Every score has a reason. A risk result is never shown on its own. It always comes with the factors, the comparison with the player's usual behaviour and links to the underlying activity.

How the assessment works

VILIKAN evaluates each player through a transparent, step-by-step process:

Activity. Gameplay, payments, sessions, limits and responsible gambling events are received from the operator's platform.
Behavioural indicators. Activity is summarised into indicators over several time periods, from the last few minutes to the player's full history.
Risk factors. Indicators are grouped into thirteen behavioural risk factors, each assessed individually.
Risk Score. Factor results are combined into a single Risk Score on a scale from 0 to 100.
Risk level. The Risk Score is assigned to one of five risk levels.
Explanation and follow-up. The result is stored together with its reasons and, where relevant, leads to an alert or an action.

The model is deterministic and fully transparent: the same activity and the same configuration always produce the same result. All weights and thresholds are part of the configuration managed by the administrator, not built into the product.

Risk levels

Every player with sufficient activity is assigned one of five risk levels. The boundaries between levels are defined in the configuration and can be adjusted by the administrator.

Level	Meaning
Very Low	No meaningful signs of risky behaviour.
Low	Isolated signals that do not form a pattern.
Medium	Emerging patterns that deserve closer attention.
High	Clear patterns of risky behaviour that call for review.
Critical	Strong indicators of harm that call for priority review.
No data	Not enough activity yet to produce a reliable assessment.

Each level is shown with a consistent colour across the product, always accompanied by the level name and score.

Risk factors

The Risk Score is built from thirteen behavioural risk factors. Each factor looks at a different aspect of the player's behaviour and receives its own result.

Factor	What it looks at
Increasing Tolerance	Stakes and spending growing compared with the player's usual behaviour
Repeated Loss of Winnings	Winnings being played back instead of kept
Binge Gambling	Long, intense periods of play without meaningful breaks
Loss Chasing	Attempts to win back losses shortly after they occur
Fluctuating Wagers	Unstable or erratic stake sizes
Impulsive Transactions	Series of quick, repeated financial actions
Cancelled Withdrawals	Withdrawals being reversed and returned to play
Payment Methods	Unusual patterns in the use of payment methods
Monetary Consumption	Overall spending and how it changes over time
Nightly Play	Activity during the player's night-time hours
Number of Days Played	How often the player gambles and for how many days in a row
Time Consumption	Time spent gambling
Game Types	A shift towards faster or higher-intensity game formats

What a factor result includes

For each factor, VILIKAN provides:

a factor score on a scale from 0 to 100;
a severity label: Normal, Elevated, High or Critical;
the main contributing indicators behind the result;
a comparison with the player's own baseline;
links to the underlying activity, such as specific bets, deposits or sessions;
the configuration version used for the calculation;
a data confidence indicator.

If a factor cannot be fully assessed because some data is unavailable, it is clearly marked with the reason and a reduced confidence level rather than silently omitted.

Personal baseline

Behaviour that is normal for one player may be a warning sign for another. VILIKAN therefore compares each player first and foremost with their own history. For every indicator, the product considers both the current value and how far it deviates from the player's usual behaviour, as well as the direction of the trend.

Time periods

Indicators are evaluated over several groups of time periods so that both sudden changes and gradual developments are captured:

Period group	Purpose
Acute	Rapid changes within minutes or the current session
Short-term	Episodes over the last day or few days
Medium-term	Behavioural trends over recent weeks
Long-term	The player's personal baseline over months and full history

All times are stored in a single standard time zone. Night-time behaviour is evaluated in the player's local time.

Acute rules

Some situations indicate immediate risk and should not wait for a gradual change in the overall score. Acute rules detect such strong indicators and can raise the risk level directly. They act as a safety mechanism on top of the weighted factor model.

Acute rules cover situations such as:

a rapid increase in deposits;
loss chasing shortly after a significant loss;
an unusually long session without breaks;
intense play during night-time hours;
a cancelled withdrawal followed by a return to play;
a series of declined payments followed by a successful one;
a sharp increase in the overall Risk Score.

The conditions of each acute rule are configurable. When an acute rule is triggered, a risk event is created promptly and, depending on the configuration, leads to an alert or a higher priority in the queue.

Explanation

Every risk result is delivered with an explanation that is available wherever the score is shown, including the player profile, alerts, cases, evidence exports and the API. An explanation includes:

the Risk Score, risk level and recent change;
the main reasons, each with its factor, contribution and supporting activity;
the comparison with the player's baseline;
the configuration version and the snapshot of indicators used.

For players with elevated risk, several reasons are always provided, each linked to the activity that supports it.

When the risk is recalculated

The Risk Score is kept up to date in three ways:

On activity – when relevant events occur, such as a settled bet, a deposit, a cancelled withdrawal or the end of a session.
On schedule – a regular full recalculation for all active players.
On request – a recalculation for a selected period, for example after a configuration change.

Each recalculation creates a new record. Previous results are never overwritten and always remain linked to the configuration that produced them, so any past decision can be reproduced exactly as it was made.

Related pages