The operator's actual levers
An operator cannot investigate a participant, cannot sanction anyone, and cannot determine whether an event was manipulated. What it can do is act on its own markets and its own accounts, and those decisions are the substance of an integrity function's daily work.
Four levers exist. Which markets to offer on which competitions. Whether to suspend a market before or during an event. Whether to void bets afterwards. And what to do with the accounts involved. Each has a commercial cost, an integrity benefit, and a customer-fairness dimension, and they are not interchangeable.
Market design: the decision that does most of the work
Most integrity outcomes are determined before any bet is placed, by the decision about what to price.
Competition selection. Offering markets on a competition is a judgement that it can be supervised: that the sport has integrity resources, that the fixtures are observable, that there is enough public information to price honestly, and that the betting interest is proportionate to the sporting interest. Low-tier competitions with no broadcast coverage, no integrity education and financially distressed participants fail that test, which is a reason for an operator to decline to price them, and some regulators restrict betting on such events outright. Germany's State Treaty on Gambling, for example, prohibits bets on events in which only or mainly minors or amateurs take part, unless they are nationally or internationally significant events.
Market type selection. Result markets on a major fixture require collective manipulation to corrupt and are heavily scrutinised. Proposition and interval markets on a minor fixture require one participant and nobody is watching. The risk sits almost entirely in the second category, and offering the full proposition set on every fixture regardless of tier is a decision to accept it. Regulators draw the same line. Germany bans bets that are especially open to manipulation, in particular on occurrences a participant can bring about at will, and in February 2024 Ohio's regulator, at the NCAA's request, prohibited prop bets on individual college athletes' performance, citing among other concerns the risk of spot-fixing.
Stake limits by tier. Capping the stake accepted on lower-tier markets caps the value of manipulating them. A fixer needs the market to be liquid enough to pay; a market that will not accept a large bet is not worth corrupting.
In-play availability. Whether in-play markets are offered at all on a given competition, and with what delay. A delay between an occurrence and bet acceptance reduces the advantage of courtsiding, betting on what is seen at the venue before the data feed updates. Limiting the in-play menu is what reduces exposure to spot-fixing: Germany allows in-play betting only on the final result and on the next goal, set or similar scoring event in sports where such events are few.
The commercial objection is that these restrictions cost turnover on products customers want. That is true, and the honest response is that the turnover at issue is small relative to the operator's book (the bets caught by the college player prop restrictions Ohio adopted in 2024 were, on the regulator's estimate, about 1.35% of all sports wagers placed in the state in 2023), the liability exposure is disproportionate, and the reputational and regulatory consequence of being the book that paid out on a fixed match in a competition nobody should have been pricing is severe. Market design is where the integrity function earns its place in the commercial conversation.
Suspension
Closing a market to new bets, before or during the event.
Pre-event suspension follows an alert that survives triage. The operator stops accepting bets on the market or the fixture because it does not trust its price, which is a trading decision, and because it does not want to be the vehicle for a manipulation it suspects, which is an integrity one.
In-play suspension is the same decision compressed into seconds. In-play markets are routinely suspended, often automatically, on significant occurrences such as a goal or a penalty; integrity-driven suspension is a manual override where the pattern of money, not the pattern of play, triggers it.
The costs. Legitimate customers cannot bet. Suspicion is signalled to the market, including to whoever is behind the pattern. And a suspended market is a visible statement about a fixture that may turn out to be entirely honest.
The discipline. Suspend on the strength of the betting pattern, not on speculation about the sport. Record the trigger and the reasoning. Reopen when the reason has been examined and does not hold. And communicate it as a trading decision, because that is what it is.
A market suspended and not reported is a lost opportunity, because the suspension protected the operator and nobody else. A market reported and not suspended can be defensible where the alert is weak and the exposure small, but it needs a reason.
Voiding
Cancelling bets after the event and returning stakes, on the basis that the market was compromised.
This is the most contested lever, because it acts on customers, most of whom did nothing wrong.
When it is defensible. Where the event or the specific occurrence was subsequently found by the sport to have been manipulated. Where the operator's terms clearly reserve the right to void bets on events it reasonably believes were manipulated, and the belief is documented and reasonable. Where a regulator or integrity body has directed it. In Britain the Gambling Commission can itself void a bet it is satisfied was substantially unfair, taking account of whether a party believed the event would be run in breach of the rules or that cheating was likely, normally within six months of the result. When Ohio withdrew college player props, it ordered bets already placed on unfinished events to be voided.
When it is not. On the operator's own unsupported conclusion about the sport. On suspicion alone, applied selectively to winning bets while losing bets are settled. As a mechanism for avoiding a liability the operator simply did not like.
The customer-fairness test applies with full force. A void applied only to winners is a confiscation dressed as an integrity measure, and it is hard to defend: in Britain an operator's terms must not be unfair within the meaning of the Consumer Rights Act 2015, and a customer can take an unresolved complaint to a free alternative dispute resolution service. A void applied to a whole market, winners and losers alike, on the basis of a sporting body's finding, is a different thing entirely.
Settlement holds are the intermediate step and are frequently the right one. Holding settlement of specific bets pending the outcome of an integrity review, with a stated timeframe, protects the operator without pre-judging the customer. The customer is told the bet is under review, not that it is void, and the review reaches a conclusion within the stated period.
Account action
What to do with the accounts behind a suspicious pattern.
Nothing, yet. Where the pattern is explained or weak. Recorded, watched, not acted on.
Stake restriction. A commercial risk decision the operator can take on its own assessment. It does not require a finding about the sport, and it should not be communicated as resting on one.
Settlement hold. As above, on the specific bets.
Closure. Where the operator's terms permit closing on suspected breach of integrity provisions, or where the commercial relationship is untenable. Again a decision the operator owns.
Referral. To the sport or the regulator, with the account-level detail the framework permits. In Britain betting licensees must pass information suggesting an offence such as cheating to the Gambling Commission, and information about breaches of a sport's betting rules to that sport's governing body where the body is one listed in the Gambling Act 2005. This is where the operator's information leaves its control, and the record of what was shared, with whom and on what basis, matters.
The constraint across all of these is tipping off. An account restricted after a report has been made, with an explanation that discloses the report, has prejudiced whatever follows. The wording has to be prepared, accurate and minimal.
Participant betting
Separate from suspicious patterns, and simpler in principle: participants, officials and insiders are generally prohibited from betting on their own sport (tennis, for example, bars every player, official and other covered person from betting on any tennis competition), and some regulators give operators the means to enforce it: Germany's State Treaty bars anyone who can directly or indirectly influence an event from betting on it, and allows operators and event organisers to exchange participants' names and dates of birth to police the ban.
The identification problem is real. An operator knows its customers' names. It does not automatically know which of them referee lower-league fixtures, play semi-professional tennis or work in a club's medical department. Screening against lists of participants exists in some sports and markets and is patchy elsewhere.
The controls that exist. Declared occupation at registration, where required. Screening against available participant lists. Detection of betting on a competition by an account whose other data suggests a connection. And reporting to the sport where a participant is identified.
Proportionality. A junior official betting a small amount on their own competition is a disciplinary matter for the sport and a reportable event for the operator. It is not a fixing case, and treating it as one wastes everyone's time.
The accumulator problem
A specific and under-discussed exposure. Manipulation of one event can be monetised through accumulators that combine it with legitimate selections, which spreads the money across many bets, disguises the concentration, and raises the payout.
Detection has to consider whether a specific event appears with unusual frequency across multi-leg bets, not only whether it attracts unusual single-bet volume. Voiding decisions have to consider whether the compromised leg is voided and the accumulator resettled, which operators' betting rules usually provide for and customers often overlook.
Deciding, and being able to explain it
The decision framework that survives review has a small number of components.
The strength of the pattern, assessed against the baselines and the innocent explanations. The exposure, in liability and in reputational terms. The reversibility of the action: a settlement hold is reversible, a void is not. The customer-fairness consequence, particularly whether the action falls only on winners. The reporting position, because market action and reporting should be decided together. And the record.
An integrity function that can show, for each market decision, that it weighed those six and recorded the reasoning is in a defensible position whatever the outcome. One that suspended because a trader had a feeling, or voided because the payout was large, is not.