In today's data-driven sports world, developing an efficient football prediction model is crucial. I have recently had the opportunity to create a predictive model in C# that is capable of anticipating upcoming match outcomes with noteworthy accuracy. The major technical achievement lies in the sophisticated back-testing process employed, enabling the model to learn from historical data prior to making predictions. By combining multiple machine learning algorithms, finely tuning the parameters, and integrating advanced statistical methods, the model's reliability has been considerably enhanced. The quest for data ended with the discovery of a new sports data API, SportDevs (https://sportdevs.com/). This powerful API is brimming with historical match data, player statistics, and other valuable pieces of information that fueled the accuracy of our model. The SportDevs API integrated smoothly into the C# environment, providing a wealth of structured data for our predictive model to parse and learn from. Utilizing the auto-encoding feature in the .Net library to render the high-dimensional data manageable, I ensured that our model could handle the breadth and complexity of data coming from SportDevs API. The implementation of the new Multilayer Perceptron (MLP) approach led to improved predictions, showcasing how SportDevs data can be harnessed to generate a model which lends remarkable insights into upcoming football matches. I found that by structuring the task in this way, the prediction model truly improved, increasing the precision of the football outcomes it predicted. I am excited to share this C#-centered success and look forward to future applications of the SportDevs API in sports analytics.