How the Safety Score works
fi-1.0 is an editorial evidence rubric, not a probability of safety, returns or solvency. Evidence-based rating; not a guarantee.
One entity, one context
A score applies to a named legal entity, residence country, retail forex/CFD product and effective date. A brand has no global score. Without a selected scope, inspect evidence and entity choices.
R1 publishes Safety only. Overall ratings, rankings and measured testing protocols belong to later releases.
Fixed dimensions
Dimensions run from 0 to 100. A documented negative is zero. Unavailable, stale, contradictory or disputed evidence is unknown. Every required subcheck must be resolved for a dimension to be known.
Verified regulatory permission · 40%
50 active applicable retail-product permission + 25 exact entity/reference match + 25 territorial permission; explicitly inactive/absent permission makes the dimension zero. Every required subcheck must be resolved.
Applicable client protections · 15%
25 each: applicable segregation, negative-balance protection, external dispute resolution, compensation eligibility. Explicit non-applicability scores zero for that subcheck; missing evidence is unknown.
Entity operating history · 10%
Completed entity incorporation years: <1=0, 1–2=25, 3–4=50, 5–9=75, >=10=100. Never use brand age.
Public documentation · 10%
20 each: current entity/contact identification, client agreement, fee schedule, risk disclosure, execution/conflicts policy. Verified absence is zero; access failure is unknown.
Attributed warning history · 10%
Active directly attributed permission/client-funds/misconduct notice=0; explicitly resolved/withdrawn notice issued within five years=50; otherwise successful scoped official searches with no matching notice in that window=100. Clone victims never reduce points.
Verified operating disclosures · 10%
25 each: current official domain, contact channel, matching entity in account terms, documented complaints procedure. Availability only, not quality or solvency.
Complaint evidence · 5%
Official adjudications/enforcement concerning the scoped entity within 24 months: unresolved adverse=0; explicitly remedied=50; successful official search with no finding=100. Unavailable sources and visitor allegations cannot establish points.
Coverage and uncertainty
Coverage C is the sum of known dimension weights. A number requires at least 90% coverage and known verified regulatory permission, applicable client protections, attributed warning history.
Let S = sum(weight × dimension points). Raw score = S / C. Uncertainty bounds = [S / 100, (S + 100 × (100 − C)) / 100]. Missing dimensions range from 0 to 100 in these bounds; unknown is never scored as zero.
Display rounding is half-up to one decimal; ordering uses unrounded values. Insufficient coverage shows “Not enough verified data”. Scores retain input revision IDs, scope, calculation time, methodology version and evidence expiry.
Caps and evidence changes
Explicitly inactive or absent applicable permission, or an active directly attributed official warning concerning permission, client funds or misconduct, caps the score and both bounds at 25. The entity is excluded from recommendations. Unknown critical evidence suppresses the score.
A clone notice about another host does not penalize the legitimate firm. Visitor allegations do not score directly or automatically hide sourced warnings. Owner-triaged fields under review and dependent ratings are suppressed; closure requires current confirmed or replacement evidence and recalculation.
No regulator tiers, colour bands or “safe” labels are inferred. Correcting evidence invalidates dependent scores immediately.
Worked examples
These are hypothetical arithmetic examples, not broker results or measurements.
Every dimension is 80
Coverage: 100%. Score: 80.0. Bounds: 80.0–80.0.
Complaint evidence unknown; other dimensions are 80
Coverage: 95%. Score: 80.0. Bounds: 76.0–81.0.
Active attributed warning; other dimensions are 100
Coverage: 100%. Score: 25.0. Bounds: 25.0–25.0. Capped at 25; excluded from recommendations.
Regulatory permission unknown; other dimensions are 80
Coverage: 60%. Score: Not enough verified data.