What We Could Verify Across Six Pilot Accounts

This study follows the evidence behind the six accounts used in our first two comparisons. It identifies the questions left open, separates rules from trading permissions, and shows how the picture changes when a question repeated across accounts is counted once.

Which six accounts did we study?

The cohort is the original pilot: one CFD/FX account and one futures account from each of FTMO, FundedNext and The5ers. These accounts were chosen for the pilot, not randomly sampled from the industry. Their nominal sizes identify the selected offers; they do not make the two markets equivalent.

The same release covers 74 accounts at these three firms: 54 CFD/FX and 20 futures. Those accounts provide context later in the study. They include the six pilot accounts, so the two groups must not be added together. Firms and accounts are listed in alphabetical order within each market, without a ranking.

For the rules themselves, use the three CFD/FX accounts compared and the three futures accounts compared. This study examines the evidence states recorded for those rules.

What exactly are we counting?

The data is the public release pfw-release-2026-10-09-pilot-3, approved on 9 October 2026, with source checks from 19 September through 9 October. We recomputed the counts from its account-rule and trading-permission exports on 10 October. That analysis date does not mean every firm's source was reread that day.

A rule row is one account paired with one field in its market's comparison table. Each CFD/FX account has 24 rows; each futures account has 23. The pilot denominators are therefore 72 and 69. A row can contain several conditions, so it is not the same as one indivisible claim.

A permission cell is one account, one activity and one stage. The four activities are news trading, holding, expert advisors and copying; the stages are evaluation and funded. These cells organize parts of the rule rows in more detail. We report them separately, without adding them to the rule-row totals.

The release's evidence states have different meanings:

For this study, open rows means the three states Partly unknown, Sources differ and Interpretation. It is a counting group, not a new site status. “Documented total” combines Documented and Documented: none. The evidence method explains why an unknown answer must stay distinct from zero or permission.

The main tables include three calculated fields per account: Opening loss allowance, Difficulty ratio and Fee per $1,000 of opening risk. This preserves the set of fields a reader sees, but those calculations are derived from other inputs and are not three additional independent pieces of evidence. We also show what happens when they are excluded. See how PFW treats its numbers.

Every percentage is its count divided by the stated denominator, multiplied by 100 and rounded to one decimal place. Rounded percentages may not sum to exactly 100.0%. Fields are counted equally; these percentages do not weight how much a missing answer matters to a particular trader.

How much of the pilot had documented states?

There were 68 documented rows out of 72 for CFD/FX (94.4%), and 64 out of 69 for futures (92.8%), including documented “none” answers.

Neither market has a Not applicable or Not verified rule row in the pilot. The four CFD/FX open rows represent 5.6% of that market's rows; the five futures open rows represent 7.2%.

All three calculated fields per pilot account have Documented states. Excluding them removes nine rows from each market, while leaving the open-row counts unchanged:

This difference comes from the counting choice. No new evidence has appeared or disappeared. Neither table establishes that one market's programs are safer, easier to trade or more likely to pay.

Which nine rule rows remained open?

The tables below identify every pilot row in the three open states. Follow the exact row to examine its evidence. The descriptions identify what a reader should check; they do not reproduce the accounts' prices, limits or payout amounts.

CFD/FX: four open rows, all on the selected $100,000 accounts.

Futures: five open rows, all on the selected $50,000 accounts.

The row-level export does not reproduce every source passage or explain every unresolved subquestion. Where that detail is absent, the appropriate next step is to inspect the cited evidence, not invent the missing explanation.

What did the stage-specific permission cells establish?

There are 24 permission cells per market in the pilot: three accounts, four activities and two stages.

There are no Partly unknown, Sources differ or Not verified permission cells in the pilot export.

Both interpretations concern copying between one's own FTMO 2-Step Standard accounts, one cell for evaluation and one for the funded stage. They describe the same support-based issue already counted in the expert-advisors-and-copying rule row. They are not two additional independent rule gaps.

Permission and evidence are separate dimensions. A documented cell can say allowed, restricted or prohibited. Conversely, a rule row can stay Partly unknown while its permission cells are Documented: FundedNext Stellar's holding row is the pilot's example, and the export does not record what remains open in it.

What changes when all 74 accounts are included?

The wider release adds sizes and programs at the same three firms. It does not add an industry-wide sample. Keep these results separate from the six-account pilot.

Including documented “none” answers, the documented totals are 1,268 of 1,296 CFD/FX rows (97.8%) and 428 of 460 futures rows (93.0%). No rule row in this export is Not applicable or Not verified.

For permissions, the denominator changes:

The 20 excluded cells belong to evaluation stages that do not exist for direct-entry accounts: FundedNext Stellar Instant in CFD/FX and FundedNext Direct in futures. They are not missing research. No applicable permission cell has another evidence state in this export.

The 60 open rule rows comprise 28 CFD/FX rows and 32 futures rows. Grouping them by market, firm and rule produces ten questions:

A firm often states a rule once for a program, or once for several programs in a market, so the same evidence state repeats across accounts. FundedNext's futures cancellation question, for example, appears on ten accounts across four programs. That is one grouped question affecting ten account rows, not ten independently discovered disputes. The grouping is descriptive; it does not establish that one answer would resolve every account's circumstances.

What can these labels and percentages miss?

A documented row can still leave a detail unanswered. In this snapshot, The5ers Day Trade's payout-minimums-and-caps row was Documented while its wording said the minimum withdrawal was unclear. We preserve that state in the counts and identify the limitation here. Moving it into Partly unknown would silently change the release being studied.

Later clarification, 10 October 2026: the next approved release reworded that row to distinguish the profit requirement before a payout from the cap on each payout. Its Documented state remained unchanged. This later clarification does not alter the 9 October counts.

Account-row links open the live site, which can show a newer release than this study. The supporting data preserves the snapshot for reproducing its results. A changed current page is not, by itself, an error in a dated study. See how changes and check dates are handled.

Other limits follow from the design:

• Selected accounts and fields. Six pilot accounts, or 74 accounts at three firms, cannot establish the state of the whole industry. This export covers the public comparison fields, not every possible question about a firm.
• Unequal practical importance. A missing answer about an essential trading method may matter more to you than several documented fields. The tables give each row one count.
• Repeated and derived information. Sizes can share a source, calculated fields reuse inputs, and permission cells overlap rule rows. None is an independent sample of a firm's behavior.
• Recorded states, not a fresh source audit. We recomputed the release's classifications; we did not independently reopen every underlying source or reclassify every sentence for this study.
• No outcome measurement. The export does not establish pass rates, realized payout rates, payment reliability, solvency or trader experience. The private research store's reading history, resolved conflicts and between-release changes are outside this analysis.

A high documented share is therefore a description of this dataset's labels. It is not a transparency score, a reliability ranking or a forecast of your outcome.

What should you check when a rule matters to your decision?

Start with the action or payment condition you need, then follow the exact account row.

Use the evidence worksheet to record what supports your conclusion. If sources disagree, follow the conflict-identification steps and the test for what counts as resolved.

These checks help you decide what remains to be established before relying on a rule. They do not replace the separate evidence needed to evaluate a claim about an actual payout.

How can you reproduce the study or report a correction?

The two Supporting data downloads below the lead contain:

The values are as the site showed them on 9 October, and some have changed since; the account pages show the current ones. FTMO's CFD/FX fee per $1,000 of opening risk was converted from euros at that day's ECB rate.

To reproduce the tables:

1. Check that every row names the study's data release. For the pilot tables, filter the pilot field to yes.
2. Split each file by market. In the rules file, count unique account-and-rule pairs by evidence state. Include all fields for the main tables.
3. Combine Documented and Documented: none for the documented total. Combine Partly unknown, Sources differ and Interpretation for the open-row total.
4. For the sensitivity check, exclude the three calculated fields identified by the metric value-source category.
5. In the permissions file, count unique account, topic and stage combinations. Exclude Not applicable cells before calculating percentages.
6. For the ten-question table, filter to open rule states and group by market, firm and rule. Retain the evidence state; do not merge unlike questions merely because their labels resemble each other.

Source counts are references attached to rows, not counts of independent confirmations. Adding them would count repeated sources more than once. The exports also carry condition flags rather than the complete text of every condition.

If you find a counting error, an incorrect classification or evidence that changes a conclusion, use PFW's correction route and identify the release, account, field and supporting record. A correction to this study's analysis should be identified as a correction. A later rule change belongs in a dated update or a new study edition.

Related reading

• How to read prop firm reviews and payout claims, for the limits of different kinds of evidence.
• When official prop firm rules conflict, for preserving and rechecking unresolved questions.
• How to compare prop firm programs, for using uncertainty in a before-you-pay decision.

Sources and method

• Public release and coverage export, 9 October 2026: the sole dataset for the study's counts and historical row states. The supporting-data instructions describe the two files and how to reproduce the tables.
• Our Approach: evidence, numbers, changes and corrections, checked on 10 October 2026 for PFW's published research method.
• Approved public release, 10 October 2026: used only for the separately dated The5ers payout-wording clarification, checked against the published account record on 10 October. It does not replace the study dataset.

The analysis reports PFW's own published classifications. The account-row links identify the relevant live records, while the downloads preserve the study's release. No private research history, reliability score or measured trading outcome is part of the study.