After calculating the probabilities for individual fixtures, the system simulates the entire NPFL season and uses the results to generate a projected final league table.
Made In Africa Sport (MIAS), a sports technology and media company, has launched an artificial intelligence-powered supercomputer designed to analyse and predict all 380 matches of the 2026/27 Nigeria Premier Football League (NPFL) season.
According to the company, the system runs 100,000 Monte Carlo simulations for each of the 380 fixtures, resulting in 38 million simulations across the season.
The predictions will be updated after each matchday to reflect new results and changes in the participating clubs’ performance.
Beyond predicting the likely winner of individual matches, the model estimates the probability of each club finishing in every position on the final table, including its chances of winning the league, qualifying for continental competitions or being relegated.
The full 2026/27 NPFL predictions, including match outcomes, expected goals, team ratings and the projected league table, are available on the MIAS Supercomputer.
Speaking at the launch, MIAS Executive Director Enitan Obadina said the product was developed to address what he described as a gap in the analysis of the NPFL.
“For a league with as much history as the NPFL, we still do not use enough of the data we have,” Mr Obadina said.
“Every week, there are conversations about which team is in form, who will win a match or who is likely to challenge for the title, but a lot of those conversations are based on opinion and eye-test.
“What we are trying to do is put a proper data layer behind those conversations. We have gone through thousands of historical matches and built a system that can turn that information into something people can actually use to understand the league.”
The model combines historical and recent performance data with head-to-head records, venue strength, goalscoring and transfer activity to determine the probability of a home win, draw or away victory for each fixture.
It also considers previous meetings between teams and their respective home and away records.
Goal-scoring form is assessed using the number of goals teams have scored and conceded in recent matches. In contrast, transfer activity is used to account for changes in squad quality that may not yet be reflected in previous results.
The final prediction provides the outcome with the highest probability, expected goals and other factors that contributed to the projection.
Mr Obadina said the system was not intended to replace football knowledge or judgement.
“We are not saying the computer knows what will happen. Football does not work like that. What it can do is show the probability based on the evidence available to it, and that gives you a much better starting point than simply saying a team will win because it looks stronger on paper.
“The important thing for us is that people can see how a prediction was reached. If we say a team has a 60 per cent chance of winning, there is data behind that number and the user can see the factors that contributed to it.”
The MIAS Supercomputer also estimates the number of goals each team is expected to score and provides ratings for attacking, defensive and transfer strength.
Transfer strength is calculated separately based on a player’s performance in the previous season. This allows the model to account for changes in squad quality that may not be captured by historical match data, particularly given the extent of squad changes made by NPFL clubs during transfer windows.
MIAS said the model was developed specifically around the characteristics of the NPFL rather than applying a generic football prediction model.
Historical data in the company’s database indicates the significance of home advantage in the competition.
Since 2003, away teams have won 601 of the 7,966 NPFL matches recorded in the database, representing 7.5 per cent of the matches.
The company said incorporating such historical patterns was intended to ensure that the predictions reflect the realities of Nigerian domestic football rather than trends from European or other international leagues.
After calculating the probabilities for individual fixtures, the system simulates the entire NPFL season and uses the results to generate a projected final league table.
The projections will change throughout the season as actual results are added to the model.
Mr Obadina said the 2026/27 edition was the first major version of a broader data project that MIAS intends to develop over time.
“This is the first major version of what we want to build. We want to keep improving it as more NPFL data becomes available, so that the model becomes more useful from one season to the next.
“The bigger ambition is to build a reliable data infrastructure around African football. We want to get to a point where journalists, clubs, analysts, supporters and other stakeholders can ask meaningful questions about our leagues and have credible data to work with,” he concluded.
