The threshold that opens a review
A rule set that raised a review for every flag would contact most of its players; one that required a signal to repeat for weeks would be too late to be useful. The sampled policy resolves that with three separate rules - two flags inside 14 days, one high-severity flag, or a spend above four times the affordability band the account declared. Different rules, different severities, one review pipeline behind them.
- rules
- 3
- review window
- 14 days
- two flags in 14 days
- 2,800
- high-severity flag
- 720
- affordability rule
- 480
- affordability multiple
- 4.00x
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.
A review opens when a rule is met, and a sampled policy uses three rules: two flags inside 14 days, which accounts for 2,800 of 4,000 reviews; one high-severity flag such as a self-reported problem or a session over 12 hours, for 720; and a 90-day spend above four times the affordability band the account declared, for 480. The thresholds are ratios and counts, not money amounts.
Why two rules and not one
Counting flags is the simplest possible trigger and the cheapest to explain: raise a review when an account has crossed two lines inside a fortnight. On its own it is also blunt, because it treats a player who stays up late twice in two weeks the same as one who states plainly that they have a problem. So a policy usually adds a second, absolute route that does not care how many flags there are: a single high-severity signal, or a spending ratio far outside the declared band, opens a review immediately.
Sample C is the result in the sampled desk. The two-flag rule produced 2,800 reviews (70.0%), the high-severity rule 720 (18.0%), and the affordability rule 480 (12.0%). The first two are behavioural and the third is financial, and they are deliberately kept apart because the evidence a review has to cite is different in each case.
The affordability ratio, in full
The financial rule is the one an account holder most often meets without knowing why, so it is worth showing the arithmetic end to end. The account declares a band - a figure it says it can afford to lose in a month. The rule compares the trailing 90-day spend with three times that band, and reviews the account when the spend exceeds four times it. The multiples are set so that a single unusual month cannot reach the trigger and a sustained one cannot avoid it.
Worked example / sample I
- declared monthly band: 200.00
- the 90-day figure a policy compares against: 200.00 x 3 = 600.00
- deposited across nine days: 900.00
- 900.00 / 200.00 = 4.50x the monthly band
- 4.50x is above the 4.00x trigger, so the review opens
- 900.00 / 600.00 = 1.50x the 90-day figure, which is the other way the same spend can be read
| Rule | What meets it | Nature | Reviews | Share |
|---|---|---|---|---|
| Two flags in 14 days | any two signals from sample B | count | 2,800 | 70.0% |
| One high-severity flag | a self-reported problem, or a session over 12 hours | absolute | 720 | 18.0% |
| Spend above the declared band | four times the band inside 90 days | ratio | 480 | 12.0% |
Why the thresholds are not money amounts
A rule written as two hundred pounds a month would review one player for an ordinary evening and never notice another whose two hundred pounds is the whole of what they have. A ratio written against the account own declaration has the opposite property: it is almost never triggered by an amount and almost always by a change. The cost is that the declaration has to be maintained, and a band declared a year ago and never revised will quietly produce either missed cases or avoidable contacts.
- 480 reviews came from the financial rule, 12.0% of all reviews.
- 240 flags in the sample - 4.0% of all flags - were absolute rather than relative, and each one opened a review by itself.
- 720 reviews came from a single flag rather than a pair, so the record cites one signal and its severity class.
- 2,800 reviews required two flags inside 14 days, so the record has to carry both crossing dates.
A player who wants to make the rule set less noisy has one lever that is entirely in their hands: keep the declared band current. It is the only input in the whole pipeline that the account holder writes.