How to read the integrity numbers
Three organisations publish counts of suspicious activity in sport, on three different bases, and the totals cannot be added together.
Three counts, three different things
On 10 February 2026 Sportradar published its 2025 integrity report. It recorded 1,116 suspicious matches across more than a million monitored events in 70 sports, a fall of about 1% on 2024. More than 99.5% of monitored events were free from suspicion.
The International Betting Integrity Association reported 70 suspicious betting alerts in the first quarter of 2026 and 76 in the second.
The International Tennis Integrity Agency reported 10 tennis match alerts in the second quarter of 2026, in a report published on 20 July.
Those three figures are all sound. They are also measurements of three different things, produced by three different methods, at three different cadences. Any total built by adding them is meaningless, and the sentence "there were X suspicious incidents in sport" has no publisher behind it.
1,116
suspicious matches identified by Sportradar through monitoring across more than a million events in 2025
70
suspicious betting alerts reported to IBIA by its member operators in Q1 2026
76
suspicious betting alerts reported to IBIA by its member operators in Q2 2026
What each basis actually counts
Sportradar counts matches. The unit is a fixture that monitoring has flagged as suspicious, and the denominator is every event monitored. That is why the 2025 report can say 1,116 and more than 99.5% clean in the same breath: both describe the same population. The figures are annual, so the most recent Sportradar publication reports the prior year.
IBIA counts alerts. The unit is a report from a member operator about betting activity it considers unusual on an event. One match can generate more than one alert, and an alert is an operator's judgement about its own customers, not a finding about a fixture. IBIA monitors more than 1.5 million events a year across more than 80 sports. Its membership runs to over 90 companies and over 200 betting brands. The reporting is quarterly, and IBIA notes that figures may be revised as investigations conclude.
ITIA counts tennis match alerts received under confidential agreements with the regulated betting industry, alongside its sanction outcomes. In the second quarter of 2026 it reported 10 alerts and nine anti-corruption sanctions, ranging from one month to 20 years. Its 2026 Tennis Anti-Corruption Program took effect on 1 January 2026.
Three units: a flagged match, an operator report, a sport-specific alert. Three thresholds. Three consequences. Nobody in the field disputes any of this, and it is still routinely lost the moment a number leaves the report it came from.
The 2025 baselines, stated properly
Sportradar's 2025 sport split was football 618, basketball 233, tennis 78, table tennis 65 and cricket 59. Europe remained the highest-volume region but recorded 66 fewer cases than in 2024, and South America 64 fewer. Asia, Africa and North and Central America rose modestly. Sportradar supported 125 sporting sanctions during the year, taking its cumulative total past 1,000.
IBIA's own full-year 2025 baseline was 300 alerts, up 29% on 2024. Those alerts led to 54 proven corrupted matches and 24 sanction outcomes, with North America accounting for 16% of the total.
Read those two paragraphs together and the divergence is obvious. One publisher recorded a 1% fall in flagged matches. The other recorded a 29% rise in operator alerts. Both describe 2025. Neither is wrong.
Why a rising count can mean better detection
IBIA's chief executive, Khalid Ali, addressed this directly when the 2025 figure was published: "The increase to 300 alerts in 2025 should not be read as a direct indication of a rising integrity risk."
That is not defensiveness. An alert count is a function of three variables at once: how much manipulation is happening, how many events are being watched, and how sensitive the watching is. If the second and third variables rise, the first can fall while the published number rises.
Sportradar's own data makes the point in a single comparison. Matches flagged with the assistance of its AI detection rose by about 56% in 2025, while the total number of suspicious matches fell by about 1%. The company attributes the increase to greater sensitivity to subtle and niche manipulation. More sensitivity, in other words, not more manipulation.
Ali has also cautioned against reading technology as the answer on its own, noting that technology is not a substitute for good governance. A detection layer that gets more sensitive without a governance layer that can act produces more alerts and the same number of outcomes.
What the first half of 2026 shows, and what it does not
On the alert basis, IBIA's first quarter covered 70 alerts across 10 sports, up 11% on the same quarter of 2025: football 25, or 36% of the total, tennis 16, esports 15 and table tennis 7. The second quarter covered 76 alerts across nine sports, up 10.1% year on year: football 25, esports 16, basketball 10 and tennis 9, with the remainder spread across darts, greyhound racing, cricket, table tennis and volleyball. Football alerts in the second quarter spanned 16 countries.
Two details in that split deserve attention. First, on the alert basis the highest-volume non-football sports in the first half of 2026 were esports, tennis and basketball. On Sportradar's 2025 match basis the order after football was basketball, tennis, table tennis and cricket. The two rankings are not in conflict, because they are not the same ranking.
Second, every table tennis alert in both quarters related to events outside the international federation's sanctioned calendar. That is a governance finding, not a betting finding: the exposure sits in competitions the federation does not run.
There is one thing these numbers cannot tell you. No full-year 2026 figures exist yet from any of the three publishers. Sportradar's 2026 publication reports 2025 data. IBIA's quarterly reports cover completed quarters and remain open to revision. Anyone presenting a 2026 annual picture today is extrapolating, and should say so.
A fourth basis, from a different direction
In evidence to the House of Lords International Agreements Committee on 21 July 2026, the Department for Culture, Media and Sport's Director of Sport and Gambling, Emma Floyd, reported 28% more corruption instances in 2025 than in 2024.
That is a fourth unit again: instances recorded by a government department, not matches, operator alerts or sport-specific alerts. It belongs in the same conversation and not in the same column.
What to do with this
Never combine. If a board paper adds figures from two publishers, it is wrong before anyone reads the conclusion.
Publish your own basis. Any federation or league reporting its own integrity numbers should state the unit, the denominator, the period and its revision policy. Doing so costs nothing and makes the figure usable.
Separate detection from outcome. Track alerts received, cases opened, cases closed with an innocent explanation and sanctions imposed as four distinct measures. A rise in the first with no movement in the others is an operational problem, not an integrity trend.
Ask what changed in the method. Before treating any year-on-year movement as a signal, establish whether the monitored population or the detection sensitivity changed. In 2025, both did.
Watch for the ITIA and IBIA reports covering the third quarter of 2026. Sportradar's next annual publication, due in early 2027, will be the first full-year picture of 2026. Until then, the honest answer about 2026 is that the year is not counted yet.
Sources
Related reading
InsightAI moves fast, and so must the system around it
Machine flagging jumped in 2025 while total suspicious matches fell. That says something about sensitivity, and something harder about who carries the burden of proof.
InsightThe market moved from results to moments
In-play micro-markets let a participant shift a price without changing a score. Monitoring built around outcomes does not see that happen.
InsightTwo bodies, 104 matches, opposite conclusions
FIFA reported nothing suspicious across the World Cup. The Group of Copenhagen recorded seven yellow notices. Sport has no shared definition of the word.