Introducing the sports betting data supply chain and the predatory integrity industry

Sports gambling is no longer just about putting on a bet before kick-off and picking the final result. In today’s betting market, in-play betting is king.

This type of betting (also known as live betting) takes place once the game has kicked off, and it means the odds will be in a constant state of flux. For gamblers, it heightens the thrill of the chase - can the underdog hold on? - and it means that modern sports betting websites are alive with incoming data.

For football alone, a single site can offer thousands of games across a weekend, each with a feast of real-time betting data. Odds are offered on multiple markets, from the next goalscorer to the number of cards to handicap options, and are updated continuously.

Up-to-second statistics for each team show shots on goal, dangerous attacks and similar information. Many sites will even provide an animation of the game, a simplified illustration of what is taking place at the venue, with the ball moving up, down and around a two-dimensional pitch.

This stock exchange of sport is made possible by a handful of sports data providers: primarily Sportradar, Genius Sports and Stats Perform. They are giant, influential and fiercely competitive rivals - with a shared interest in stimulating the already insatiable appetites of the world’s gambling companies for fresh games to offer.

Live bets are primarily offered on lower league matches

While they battle in court over the exclusivity to cover competitions like the Premier League, it's down in the lower reaches of the football pyramid they find most of the data they distribute.

That is how you end up with the absurd situation where Australia - a country with just a dozen professional teams - is the dominant source of betting data for many parts of the year. One betting site alone, which operates illegally in a number of countries, can offer almost 250 games from Australia on a typical weekend. 

The title of one competition, the Sunday League, gives an insight into the community nature of many of these matches. During the week, you might find games from a regional competition where 60-year-olds can play, thanks to the use of unlimited substitutions. 

This is truly a social event, and therefore it is so much more remarkable that one sports data provider (Genius Sports) has actually signed a deal for the rights to collect data from these matches for betting purposes, while another prefers to just turn up and take it for free. While the company can be kicked out of a ticketed venue, kicking them out of a public park is another matter altogether. 

It echoes the predatory behaviour of the betting industry that these companies serve. It also creates a major integrity risk: when amateur sports are on the betting market, amateur athletes become targets of dangerous organised criminals. 

Players at the lower reaches are cheaper to buy off, and with less scrutiny of their games, may feel uninhibited to cheat. They also receive little to no training in how to deal with the approaches of fixers, who may spend months grooming their targets and then threaten them if they do not comply. 

The susceptibility and corruptibility of these players is borne out by Sportradar’s own data: it found more than half of all suspicious games came from the third tier or lower in 2022, including youth games.

Our analysis of the European football market found that between them, the three rival companies provided data for four out of five games on the betting market. Subsidiary companies may have provided others. 

More than half of all games offered for live betting were from semi-professional, amateur, youth and regional leagues. In an audit we undertook of the betting market during the pandemic, more than 800 sub-professional games from Europe were available for live betting on one weekend alone. 

In countries like Spain and Germany, which have the highest number of games on betting markets most weekends, just one in four games offered would be classed as professional. The numbers are similar in England.

The proportion of lower league games is less pronounced in countries like France, where authorities are more hostile to the gambling industry. 

That suggests it is not an inevitability that community-level games will end up on the international betting market, but that their presence is a reflection of local attitudes. It also contrasts with claims by some in the live sports betting data industry that “If we don’t supply this data to bookmakers, really bad people will”. 

Some companies' data scouts use an earpiece to transmit results, speaking the action into a phone connected to a call centre somewhere in the world. Most companies have specialised apps featuring a football pitch. Data scouts punch the movements into the app and press other buttons for significant events like goals, cards and corners.

It is a surprisingly inexact science, open to much interpretation, meaning that each data feed carries a distinct fingerprint. These fingerprints can allow you to follow the data further up the data supply chain.

Conflicts of interest when data companies also provide integrity services

So where does all this data end up? Live sports betting data companies are generally reluctant to reveal their client lists, but there are multiple reasons why this is important to know. Chief among them is that many of these same companies also offer “integrity services” to sports.

Through monitoring the betting market for signs of suspicious activity, such as a large and unexplainable shift in the odds that could suggest signs of manipulation, they promise to protect these competitions’ games from match-fixing. 

When these movements are suspicious, the companies issue alerts to the relevant bodies. These are not smoking guns, but they are red flags worthy of investigation, and sports data companies employ talented investigators who can collaborate with leagues and authorities. 

This system can produce significant results and the sports data companies undoubtedly have natural advantages in this space. But there are clear and significant conflicts of interest here, especially when these same companies are making live, in-play betting involving vulnerable lower-tier games available to bookmakers, and often without the knowledge of those taking part. 

“This is like putting a cigarette producer as the chief of an oncology surveillance system,” the chairman of Ukrainian football’s Ethics and Fair Play Committee, Francesco Baranca, tells Play the Game.

It’s an effective business model: first, you create the poison, and then you supply the medicine. Some data companies have even been known to inflate the number and importance of their match-fixing alerts, to win more integrity contracts. Yet few in sports administration seem to recognise this conflict of interest, and FIFA even abandoned its  in-house system many years ago to rely on an external integrity partner.

The value of match-fixing alerts is limited

Furthermore, the impact of this integrity work has been diminished by multiple recent findings of the Court of Arbitration for Sport. On at least four occasions, CAS has overturned bans on players for alleged match-fixing after finding that monitoring of bets is not of itself sufficient evidence. Odds analysis also needs to be supported by an actual investigation, it has concluded.

“The Sportradar Report cannot be deemed as having evidentiary value to support any specific wrongdoing of the appellants,” it said in one case in August this year.

In another, the global players’ union, FIFPRO, successfully argued that a ban against one player “follows a worrying trend in football where players are sanctioned without proper evidence for alleged involvement in match-fixing”. That evidence, FIFPRO said, “is mainly based on reports of sports data companies and a subjective analysis of the match.”

In other words, betting monitoring can help identify issues to investigate, but without further evidence, they alone are of limited value. 

Yet investigating cases of match-fixing can be extremely hard to do, especially retrospectively. As one expert witness in another CAS case pointed out: “It is quite difficult to ascertain whether the many errors committed in the latter stages of the games concerned are deliberate and malicious or merely manifestations of inadequacy and fatigue.”

This is why police in Australia once put a microphone into a goal frame as they investigated a major match-fixing ring: they needed actual evidence of a conspiracy, such as conversations between players to let goals in. So when the fix is a one-off, it may be too late to prove anything untoward took place. 

It means these games might have little real protection against match-fixing, and that even a country like Australia, which is generally considered a well-regulated environment, could have become a match-fixers’ paradise. If it hasn’t already.  

According to an analysis of betting alerts conducted by United Lotteries for Integrity in Sports (ULIS), a body representing many international state-backed gambling ventures, only three countries have had more issues in 2023 (to September) than Australia, with all alerts being for lower-tier or cup competitions featuring lower-tier teams. 

Unlike data companies, ULIS is not in the business of selling live data, which may explain why its findings about Australia look so different to those provided by a data company the previous year. It found no suspicious bets in Australia or anywhere in Oceania.

Shady endpoints

So if data companies are supplying integrity services, are they also supplying dodgy bookmakers such as 1XBet at the same time?

1XBet was founded in Russia but is now based in Cyprus after being banned in its homeland. It is sanctioned in Ukraine, had its licence cancelled in England, and was recently run out of Morocco. One of its main brand ambassadors is a popular porn star and it is the website that last year streamed a fake version of the Indian Premier League competition until police in India arrested the organisers of the competition. 

It is therefore of great public interest to know which data companies might be doing business with a bookmaker like this.

While the source code of many websites will reveal the origin of its data, this is obscured on 1XBet. So we instead looked at the unique “fingerprint” in its data. The “dangerous attack” count is one of the best methods for assessing this. 

A “dangerous attack” occurs when a team takes the ball into the final third of the pitch. A data scout’s decision to call a move a “dangerous attack” is therefore based on some subjectivity, and also perhaps on how seriously they are paying attention to the game.

If 1XBet shows the home team with 50 dangerous attacks, and a website using data from Sportradar shows 25, then evidently, 1XBet’s data cannot be coming from Sportradar. But if a site using data from Stats Perform also shows 50, this could indicate the origin of the data. 

The graph below shows a situation just like this. In a game in an Australian state league between the youth teams of the Wollongong Wolves and the Sutherland Sharks, the number of dangerous attacks shown on 1XBet was consistently aligned with the number of attacks seen on sites using Stats Perform’s data. This was true for both the home and away teams, and these feeds were in sync for the entirety of the match.

This trend of 1XBet data aligning with Stats Perform’s was observed on a minority of matches. Data from Sportradar or Genius Sports was never seen to align with that used by 1XBet. These data streams looked different to 1XBets, and often wildly so.

However, we also looked at the data used in the on-screen animations that illustrate the movement of the ball around the pitch. In the passage of play below, from a game in the Hong Kong League, it was observed that the coordinates displayed on 1XBet had an almost perfect degree of alignment with data sent by Sportradar to sites. (The one point where they do not match is likely due to a rounding error.)

This data is so aligned that it would seem unexplainable by chance. Two data scouts simply would not be able to replicate each other's feeds with this level of precision. This does not rule out the possibility that data was sold to 1XBet by a third party, or that the bookmaker stole the data.

Nonetheless, the URL of a Sportradar data file also gives another insight into a possible business connection between the data company and 1XBet. In the links below, the information between the first slashes represents the name of the client that the file is created for. In these cases, Bet365 and Parimatch.  

https://lmt.fn.sportradar.com/bet365/en/Etc:UTC/gismo/match_info/43255167

https://lmt.fn.sportradar.com/parimatch/en/Etc:UTC/gismo/match_info/43255167

When that client is changed to a non-existent client, the file will return an error message. Here, we attempted to use playthegame and this_is_not_a_real_company as the client. That resulted in an error message. 

https://lmt.fn.sportradar.com/playthegame/en/Etc:UTC/gismo/match_info/43255167

https://lmt.fn.sportradar.com/this_is_not_a_real_company/en/Etc:UTC/gismo/match_info/43255167

However, if you enter 1XBet as the client, the normal files appear. It shows that the Russian bookmaker is in the Sportradar system, even if it is not an active client.

https://lmt.fn.sportradar.com/1XBet/en/Etc:UTC/gismo/match_info/43255167

8XBet’s data also mirrors those of Sportradar

Another website of interest is 8XBet. Unrelated to the Russian-Cypriot bookmaker, 8XBet is the mysterious betting partner of Manchester City which appears to be operated by a “marketing firm” in Dubai and connected to a “cyber slavery” compound in Cambodia. Judging by the absence of one on its website, it currently operates unlicensed.

Through a basic visual comparison, it can be seen that 8XBet uses the same data as sites taking a Sportradar feed. For example, in a recent Champions League match between Union Berlin and Real Madrid, match statistics (e.g., “shots on target” and “ball possession”) were aligned between 8XBet and Sportradar, as were the movements of the ball on the sites’ animations. In some instances, 8XBet’s ball was even ahead of that on the Sportradar site.

The source code of 8XBet reveals that the live-feed data files it receives are identical to those received by the Sportradar client - with one key exception. The files used by 8XBet are not from sportradar.com, but from an obscure domain whose ownership is not publicly known. Does it raise the possibility that a “secret server” is used to service an unsavoury client? 

An open-source integrity service

It would appear that the current business model of sports betting data companies is incompatible with the Macolin Convention. This international treaty, designed in Europe, is aimed at tackling the global threat of match manipulation. 

It calls for, among other things, “the limitation of the supply of sports betting” on the lower-level football competitions that we have demonstrated are the dominant offering on the global betting market.

Bets on under-18s competitions should also not be offered, it says, yet as we have compiled this report, we have witnessed FIFA’s Under 17 championship being available for live betting around the world, thanks to one of the companies described in this article.

Article 10 of the convention calls for the prevention of “competition stakeholders being involved in compiling betting odds for the competition in which they are involved”, in order “to prevent conflicts of interest”. Stakeholders here include “other accredited persons”, which could be read to include anybody authorised by a competition - such as a data company or integrity provider.

Meanwhile, the following article in the convention calls for the “direct and indirect restriction of access to illegal remote sports betting operators” and the “raising of consumers’ awareness of the risks associated with illegal sports betting”. 

The convention’s definition of “illegal sports betting” is clear and unforgiving: it means “any sports betting activity whose type or operator is not allowed under the applicable law of the jurisdiction where the consumer is located”. 

Having a license to operate in Curacao - a popular jurisdiction for bookmakers looking to operate with limited regulation - is not a free pass to flout the laws of other nations and offer your services to their residents. Yet when we made one sports betting data company aware that several Curacao licensed bookmakers that it was supplying data to appeared to operate in breach of Australia’s Interactive Gambling Act, it insisted that it could not police the activities of its clients. 

All things taken into account, it is hard to see how sports betting data companies can be considered an appropriate choice of integrity partner for a sport or a nation that believes in the Macolin Convention.

Recognising this, the Canadian province of Ontario moved to ban sports data companies from providing integrity services.

“Independent integrity monitors shall not have any perceived or real conflicts of interests in performing the independent integrity monitor role, including such as acting as an operator or as an oddsmaker”, reads its legislation. 

A possible solution to the problems raised in this article could be an open-source alert system for match-fixing. Such a system would operate independently and free of the commercial conflicts of interest that cast a shadow on the work of the integrity industry as it stands. 

It would operate in the public sphere, meaning that all stakeholders - from officials to fans to players - can see for themselves how the markets moved on their matches, and if it flagged suspicious activity. This type of information is not currently publicly available, despite being based on odds which very much are. For the safety of the game, it really needs to be.