Industries / Retail, hospitality and personal services / Hospitality and tourism / Gaming and casinos
Sports betting
Taking bets on the results of sporting events, and the work of keeping those events from being fixed for bettors.
7
Milestones achieved, from 1920 to 2026.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1920
A commissioner after a fix
Eight Chicago White Sox players were accused of fixing the 1919 World Series for gamblers linked to Arnold Rothstein. Baseball named Kenesaw Mountain Landis, a former federal judge, its first commissioner in 1920. His first major act was to ban the eight for life, and he kept the ban after a Chicago jury acquitted them.
LessonA sport can punish fixing under its own rules even when a criminal court does not convict.
Credit
Kenesaw Mountain Landis and Major League Baseball
2009
Tennis builds an integrity unit
Heavy betting against Nikolay Davydenko in an August 2007 match in Sopot, Poland, led tennis's governing bodies to hire Ben Gunn and Jeff Rees to review the sport. Their May 2008 review examined 73 matches flagged for suspicious betting. The bodies adopted one anti-corruption code and jointly funded the Tennis Integrity Unit (TIU), which began work in January 2009. The ATP also signed agreements with Betfair, a betting exchange where customers bet against each other, to get reports of suspicious betting.
LessonThe 2018 independent review found the new unit never followed up the Sopot inquiry's intelligence on other players, some of whom were later involved in suspicious activity.
Credit
ATP, WTA, ITF and the Grand Slam Board; Ben Gunn and Jeff Rees
2016
Fraud alerts pass an audit
Sportradar's Fraud Detection System watches betting markets for UEFA and flags irregular odds movements before and during matches. It covers more than 31,000 matches a season across the top two leagues and main cup of each UEFA member country. An outside study by David Forrest (University of Liverpool) and Ian McHale (University of Salford) found that matches it flagged as suspicious were "very likely to have been manipulated".
LessonThe auditors were struck by how rarely flagged matches led to action by sports bodies or police.
Credit
Sportradar and UEFA; audited by David Forrest and Ian McHale
2018
States may legalise sports betting
On 14 May 2018 the US Supreme Court, in Murphy v. NCAA, struck down the Professional and Amateur Sports Protection Act (PASPA), the federal law that barred states from authorising sports betting outside Nevada. Justice Alito wrote for the majority that Congress cannot dictate what state legislatures may and may not do.
Credit
State of New Jersey, which brought the case; US Supreme Court
2018
Tennis limits live data
An independent review panel found that betting-related corruption in tennis was "particularly acute" at the lowest professional level, where a TIU investigator called it a "tsunami". It blamed in part the ITF's 2011 and 2015 deals selling official live scores to Sportradar, which made tens of thousands of low-level matches available to bet on. It recommended ending official live data for the new $15,000 prize-money tier from 1 January 2019.
LessonSelling data on minor matches multiplied the matches that could be fixed for bettors.
Credit
Independent Review Panel of Adam Lewis, Beth Wilkinson and Marc Henzelin
2025
Caps on single-pitch bets
On 9 November 2025 federal prosecutors in New York's Eastern District charged two Cleveland Guardians pitchers, Emmanuel Clase and Luis Ortiz, with rigging individual pitches for bettors who wagered on whether a pitch would be a ball or below a set speed. Both deny the charges. The next day MLB said operators with more than 98% of the US betting market would cap bets on single pitches at $200 and keep them out of parlays, bets that combine several wagers.
LessonMLB's own statement says bets on one-off events that a single player decides carry the highest risk of fixing.
Credit
Major League Baseball and its sportsbook partners
2026
Fantasy apps bypass prop bans
A prop bet is a bet on one player's statistics. Fifteen US states with legal sports betting ban them on college athletes, partly because losing bettors harass players. Researchers at the University of Maryland's Povich Center and Howard Center collected nearly 21,000 chances to bet on individual athletes during the 2026 men's March Madness tournament from 14 sportsbooks and fantasy platforms. The bans applied to just 38% of them, because "pick 'em" fantasy sites such as PrizePicks and Underdog are often not regulated as betting.
LessonFor every 100 such chances offered by sportsbooks, pick 'em sites offered 163, outside the rules written for sportsbooks.
Credit
Shirley Povich Center for Sports Journalism and Howard Center for Investigative Journalism, University of Maryland; published by Capital News Service, October 2026
Watch for
Sports bets sold as swaps
Kalshi, an exchange registered with the Commodity Futures Trading Commission (CFTC), says its sports event contracts are swaps, financial contracts overseen only by the CFTC, so state gambling laws cannot apply to them. The Third Circuit sided with Kalshi 2-1 against New Jersey in April 2026. The Ninth Circuit (Nevada, 28 August 2026) and the Sixth Circuit (Ohio and Tennessee, September 2026) ruled against it, and the CFTC filed a brief backing Kalshi in May 2026. New Jersey asked the Supreme Court to review its case on 2 September 2026. On 30 October 2026 the Court extended Kalshi's time to respond to 7 December 2026; it has not said whether it will hear the case.
Achieved whenThe Supreme Court decides whether federal commodities law overrides state gambling law for these contracts.
Claimed by
Kalshi, in court cases decided April to September 2026
Machine learning flags fixes
Sportradar says it detected 1,116 suspicious matches across 12 sports in 94 countries in 2025, and that its machine-learning system, UFDS AI, identified 682 of them, up from 436 in 2024. It calls its figures indicative because it does not monitor every competition the same way.
Achieved whenAn independent audit, like Forrest and McHale's study of the earlier system, reports what share of the matches the AI flagged were confirmed as fixed.
Claimed by
Sportradar, February 2026
Open
01
Proven fixes in one table
An open, match-level table of fixes proven in disciplinary cases, starting with the published decisions of the International Tennis Integrity Agency (ITIA). For each corruption case, record the player, the match, the date, the tournament tier, the offence, the sanction and the page of the decision it comes from. Then add the public result of each match and, where they exist, public pre-match odds.
The briefCredit · unclaimed02
A register of prop rules
A public table, updated every month, with one row for each US state and Washington, DC. It records whether sportsbooks may offer player props on college athletes (all, out-of-state players only, or none), whether fantasy pick 'em contests on college athletes are allowed, and links to the law or regulator order behind each answer. Every change is logged with its date.
The briefCredit · unclaimed
Open 01 · the brief
Proven fixes in one table
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open, match-level table of fixes proven in disciplinary cases, starting with the published decisions of the International Tennis Integrity Agency (ITIA). For each corruption case, record the player, the match, the date, the tournament tier, the offence, the sanction and the page of the decision it comes from. Then add the public result of each match and, where they exist, public pre-match odds.
Why it is still open
The ITIA posts its decisions as separate PDF files, 86 of them linked from its sanctions page, with anti-doping and corruption cases mixed together and no table of the matches involved. An open-source match-fixing detector published on GitHub in May 2026 says that testing it against known fixing cases is its "next milestone", and Sportradar's own detection data are private. This research found no public match-level list of proven fixes to test such tools against.
Built with what exists
The ITIA's decision PDFs; pdftotext to turn them into text; a language model such as Claude to draft each row for a person to check (Note "Seven more flagged"); a public Git repository for the CSV file and its change log.
First step
Download every decision linked from the ITIA sanctions page, sort corruption cases from anti-doping cases, and extract the matches from ten decisions by hand to settle the table's columns.
Done looks like
At least 150 match-level offences in a public CSV file, every row checked by a person against the decision with a page number, and the share of those matches that at least one open detector flags, wherever public odds exist.
Open 02 · the brief
A register of prop rules
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public table, updated every month, with one row for each US state and Washington, DC. It records whether sportsbooks may offer player props on college athletes (all, out-of-state players only, or none), whether fantasy pick 'em contests on college athletes are allowed, and links to the law or regulator order behind each answer. Every change is logged with its date.
Why it is still open
The University of Maryland analysis that counted 15 states banning all college player props was a one-time snapshot of the 2026 March Madness tournament. It found that the bans reached only 38% of chances to bet on individual college athletes, because pick 'em fantasy contests fall under different laws in each state.
Built with what exists
State statutes and gaming regulators' published rules and orders; a browser agent such as Browser Use 1.0 to fetch them on a schedule (Note "Deleting 80% to start over"); a diff tool and a public Git repository.
First step
Start with Maryland, which the Maryland analysis names as one of the 15 banning states. Find the law or regulator order behind its ban and its rule on fantasy contests, and record both in the table format.
Done looks like
51 rows, each with two answers that link to a primary legal text, re-checked every month for six months, with each change dated.