Key points
- A ranked list hides whose priorities produced the order.
- Start from your constraints, not from a shortlist someone else built.
- Compare like with like: same currency, same fee type, same unit, same measurement.
- The fields that end accounts matter more than the fields firms advertise.
What is wrong with a ranked list
A ranking is a single order produced from many incompatible measurements. To produce it, somebody decided how much a 90% split is worth against a static drawdown, how much a low fee is worth against a late first payout, and how much a firm's age is worth against its published terms. Those weightings are the whole answer, and they are almost never published.
The problem is worse in this industry than most, because the fields genuinely are incompatible. A percentage limit and a currency limit are different units; a one-time fee and a monthly subscription are different cost structures; lots and contracts measure different things. An order across them is not a summary, it is an assertion.
Rex therefore publishes no ranking and no score, and will not until it can explain exactly how one was produced. What it publishes instead is the evidence and the incompatibilities. Why there is no ranking sets out what a future score would and would not measure.
A method that works instead
- Fix your constraints firstMarket, account size, budget per charge and fee type. These are facts about you, and they eliminate most of the field before any judgement is involved.
- Choose the rule your strategy cannot live withFor most people this is the drawdown model. Intraday trailing, end-of-day trailing or static — pick the one that matches how your equity curve actually behaves.
- Filter, do not browseApply those as filters and read the count. If four options match, you have a shortlist. If forty do, you have not been specific enough.
- Compare two or three in fullSide by side, field by field, with the incompatible fields marked rather than averaged.
- Read the source pages for the survivorsEvery value in Rex links to the firm page it came from. Read the two or three that matter on the firm's own site before buying.
The whole process takes about twenty minutes and produces a shortlist you can explain, which is the part a ranking cannot give you.
Comparing like with like
Four pairs of values look comparable and are not. Rex marks each of them incompatible rather than producing a number, and the reason is always shown.
| Looks comparable | Why it is not |
|---|---|
| €345 against $299 | Different currencies. No conversion is applied, because the rate would be an assumption inside a comparison |
| $100/month against $299 one-time | Different cost structures. The answer depends on how long the evaluation takes |
| 8% max loss against $2,000 max loss | Different units, and the percentage scales with the account while the amount does not |
| A $2,000 intraday trailing limit against a $2,000 end-of-day one | Same number, materially different amount of room |
When Compare tells you two options cannot be compared on price, that is a real finding about the decision, not a failure of the tool.
Which fields actually decide the outcome
Firms advertise the profit split and the price. Accounts are lost to the drawdown model and the daily loss, and payouts are delayed by consistency rules and eligibility conditions. That mismatch is worth correcting for deliberately.
- Decides whether you keep the account: maximum loss type, stop point, measurement, daily loss basis and breach effect.
- Decides whether you get paid, and when: first payout eligibility, consistency scope, base profit split.
- Decides what it costs: fee type, activation fee, reset fee, refund condition.
- Decides very little on its own: the advertised split ceiling, the account size, the firm's marketing claims.
What Data Confidence is, and is not
Rex shows a Data Confidence percentage on every firm and option. It measures how completely Rex has documented that object — how many applicable fields carry a source, weighted by how critical each field is and how fresh the evidence is.
It is not a quality score and must not be read as one. A firm with a low figure may publish less, or may simply be one Rex has read less of. Sorting by it finds the firms you can check most thoroughly, which is useful and is not a league table.
The same applies to Research Depth. "Full research" means published facts cite at least two distinct authoritative pages from that firm. It is a statement about Rex's coverage, not about the firm.
Turning a shortlist into a decision
Once two or three options survive, three questions usually separate them, and all three are answerable from published terms.
- On the field my strategy cannot live with, which is most forgiving? Not a general verdict — most forgiving for the way this particular strategy behaves.
- What is the realistic total cost, including an activation fee and one reset?
- How much of each plan is actually published? An option missing its drawdown type or its payout eligibility is an option you are partly guessing at.
Start from a filtered set: discovery pages group options by an objective criterion and state the criterion in full, or filter directly in the challenge explorer and open two in Compare.
Common questions
Will Rex ever publish a ranking?
Only with a published methodology, a version number and a per-firm breakdown of how the figure was produced. Until then, no.
Can I sort by Data Confidence to find the strongest firms?
You can sort by it, but it measures documentation, not quality. Rex labels it that way on every surface for exactly this reason.
Why does Compare refuse to tell me which option is cheaper?
It will, inside a currency and fee type. Across them, the honest answer depends on assumptions you should make yourself rather than have made for you.
How many options should be on a shortlist?
Two or three. If a filter leaves you with twenty, add the rule your strategy cares about most; that is usually the drawdown model.
Compare the options this guide mentions
Opens 3 challenge options side by side, with every field that cannot honestly be compared marked rather than averaged.