Source firstEvery case should be traceable to a public, checkable source.
Context mattersEvidence, severity and operator response affect the weight.
Outcome mattersResolved, rejected and open cases are not treated equally.
1
Source
Every published case must be traceable.
2
Context
Severity and available evidence change the weight.
3
Outcome
Resolved, rejected and unresolved cases are not equal.

Why we don't just count complaints

Counting complaints is easy. Understanding what they actually say about an operator is harder.

A large sportsbook that has operated for years may naturally accumulate more complaints than a recently launched operator with a much smaller player base. A raw complaint count therefore tells us very little on its own.

IceCasino is a good example. At the time of our research, Casino Guru indexed 556 complaints associated with the operator: 314 rejected, 237 resolved and 5 open.

Looking only at "556 complaints" produces a very different impression from looking at the outcomes behind that number.

That is why we don't treat every complaint as an equal negative signal. A complaint does not automatically reduce a Trust Score — its impact depends on what happened, how well the claim is documented, the amount and severity involved, whether the operator responded, and how the dispute ultimately ended.

A resolved withdrawal delay should not carry the same weight as an unresolved five-figure winnings confiscation.

How we classify complaint evidence

Every complaint we publish must be traceable to a public source. But not every source provides the same level of evidence. We broadly separate complaint evidence into three categories.

1. Documented individual cases

These are specific disputes with identifiable details such as the disputed amount, the operator's response, supporting information or a recorded outcome.

Complaint platforms such as Casino Guru and AskGamblers can provide particularly useful documentation because the correspondence between the player, operator and mediator is often preserved and the final outcome is recorded. However, a complaint appearing on one of these platforms does not automatically mean the player's allegations were proven — we record the outcome separately.

2. Public individual reports

We also monitor sources such as Bitcointalk, Trustpilot and Reddit. These reports can reveal issues that never reach a formal complaint platform, including sportsbook-specific disputes involving account limits, arbitrage allegations, value betting, KYC or confiscated winnings.

But a public allegation is still an allegation. Where possible, we look for additional evidence such as screenshots, transaction details, betting history, an operator response or confirmation from another source. Cases with limited supporting evidence are identified accordingly rather than presented as established facts.

3. Aggregate complaint patterns

Sometimes the important signal is not one complaint but a repeated pattern. An operator may have dozens of reports describing essentially the same issue: withdrawals remaining pending, repeated KYC requests, accounts being restricted after winning, or the same generic explanation being given to multiple players.

Adding every near-identical report as a separate complaint can artificially inflate the apparent size of the problem. In those situations, we may record an aggregate pattern instead — the underlying source remains linked, and where available we preserve the breakdown between open, resolved and rejected complaints.

3
Evidence categories
Every complaint
Traceable to a public source
Separate
Allegation from outcome

What matters more than the complaint count

When assessing a complaint, we look at several factors together:

  • Source quality — where did the complaint originate and what evidence is available?
  • Severity — was this a minor delay or a substantial balance confiscation?
  • Amount involved — larger financial disputes generally deserve greater scrutiny.
  • Operator response — did the operator explain its decision or simply provide a generic response?
  • Evidence provided — were specific transactions, bets or rule violations identified?
  • Outcome — was the player paid, was the complaint rejected, or does the dispute remain unresolved?
  • Pattern — are similar complaints appearing repeatedly?

This prevents very different situations from being treated as equivalent. For example, an operator that receives many complaints but consistently resolves legitimate withdrawal issues should not necessarily be assessed more negatively than an operator with only a handful of complaints where substantial balances remain confiscated without a clear explanation.

Case study: GG.BET withdrawal delays

GG.BET provides a useful example of why we look beyond the headline complaint count.

Our initial research identified six sportsbook-specific complaints on AskGamblers involving withdrawal delays. Most were eventually resolved, but the cases showed a recurring pattern of withdrawals taking considerably longer than players expected, sometimes accompanied by explanations referring to technical or payment-processing issues.

A second research pass identified 24 additional complaints on the casino side of the same AskGamblers profile.

What we found

24 further complaints, same pattern

Twenty-two were listed as resolved and two as rejected. Rather than adding 24 similar entries and presenting them as 24 separate negative signals, we recorded the broader withdrawal-delay pattern and preserved the outcome breakdown. That tells us more than the raw number. See it on GG.BET's profile.

The relevant finding is not simply "GG.BET has many complaints." It is:

"GG.BET has a documented pattern of withdrawal-delay complaints, but a large majority of the cases reviewed were ultimately resolved."

Those are materially different conclusions.

What we don't publish

We do not publish every accusation we find. Claims without a traceable public source are excluded. Reports that contain too little information to understand what allegedly happened may also be excluded or kept out of the scoring process until additional evidence becomes available.

We are particularly careful with serious allegations involving confiscated funds, fraud, multi-accounting, arbitrage or sportsbook abuse. If the available evidence only establishes that a player made an allegation, that is how we describe it. We do not turn an allegation into a finding simply because it appears online.

How complaints affect the Trust Score

There is no automatic rule that one complaint equals a fixed deduction. A complaint's impact depends on its severity, evidence, outcome and broader context.

A €500 withdrawal that was delayed and subsequently paid is fundamentally different from an unresolved €50,000 balance confiscation.

Likewise, a complaint rejected after the operator provides convincing evidence of multi-accounting should not be treated the same as a case where an operator confiscates winnings but provides neither the player nor the mediator with a meaningful explanation.

Patterns also matter. One isolated withdrawal delay may tell us very little. Twenty similar reports over a relatively short period may indicate an operational issue even if many are eventually resolved. This is why we evaluate what happened and how the operator handled it, rather than simply displaying a complaint counter.

Transparency matters

Complaint research is imperfect. We don't have access to operators' internal risk systems, sportsbook-provider data or private KYC investigations. In many disputes, neither side makes all evidence public.

Our objective therefore isn't to declare a winner whenever two sides disagree. It is to document what can actually be established:

  • What did the player allege?
  • What did the operator say?
  • What evidence was made public?
  • What was the outcome?

That distinction is central to how we evaluate sportsbook complaints — and why 500 complaints do not automatically make an operator worse than five.

See the full scoring methodology, including how complaint records feed into the final Trust Score.

How scoring works →