ระเบียบวิธี
| แกน | น้ำหนัก | Basis |
|---|---|---|
| reviews | 25 % | reviews |
| cost | 20 % | round_turn_cost_usd |
| safety | 20 % | measured |
| payouts | 15 % | withdrawal_hours |
| track_record | 10 % | year_founded |
| platforms | 10 % | measured |
What each axis measures
What happens when a figure is missing
An axis with no data is dropped and the remaining weights are renormalised. Below 55% of the weight no score is published at all: a thin score is not a low score, and publishing one would rank a broker on its best axes.
A broker enters a ranking only when it has both a review rating and a round-turn price. Others appear in the table with the score column blank — which is why blanks are visible on these pages rather than filled with a dash.
We never fill a gap with an estimate. Rule four of the method is that an unknown value stays unknown.
Where the figures come from
Every value carries the address it was read from and the date it was checked. Both are printed at the foot of each broker page, with a marker saying whether the source points at a document or only at a domain.
A figure whose size is set by a national regulator — leverage, minimum deposit — is used for this country only when it was published for this country. A group-wide figure is shown as absent here rather than borrowed: a worldwide leverage cap in a national ranking would put the most permissive offshore entity at the top of a market where the regulator does not allow it.
Figures are recalculated on a schedule and the date is on every page. The current base for ไทย is 100 brokers, of which 60 are priced.
What we do not do
We do not change a score for a partner, use promotional prices, or let a paid placement take a computed position. A paid row is labelled and sits outside the order.
We do not publish a ranking that rests on fewer than eight brokers: a top three drawn from four candidates is a list with a headline, not a comparison.