Trade paper

Markets Texas Capital puts Sands China's 2028 Macau earnings just under US$2.28 billion.

Regulation

Illegal payments in this bank sample sit inside licensed play.

Of the 146 with a problem-gambling score, 45 percent scored 8 or more, against 6 percent of licensed non-lottery gamblers in the same file. The study does not show that the illegal payment caused the harm.

bookmaker.solutions

An empty casino pit, a crimson roulette table under gold light, a broad scatter of chips with one tighter stack nested in the middle, emerald chairs, magenta and teal lamps.
Our picture. It is not a photograph from the source. The frame is this paper's mark.
What was matched
3,426 transactions by 179 people. Average age 44, and 115 were male. Scores exist for 146. The panel is 38,000 people, July 2018 to July 2026.
Licensed play alongside
Among people with repeated detected illegal transactions, 91 percent also used a licensed operator. In the first month after a first detection, 122 people with at least two unlicensed payments averaged 4.9 unlicensed transactions and 23.7 licensed ones.
Scores in this file
A score of 8 or more is 45 percent of the illegal-payment group and 6 percent of licensed non-lottery gamblers in the same data, 12,543 people. A score of 0 is 26 percent against 68 percent.
The account
Credit cards were 138 transactions, 4 percent, by 15 people. Current accounts were 3,288 transactions, 96 percent, by 172 people.
What the file can settle
Open-banking payments at participating institutions, and only operators the Commission could already name. It cannot size the illegal market, and it cannot show cause.

The Gambling Commission published an open-banking study on 28 October 2026. SBC News reported it on 30 October. The counts below come from the Commission’s report, which this desk opened. YouGov gathered the banking records with consent, from 38,000 people, between July 2018 and July 2026. The records were anonymised and linked to survey answers those people had already given, including a Problem Gambling Severity Index score where one existed. The panel is not a cross-section of Britain. It leans toward men, and women over 55 are thin in it. Everyone in it agreed to share banking data for a small incentive.

A hunt for the names of known illegal operators inside payment descriptions found nothing. The Commission takes that miss to mean those firms seldom put a trading name on the line a bank shows. Two later passes used text that was already tied to illegal play. Matching descriptions taken from known statements produced 821 transactions. Matching merchant-name fields against names strongly suspected to belong to illegal operators produced 2,628, and 23 of those had already been counted. The unique total is 3,426 transactions, made by 179 people. Their average age was 44, and 115 were male. For scale, the same file holds 4.8 million transactions with licensed operators once lotteries are left out, and 157.2 million transactions of every kind.

The illegal payments turn up inside a licensed habit. Among people with repeated detected illegal transactions, 91 percent also used a licensed operator. A narrower cut, 122 people with at least two unlicensed transactions, shows the mix month by month. In the six months before the first detected illegal payment, licensed transactions average 28.8, 26.9, 23.4, 23.1, 25.3 and 24.8, and the unlicensed count is zero because nothing illegal has been detected yet. In the first month after that detection the averages are 23.7 licensed and 4.9 unlicensed. Over the five months after that, unlicensed averages 1.8, 1.4, 1.9, 1.2 and 1.5, while licensed averages 23.5, 25.3, 23.2, 19.4 and 23.6. The illegal payments arrive in short runs, often close in time to a licensed one. The licensed count stays in the same band.

The harm scores in this file are far higher on the illegal-payment side, and the Commission says the design cannot show that the illegal spending caused them. Scores exist for 146 of the 179. Of those, 45 percent score 8 or more, 17 percent score 3 to 7, 12 percent score 1 or 2, and 26 percent score 0. Among the 12,543 people in the dataset with any licensed non-lottery gambling transaction, the same bands are 6 percent, 9 percent, 16 percent and 68 percent. The report limits that comparison to this sample. It is not a figure to set beside the Gambling Survey for Great Britain.

Credit cards are a thin slice of the detected illegal payments. Of the 3,426, 138 were on a credit card, 4 percent, made by 15 people, listed as 8 percent of the group. Current-account payments were 3,288, 96 percent, involving 172 people, so some people used both. The file sees cash movements only at institutions on open banking. A payment in cryptocurrency is outside it. So is any illegal operator whose payment text the Commission did not already have. Some of the operators that did match are not believed to be taking British customers now. Because licensed payments are much easier to recognise than illegal ones, the report says these transactions cannot be used to size the illegal market, in absolute terms or against the licensed one. Licensed payments in the wider set were labelled by a Yaspa model that called a transaction gambling once its probability passed one half. The Commission ran the work in Python and SQL, with support under its agreement with Warwick Business School. The same method cannot go looking for an operator it does not already know how to name.

Sources

  1. Gambling Commission, open-banking analysis of illegal gambling, 28 October 2026
  2. SBC News, 30 October 2026