Methodology limitations
What this study cannot do, where the gaps are, and what a reader should weigh against the conclusions.
What this study does not cover
It does not measure outcomes. It does not measure slippage, execution quality, withdrawal speed, customer service, or whether a particular trader passed an evaluation. It reads published rules and produces a verdict on the rule — a verdict on outcomes belongs to the trader who took the trade.
What this study cannot see
It cannot see rule changes between the date we read a firm and today. The vigia (nightly crawler) re-reads every page every night, but a firm can change a clause in the hours between two reads. It cannot see a forced live conversion that was sent by email to one trader and not published. It cannot see a discretionary refusal of a payout that the firm did not record. It calls these "things the firm has not written down" and treats them as missing data, not as accusations.
Where the data is young
Some firms in the register have been read once. Some have been read many times. Some have not been read at all, and live in the directory as "not judged". A verdict that says "this firm is safe" requires that we read it; a verdict that says "this firm is too new to judge" is the honest substitute when we have not.
Where the data is wrong
When a firm's own pages contradict each other, we cite both. When the contradiction costs the trader money, it is graded down. When the contradiction is cosmetic — a stale marketing figure nobody updated — it is not. The grading rubric is published on the methodology page, and the contradictions are listed on the row of every firm that carries one.