fxgnosis

Broker Rating Methodology: 7 Factors and Weights

Last updated: July 18, 2026

Each broker in the fxgnosis rating receives a score from 0 to 5 across seven independent factors. The final position is determined by a weighted average of these factors.

Five factors rest on objective data (Regulation, Trading Conditions, Payments, Reviews, Localization): licences, fees, the set of payment methods and platforms, public ratings, verifiable localization signals. Two factors, Reputation and Support, include an editorial assessment, described below. The methodology is the same for all brokers regardless of any affiliate relationship; any update is dated below and applied to the entire rating simultaneously.

The 0 to 5 rating scale

The score is an absolute rating on a strict scale, not a broker's position relative to others in the list. A five means a broker with no weak spots on any factor: a set of Tier-1 licences, service to clients from South Africa under a top-tier licence, best-in-class trading conditions and a spotless reputation. South Africa has a real local regulator (the FSCA), so serving local clients under a recognised licence is achievable, and the strongest brokers here reach the Exceptional band. A flawless five, though, stays out of reach: it would need Tier-1 regulation for the local entity too, and South Africa's own regulator sits at Tier-2.

012345
  • 4.0–5.0 Exceptional. A broad set of top-tier group licences together with a local FSCA licence for clients from South Africa, and strong results across the board. Within reach of the strongest brokers here, because the local FSCA licence lifts the per-geo side of regulation.
  • 3.0–4.0 A strong, reliable broker for the South African market, built on a solid set of international licences. What keeps the score in this band rather than higher varies: a broker may serve South African clients through an offshore entity instead of a local FSCA licence, which weakens the per-geo half of regulation, or it may hold the local FSCA licence yet be held back by weaker reputation, support or localization. Sound on the fundamentals, short of the maximum on one or more factors.
  • 2.0–3.0 A workable option with more serious limitations in the fundamentals themselves: a modest set of international licences, average trading conditions, or little data on reviews and support. Here it is not only about which licence the broker operates under in the South African market, but about weaker baseline indicators.
  • 1.0–2.0 A weak result with serious gaps on key factors.
  • 0.0–1.0 Fails the basic criteria. Brokers with confirmed systematic problems do not make the rating at all.

A broker that serves South African clients under a local FSCA licence and is strong across the board reaches into the 4s, the Exceptional band; one on an offshore licence, or with a weaker reputation or support, settles in the 3–4 range. A flawless five stays out of reach: it would need Tier-1 regulation for the local entity, and the FSCA is a Tier-2 regulator. These scores also run lower than the usual 4.8–4.9 on review sites that grade on a lenient scale, where almost every broker sits in the top band; here a five is a global benchmark, not an average. A low score reflects real limits on verifiable factors, not editorial bias.

Seven rating factors

Regulation (weight 25%)

Composed of two parts.

Group-level (0–3 points): the broker's full licence portfolio.

  • +1.5 if any Tier-1 licence is present (FCA, ASIC, BaFin, FINMA, MAS, SEC, CFTC)
  • +0.9 if two or more Tier-1 licences
  • +0.3 per Tier-2 licence (CySEC, DFSA, FSCA), max 0.6
  • +0.15 per Tier-3 licence (FSA Seychelles, IFSC, FSC Mauritius, MISA, SCB, VFSC), max 0.3

Per-geo (0–2 points): the licence under which the broker actually serves local clients. Tier-1 → 2, Tier-2 → 1, Tier-3 → 0.5, no licence → 0.

Final score: sum of both parts, capped at 5. A broker with no licences legitimately gets 0 (a meaningful result, not missing data), so this factor always participates in the calculation.

Reputation (weight 18%)

Composed of two parts.

Age bonus (0–1) is calculated automatically: 20+ years → 1.0, 10 years → 0.5, 5 years → 0.25, under 3 years → close to 0.

Editorial score (0–4) is assigned based on monitoring of independent platforms (ForexPeaceArmy, BrokerScam, industry forums), regulatory actions, and fines:

  • 4 – Spotless: no regulatory penalties, no mass complaints about non-payments, transparent ownership
  • 3 – Good: solid brand, occasional minor complaints
  • 2 – Average: isolated complaints or minor past fines
  • 1 – Questionable: notable withdrawal complaints, accumulated regulatory orders
  • 0 – Poor: systematic complaints, major fines, documented incidents

Without an editorial score, the factor is dropped (excluded with weight redistribution).

Brokers with confirmed systematic complaints about withdrawal refusals or price manipulation are excluded from the fxgnosis rating regardless of other factors.

Trading conditions (weight 22%)

Weighted sum of eight sub-factors; a sub-factor with no data is dropped and its weight is redistributed to the others:

  • 15% Minimum deposit: ≤$100 → 5, ≤$200 → 4, ≤$500 → 3, ≤$1000 → 2, higher → 1
  • 30% Trading cost (best account): the cheapest account is used, i.e. the better score of the spread and the commission, not their sum or average. Each on a pip ladder (commission = commission/10): ≤0.3 → 5, ≤0.6 → 4, ≤1.0 → 3, ≤1.5 → 2, higher → 1
  • 10% Number of instruments: <100 → 1, 100–300 → 2, 300–700 → 3, 700–1500 → 4, >1500 → 5
  • 15% Platforms: +1 for MT4, +1 for MT5, +1 for cTrader, +0.5 for TradingView, +0.5 for TradeLocker, +1 for a proprietary platform
  • 3% Swap-free: available → 5, no → 0, unknown → not counted
  • 7% Copy trading: available → 5, no → 0, unknown → not counted
  • 10% Adjustable leverage: the trader can set the leverage in account settings (useful for risk management): available → 5, no → 0, unknown → not counted. What counts is the ability to adjust, not the size of the maximum leverage.
  • 10% Negative balance protection: the account cannot go below zero, so a trader never owes more than was deposited: available → 5, no → 0, unknown → not counted.

Payment methods (weight 14%)

The score depends on which tiers of payment methods are present and which direction they support, not on the number of listed methods. Within a single tier, deposit and withdrawal are combined as “at least one method,” so five cards of the same tier give no advantage over one.

Each present tier contributes:

contribution = tier weight × (0.4 for deposit + 0.6 for withdrawal)

Withdrawal carries more weight: being able to take money out matters more to a trader. Tier weights: Tier 2 = 3.5 (base), Tier 3 = 1.6 (premium), Tier 1 = 0.9 (weak reserve). The sum of contributions is capped at 5.

Payment method tiers for South Africa:

  • Tier 2 (base) – international cards and stablecoins: Visa, Mastercard, USDT (TRC20). Available at almost every broker and work everywhere, but require conversion to rand.
  • Tier 3 (premium) – local South African methods: instant EFT via Ozow or PayFast, Capitec Pay, and local bank EFT (Capitec, FNB, Standard Bank, ABSA, Nedbank). No USD↔ZAR conversion, no bank fees, usually instant. Maximum value for a South African trader, so this is awarded on top of the base.
  • Tier 1 (reserve) – methods requiring extra steps: SWIFT (slow and expensive), e-wallets Skrill/Neteller (separate KYC account), Bitcoin, Perfect Money. They work, but offer less convenience.

There is no separate penalty for missing local methods: Tier-3 is rewarded with its own weight, and its absence simply skips the premium.

Customer reviews (weight 9%)

The Trustpilot rating is used, on a 0 to 5 scale.

TradingView is shown on the broker page for reference but is excluded from the score: the platform integrates brokers as paying partners, so its rating is not an independent signal. If a broker has no Trustpilot rating, the factor is dropped and its weight is redistributed across the others.

Customer support (weight 5%)

A score from 0 to 5 that assesses the broker's customer-support channels and their availability: which channels are offered (live chat, email, phone, messengers), the hours and languages they cover, and how promptly and substantively support replies. It is a qualitative assessment, not a measured benchmark, so the factor carries a low weight.

  • 5 – several channels including live chat, broad hours, prompt and relevant replies
  • 3–4 – workable channels, but narrower hours, slower or more scripted
  • 1–2 – limited channels, email-only or replies that take days
  • 0 – no working channel, or autoreplies only

Without an assessment, the factor is dropped and its weight is redistributed.

Localization for the South African market (weight 7%)

This factor measures only the localization that is not already captured by other factors: local payment methods are part of Payments, support is part of Support, and the local regulator is part of the per-geo half of Regulation. What remains is a residual of four verifiable signals (weighted sum; a sub-signal with no data is dropped):

  • 35% Localized website: a South Africa-focused site or section, not merely the default international English site
  • 30% ZAR account: South African rand as the trading account currency
  • 20% Localized education: learning materials aimed at South African traders
  • 15% Local contact: local South African phone number or office in the country

Each signal: present → 5, absent → 0, unknown → not counted. If no signal is set at all, the factor is dropped and its 7% weight is redistributed.

How missing data is handled

Missing data on a factor must not silently turn into either a penalty or a reward. If there is no data on a factor or a sub-signal, it does not get 0 and does not get 5: it is dropped from the calculation, and its weight is proportionally redistributed across the factors that do have data. The final score is therefore always based only on verifiable information. For example, if a broker's spread is not confirmed, the spread sub-factor does not pull the score down to a token value; it simply does not participate.

Data sources

Licences and licence numbers are verified directly against public regulator registers (FCA Register, CySEC public list, FSC Mauritius licensees, etc.). Trading conditions come from brokers' official websites and live client cabinets at the time of verification. The review rating is taken from the public Trustpilot listing, unmodified; the TradingView rating is shown for reference but is not scored. Localization signals (site and education language, account currency, local contact) are checked against the website and the client cabinet. Editorial scores for reputation and support are assigned by fxgnosis: reputation from monitoring independent platforms and regulatory records, support from the broker's customer-support channels and their availability.

Rating updates

A full recalculation runs automatically on the last day of each month: all seven factor values are recomputed and positions in the ranking are refreshed. Off-cycle recalculations are triggered by urgent events (licence suspension by a regulator, mass confirmed complaints); such cases are noted in the broker's card.