The duty to step in
Some gambling decisions are not the player own. A licence usually obliges an operator to watch for signs that play has stopped being a choice, to act on what it finds, and to write down what it did. This desk takes that obligation apart: the five signals a protection rule is written around, the threshold that opens a review, the contact the operator has to make, the limits and breaks it can impose for protection rather than for its own commercial reasons, the record it keeps, and what the player is owed when it gets this wrong.
- accounts sampled
- 40,000
- met a review trigger
- 10.0%
- reviewed
- 4,000
- imposed a limit or break
- 640
- restricted or closed
- 320
- runtime javascript
- none
Read the signals
Five comparisons the account already supports, run continuously rather than after a complaint. 6,000 flags in the sample, 96.0% of them relative to the account own history.
Choose a response
Nothing, a message, a limit or a break, or a restriction - one of four, graded by how much it takes from the player. 960 of 4,000 reviews in the sample changed access.
Write it down
Signals, crossing dates, rule version, path, response and outcome - one record per review, including the 1,600 that ended with nothing done.
Take the appeal
A route outside the desk that decided. 480 appeals in the sample, 176 of them upheld, which is 36.7% - high enough that the route does real work.
An operator protection duty is a licence obligation, not a courtesy: the operator must monitor play for signs of harm, act on what it finds, and keep a record of the decision. The signals are read from the account rather than from the person, so a review begins when a rule is met. The operator can then do nothing, send a message, impose a limit or a break, or restrict the account - and it owes the player notice, a reason and a route to appeal.
40,000 active accounts at one desk over 90 days. Of them, 36,000 (90.0%) met no protective review at all and 4,000 (10.0%) met a trigger. The desk is about the 4,000: who watched, what they read, what the desk was required to do next, and what the record shows afterwards.
- accounts sampled
- 40,000
- window
- 90 days
- met no review
- 36,000
- met a trigger
- 4,000
- share reviewed
- 10.0%
The five families of signal a protection rule is written around, and the flags they raised across the 40,000 accounts: a session longer than six hours 1,480; a single session staked at three times the account own 90-day mean 1,120; a deposit made within ten minutes of a loss 2,640; a request to raise a deposit limit shortly after a loss 520; and a self-reported problem flag 240. That is 6,000 flags, and 1.5 flags for every account that raised any.
- flags raised
- 6,000
- long session
- 1,480
- spend above the mean
- 1,120
- deposit after a loss
- 2,640
- limit-lift after a loss
- 520
- self-reported problem
- 240
- flags per flagged account
- 1.5
A review starts when a rule is met, not when a person becomes worried. In the sampled policy three rules do it: two flags inside 14 days (2,800 accounts, 70.0%), one high-severity flag such as a self-reported problem or a session over 12 hours (720, 18.0%), or a 90-day spend above four times the affordability band the account itself declared (480, 12.0%).
- accounts reviewed
- 4,000
- two flags in 14 days
- 2,800
- share, two-flag rule
- 70.0%
- a high-severity flag
- 720
- share, high severity
- 18.0%
- above the affordability band
- 480
- share, affordability
- 12.0%
How the 4,000 reviews ended: 1,600 (40.0%) with no action, the pattern having been explained; 1,440 (36.0%) with a safer-play message and no change to the account; 640 (16.0%) with a limit imposed or a break placed; and 320 (8.0%) with the account restricted pending documents, or closed for protection. The first two are the majority, which is the point of a rule written around signals rather than around outcomes.
- reviews
- 4,000
- no action
- 1,600
- share, no action
- 40.0%
- message only
- 1,440
- a limit or a break
- 640
- restricted or closed
- 320
- accounts changed
- 960
2,400 contacts asked the player to respond or changed their access. A response arrived within seven days on 1,680 (70.0%); of those, 480 (28.6%) had the action lifted or reduced and 1,200 (71.4%) had it confirmed. On the 720 (30.0%) that never answered, the action stood in every case. A duty that ends in silence is still a decision.
- contacts
- 2,400
- response within 7 days
- 1,680
- response share
- 70.0%
- lifted or reduced
- 480
- confirmed
- 1,200
- no response
- 720
- actions standing, no answer
- 720
Of the 4,000 reviewed accounts, 480 (12.0%) were placed on a mandatory break: 320 for 24 hours, 112 for seven days and 48 for 30 days. The sampled policy allows a break to be placed without the player consent, and none of the 480 was shortened before it expired. A break placed for protection is recorded on the account for 12 months in 96 of the 480 cases (20.0%).
- accounts reviewed
- 4,000
- placed on a break
- 480
- share on a break
- 12.0%
- 24 hours
- 320
- seven days
- 112
- 30 days
- 48
- shortened before expiry
- 0
- recorded for 12 months
- 96
The 1,600 reviews that ended with no action are the ones the rule got wrong. Of them, 1,280 (80.0%) were explained by a one-off event - a holiday, a bonus, a single large deposit; 224 (14.0%) by a data error, such as one device counted as two or a converted currency figure read as a spend; and 96 (6.0%) were never explained and the account was left unchanged anyway. Every one of the 1,600 cost a contact, and 640 of them had already produced a limit or a break before the review closed.
- no-action reviews
- 1,600
- a one-off event
- 1,280
- share, one-off
- 80.0%
- a data error
- 224
- share, data error
- 14.0%
- never explained
- 96
- share, unexplained
- 6.0%
The failure in the other direction. A retrospective audit of the 36,000 accounts that met no review found 360 (1.0%) that should have met a trigger and did not: 288 (80.0%) because the flag was raised after the review window had closed, and 72 (20.0%) because the pattern was split across two accounts and neither alone reached the rule. A rule can be too eager and too quiet at once.
- accounts not reviewed
- 36,000
- should have been reviewed
- 360
- share missed
- 1.0%
- window closed first
- 288
- split across accounts
- 72
One account, one review. A declared affordability band of 200.00 a month. Fourteen deposits over nine days totalling 900.00, which is 4.50 times the band; a single session on day seven that ran 7.2 hours; and a deposit made 4 minutes after a 60.00 loss on day eight. The spend crossed the affordability rule on day nine, the review opened the same day, a seven-day break and a message were placed on day ten, the player replied on day eleven and the break was confirmed.
- declared monthly band
- 200.00
- deposits in nine days
- 14
- deposited
- 900.00
- times the band
- 4.50x
- longest session
- 7.2 h
- deposit after a loss
- 4 min
- trigger to review
- 0 days
- trigger to break
- 1 day
- break length
- 7 days
- answer
- day 11
A quarter at the desk: 40,000 accounts placed 9,600,000 rounds and staked 48,000,000.00, a mean of 240 rounds and 1,200.00 each. The 4,000 reviewed accounts were 10.0% of the accounts, 25.0% of the rounds and 30.0% of the money. Reviews 4,000; contacts 2,400; breaks 480; restrictions and closures 320; no action 1,600. Of the 4,000 reviewed, 3,424 (85.6%) were still playing 90 days later and 576 (14.4%) had closed the account themselves.
- accounts
- 40,000
- rounds placed
- 9,600,000
- staked
- 48,000,000.00
- mean stake per account
- 1,200.00
- reviewed accounts, share of money
- 30.0%
- reviews
- 4,000
- contacts
- 2,400
- still playing at 90 days
- 3,424
- closed their own account
- 576
The review is about a pattern, and a pattern is not a person
When a player is contacted about their play, the natural reading is that somebody was watching them. That reading is almost never right. A protection rule reads numbers that the account already produces - how long a session ran, how a spend compares with the account own history, whether a deposit followed a loss - and adds them up. The rule does not know who the player is, what the money means to them, or why this week looked different from last week. It knows that a threshold was crossed.
That is the whole reason this duty is worth understanding before it touches you. A review is not an accusation, and a limit is not a punishment: both are the output of a rule that was written in advance, applied to an account, and - in a well-run desk - recorded with its reasoning. The rest of this desk is about what those rules contain, what the operator is allowed to do with the result, and what you are owed while it happens.
Every figure on this site derives from ten invented samples defined in the cards below. Nothing here is observed from a live account, and no operator, regulator, charity or person is named. Every arithmetic step is shown so it can be re-derived.
The arithmetic in one line, and then in a paragraph
Worked example / sample I
- the affordability band the account declared: 200.00 a month
- deposits over nine days: fourteen, totalling 900.00
- 900.00 against a 200.00 monthly band: 4.50x
- the policy trigger: spending above four times the declared band
- 4.50x clears 4.00x, so the review opens
- the decision: a seven-day break plus a message, one day later
- the player reply arrives on day eleven, and the break is confirmed
Read the last line carefully. The rule that reviewed this account had no view on whether the spending was a problem. It had a band the account itself had declared and a spend figure, and those two numbers produced a ratio. Understanding that is what turns a confusing message into a legible one - and it is also what makes the false-positive problem, covered on this desk, so easy to predict.
What the desk covers, in the order the questions arrive
| Question | Where it is answered | Figure |
|---|---|---|
| What is the operator actually obliged to do? | what the duty is | 3 duties |
| What does it read to decide? | the pattern it reads | 5 signals |
| What makes it open a review? | the threshold | 4.00x |
| What can it do to the account? | what an intervention can do | 4 outcomes |
| What am I owed if it is wrong? | what you are owed | 3 rights |
The order matters. A reader who starts by asking why they were restricted arrives at the thresholds already knowing who set them and what the operator is required to do with them - which is the difference between a policy and an arbitrary decision.
How to use this desk
- Read the answer block at the top of a page first; it is written to be quoted and it contains no arithmetic.
- Read the worked example second; every number in it comes from the ten samples and every step is shown.
- Read the outcome you actually received before the ones you did not. A message, a limit and a restriction are three different decisions with three different routes to appeal.
- Treat the figures as illustrative throughout. They describe a sampled policy, not any real licence condition, and they are labelled as invented everywhere they appear.
- If you are reading this because of your own play rather than curiosity, the support routes on the legal page matter more than any figure here.