Canon

Solutions · Bonus-abuse defence

Don't pay bonuses to players who only deposited for the bonus.

Bonus-abuse hunters take a real slice of every operator's bonus budget. Canon detects the pattern at decision time and routes hunters to cooldown or no-action — not to a bigger nudge.

Hunter detection · live decision feed
Bots claim. Canon catches. Bonus saved.
Hunters caught
0
Allowed
0
Bonus saved
£0

Today

Detection comes after the bonus is paid.

Post-hoc detection only

Most operators detect bonus abuse in finance reconciliation, weeks after the bonus has been paid out and withdrawn.

Block-list whack-a-mole

Block one device or IP, the hunter spins up another. Account-level blocking is a treadmill that never wins.

False positives hurt real players

Aggressive abuse rules also block legitimate players with overlapping signal patterns — withdrawal-heavy, multi-account households, etc.

With Canon

Detection at decision time, not post-hoc.

Pattern detection

Canon scores every player on a hunter likelihood derived from deposit, wagering, withdrawal, and time-of-day patterns.

Decision-time routing

Suspected hunters get cooldown or no-action decisions — never a bonus. The cost of a wrong call is zero spend, not a payout.

Continuous learning

Confirmed hunters feed the model. Detection improves operator-by-operator as the signal pattern library grows.

Low false positives

Decisions are scored with confidence intervals; borderline players get held for review rather than blanket-blocked.

Outcomes
−72%
Bonuses paid to suspected hunters

The hunters were a known leak. We never expected to recover this much without false-positive complaints.

Head of Risk, Tier-1 European sportsbook · anonymised pilot
0.4%
False-positive rate, vs. operator-confirmed legitimate cohort
£11K
Median monthly bonus saved, per million treated decisions

Run hunter defence against your last quarter.

We can replay your historical event stream and report how many suspected-hunter bonuses Canon would have caught — no integration needed.