, only 3 colleagues finished better than the top AI bot, which finished in fourth place! This was even better than expected — better than the AI bots fared during the UEFA tournament. A great result!
Looking at the data, we found some interesting results:
As you can see, a lot of predictions were made! On average, the AI fared a little better in the average predictions.
The big surprise of the tournament was Argentina losing to Saudi Arabia, and we can see that reflected in the data. In this game, the least number of points were scored — only 4 out of a possible 730. On the other hand, The Netherlands versus Qatar game was the most predictable.
Here is a look at the progression of the leaderboard during the tournament. The names of participants have been anonymized.
It has been a joy to work with our Codit colleagues on this project. Coming together, we have shown that we can create something that combines many of our practices and standards, as well as achieve a great result.
We did what we originally set out to do and more, as during the tournament we had feature requests coming in and added some new functionality. We opened up the application to our customers and got some enthusiastic feedback from them.
It was awesome to see the number of people using our application, not only within Codit but among our customers as well. It certainly generated some friendly competition! 😉
A huge thank you to our development team at Codit, and everyone who joined in to make their predictions!
See you at the next tournament!
Subscribe to our RSS feed