The pandemic’s impact on sports events has not only affected fans but also the gambling syndicates’ evasion tactics.

Criminals involved in sports-gambling and fraud have had to diversify and improvise due to the massive impact of the pandemic. In the past, match-fixers targeted sports and leagues such as football, cricket tennis and basketball — where profits and turnover were higher.

Now they have diversified into other sports and leagues that receive less attention traditionally. In response, advanced data analytics and monitoring technologies are helping capture, control and analyze massive volumes of data to detect fraud and prevent match-fixing.

Artificial intelligence (AI) is also being used to track manipulation of sports events. Here are more insights on how technology is being used to keep fraudsters in check, gleaned from an interview with Andreas Krannich, Integrity Services Managing Director of Sportradar AG, a firm specializing in collection and analysis of sports data.

Andreas Krannich, Managing Director, Integrity Services, Sportradar AG

What tech solutions do you use to combat match-fixing?

Andreas Krannich (AK): Since 2005, Sportradar has used its technically advanced bet monitoring system — the Universal Fraud Detection System (UFDS) — to detect match-fixing across global sport.

The UFDS comprises sophisticated algorithms and a constantly maintained database of odds that are used for the purpose of detecting match-fixing. With this system we support sport, law enforcement and state authorities to monitor, detect and analyze betting-related manipulation and other types of corruption. Currently, the UFDS team at Sportradar has expert analyst teams located in London, Sydney, Singapore, Montevideo, Minneapolis and Las Vegas; allowing for 24/7/365 oversight of global betting activity.

From a technological perspective, the UFDS collects in real-time information from the broadest range of relevant betting operators and sources around the world, with over 7.5 billion datasets processed from across over 600 betting operators.

The UFDS is also supported by account-level individual betting data at more than 130 operators worldwide, allowing for the expansive identification of suspect betting activity to be used effectively for integrity purposes.

Most people’s conception of match-fixing is that it is usually run by big criminal organizations. How can technology be used to cut match-fixing at the source, or can it only be minimized when it is already happening?

AK: Match fixers can come from any part of the world and may belong to small or big criminal organizations, or they may be independent operators.

Policing and prevention in equal measure, is the way forward in combating match fixing. Policing includes monitoring, detection, investigation and prosecution.

At Sportradar, we have monitored over 600,000 matches a year, and have detected more than 6,000 suspicious matches involving major sporting leagues in football, basketball, e-sports, tennis, volleyball, and cricket in the last 15 years through UFDS. It has resulted in 456 sports disciplinary sanctions and 51 criminal convictions.

Prevention, which includes education and disruption, is the other side of the coin. If players are informed of the rules and have regular exposure and access to materials about the dangers of corruption, an integrity unit will be able to disrupt attempts at foul play and consequently, the number of issues will decrease.

For match-fixing to happen, players would have had to be involved, such as in a televised basketball game in the Philippines that was stopped during halftime. With the technology now available, could such a game have been stopped preemptively?

AK: With a data-driven approach to utilizing the latest technology, match-fixing can be detected early on… in real time.

However, it is generally recommended that a full analysis is completed on a match and the associated betting data, because concluding that a match has been definitively fixed is a serious judgment.

So, while there are some cases where preventative action such as postponing a match before kick-off may be the right approach, it should be left up to the competition’s governing body to determine the right course of action based on the evidence available at the time.

How are machine learning and AI used for identifying match-fixing?

AK: Data is used to detect anomalies and suspicious betting patterns in games. While historically this would have involved data analysts examining data streams in the search of indicative outliers, today it is a combination of machine learning algorithms alongside data analysts who can confirm suspicious behavior patterns.

Another area where we are building our capabilities is AI to improve fraud detection, prevention and prediction. With predictive insights and real-time analysis you can beat fraudsters before they strike.

Our UFDS technology uses both AI and machine learning, and we keep constantly evolving and improving it. This has resulted in developments such as: smarter alerting, learning from the betting data that has been gathered over the past 10 plus years, customized alerting developed for different sports and for different levels of competitions within the sports.

In fact, we are offering UFDS to all sporting federations and leagues at no cost, to enable even the lower-tiered clubs that cannot afford these services to examine punting trends on their particular sport.