First-Party Research · Rule-Structure Statistics

Prop-firm rule-structure statistics 2026: 12 measures across 247 configs, 60 firm rule sets

Published 2026-09-23 · corpus last re-read 2026-09-27 · coverage 2026-07-16 · derived from rules/**, freshness.json and the rule-volatility study · 60 firm rule sets · verification log

Computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · method →. Structural counts and ratios only — no firm named, no firm ranked, no pass rate. Firm rule sets are not a count of unique operators; first modeled phases include evaluations and direct-funded plans.

1. Drawdown-type mix

In the first modeled phase, 160 of 247 configs (64.8%) encode a trailing drawdown floor; 87 (35.2%) encode a static floor. These are recorded model types, including source approximations.

Cite this › method → · “In the first modeled phase, 160 of 247 configs (64.8%) encode a trailing drawdown floor; 87 (35.2%) encode a static floor. These are recorded model types, including source approximations.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-1

How computed: Counted from phases[0].max_drawdown.type across 247 account-size configurations in the live corpus; phases[0] is the first modeled phase, including direct-funded accounts. Coverage date: 2026-07-16.

2. Recorded trailing-floor lock settings

In the first modeled phase, 97 of 160 trailing-floor configs (60.6%) have a lock setting recorded; 63 (39.4%) have no lock recorded. Of all trailing-floor configs: 73 (45.6%) record a lock at the starting balance; 24 (15.0%) record a lock above it. No lock recorded does not establish that a floor never locks; modeled settings can retain platform-specific source approximations. Excluded: 87 static-floor configs and 0 with another or unrecorded type.

Cite this › method → · “In the first modeled phase, 97 of 160 trailing-floor configs (60.6%) have a lock setting recorded; 63 (39.4%) have no lock recorded. Of all trailing-floor configs: 73 (45.6%) record a lock at the starting balance; 24 (15.0%) record a lock above it. No lock recorded does not establish that a floor never locks; modeled settings can retain platform-specific source approximations. Excluded: 87 static-floor configs and 0 with another or unrecorded type.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-2

How computed: Counted from phases[0].max_drawdown.locks across the 160 trailing-floor configs; non-null settings record the modeled lock level (at, value). All percentages use the full trailing-floor denominator. Coverage date: 2026-07-16.

3. Recorded drawdown reference basis

Among 244 configs with a recorded drawdown reference basis in the first modeled phase: end-of-day balance 110 (45.1%), intraday equity 112 (45.9%), and closed-trade balance 22 (9.0%). These reference/ratchet categories do not establish when breaches are monitored.

Cite this › method → · “Among 244 configs with a recorded drawdown reference basis in the first modeled phase: end-of-day balance 110 (45.1%), intraday equity 112 (45.9%), and closed-trade balance 22 (9.0%). These reference/ratchet categories do not establish when breaches are monitored.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-3

How computed: Counted from phases[0].max_drawdown.basis across the 244 of 247 configs that record it. This is the stored reference/ratchet category, not the breach-monitoring schedule. Coverage date: 2026-07-16.

4. Recorded daily-loss percentage distribution

Recorded daily-loss values in the first modeled phase: 0% encoded 86 (34.8%); > 0–2% 29 (11.7%); > 2–3% 55 (22.3%); > 3–4% 42 (17.0%); > 4–5% 30 (12.1%); above 5% 5 (2.0%). Absolute cap without a positive percentage: 0; no percentage or positive absolute cap recorded: 0 (all percentages use n=247). Source-specific percentage bases and optional caps are retained, so these are not uniform mandatory limits; missing values do not establish no limit.

Cite this › method → · “Recorded daily-loss values in the first modeled phase: 0% encoded 86 (34.8%); > 0–2% 29 (11.7%); > 2–3% 55 (22.3%); > 3–4% 42 (17.0%); > 4–5% 30 (12.1%); above 5% 5 (2.0%). Absolute cap without a positive percentage: 0; no percentage or positive absolute cap recorded: 0 (all percentages use n=247). Source-specific percentage bases and optional caps are retained, so these are not uniform mandatory limits; missing values do not establish no limit.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-4

How computed: Bucketed from phases[0].daily_loss.amount_pct across 247 configs. Explicit 0% is separate from missing data; a positive amount_abs without a positive percentage is counted separately. Positive percentages retain their own amount_basis and optional settings; no base conversion is performed. Coverage date: 2026-07-16.

5. Encoded consistency items + normalized thresholds

63 of 247 configs (25.5%) have a consistency item encoded in the first modeled phase; 184 (74.5%) have no item encoded, which does not establish absence of a real-world rule. Among 60 items with the normalized best-day metric, recorded thresholds range 20%–55%, most commonly 40% (16 items) and 50% (13 items). A further 3 items use another normalized metric. These stored thresholds do not establish a shared source profit denominator or real-world enforcement stage.

Cite this › method → · “63 of 247 configs (25.5%) have a consistency item encoded in the first modeled phase; 184 (74.5%) have no item encoded, which does not establish absence of a real-world rule. Among 60 items with the normalized best-day metric, recorded thresholds range 20%–55%, most commonly 40% (16 items) and 50% (13 items). A further 3 items use another normalized metric. These stored thresholds do not establish a shared source profit denominator or real-world enforcement stage.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-5

How computed: Item presence counted from phases[0].consistency across 247 configs; thresholds are .limit values on items normalized as best_day_pct_of_total_profit. Threshold and other-metric counts count items, not distinct configs. Source profit denominators and enforcement stages can differ despite the shared normalized metric. Coverage date: 2026-07-16.

6. Recorded payout-cycle day values

≤ 2 recorded days: 16 (6.5%). 3–9 recorded days: 42 (17.0%). 10–16 recorded days: 43 (17.4%). 17–35 recorded days: 11 (4.5%). > 35 recorded days: 0 (0.0%). No scalar cycle value recorded: 135 (54.7%). Source units include calendar, business, trading or qualifying days and conditions differ; these stored values cannot establish a uniform calendar payout cadence.

Cite this › method → · “≤ 2 recorded days: 16 (6.5%). 3–9 recorded days: 42 (17.0%). 10–16 recorded days: 43 (17.4%). 17–35 recorded days: 11 (4.5%). > 35 recorded days: 0 (0.0%). No scalar cycle value recorded: 135 (54.7%). Source units include calendar, business, trading or qualifying days and conditions differ; these stored values cannot establish a uniform calendar payout cadence.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-6

How computed: Numeric buckets of payout.cycle_days across 247 configs. Null means no scalar cycle value recorded, including conditional choices or on-request arrangements; it does not establish an unknown payout policy. Coverage date: 2026-07-16.

7. Minimum-trading-day distribution

Recorded minimum trading days in the first modeled phase: 0 days: 135 (54.7%); 1 day: 24 (9.7%); 2 days: 3 (1.2%); 3 days: 47 (19.0%); 4 days: 3 (1.2%); 5 days: 18 (7.3%); 10 days: 11 (4.5%); not recorded: 6 (2.4%).

Cite this › method → · “Recorded minimum trading days in the first modeled phase: 0 days: 135 (54.7%); 1 day: 24 (9.7%); 2 days: 3 (1.2%); 3 days: 47 (19.0%); 4 days: 3 (1.2%); 5 days: 18 (7.3%); 10 days: 11 (4.5%); not recorded: 6 (2.4%).” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-7

How computed: Counted from phases[0].min_trading_days across 247 configs, including direct-funded first phases; explicit zero is separate from an unrecorded value. Coverage date: 2026-07-16.

8. Recorded initial fee per $1 of initial modeled drawdown allowance

Recorded initial fee per $1 of initial modeled drawdown allowance ranges 5.0¢–70.4¢, median 9.9¢ (n=63; 184 configs excluded). This uses the first modeled phase, including direct-funded plans; an initial fee is not necessarily an evaluation fee or a total lifetime cost.

Cite this › method → · “Recorded initial fee per $1 of initial modeled drawdown allowance ranges 5.0¢–70.4¢, median 9.9¢ (n=63; 184 configs excluded). This uses the first modeled phase, including direct-funded plans; an initial fee is not necessarily an evaluation fee or a total lifetime cost.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-8

How computed: Ratio of pricing.initial_fee to phases[0].max_drawdown.amount_abs, across the 63 of 247 configs where the fee and a positive initial modeled allowance are recorded; 184 excluded. Coverage date: 2026-07-16.

9. Rule-change rate

11 canonical rule-value changes recorded across 60 firm rule sets, per the rule-volatility study dated 2026-09-11 (10 of 60 firm rule sets, 16.7%, show at least one; 0.058 value changes per raw touch).

Cite this › method → · “11 canonical rule-value changes recorded across 60 firm rule sets, per the rule-volatility study dated 2026-09-11 (10 of 60 firm rule sets, 16.7%, show at least one; 0.058 value changes per raw touch).” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-9

How computed: Read from research/rule-volatility-data.json findings (its own generator, its own dated snapshot) — canonical value-change and firm-rule-set counts, never recomputed here; rule sets are not a count of unique operators. Source dated: 2026-09-11.

10. Recheck disconfirmation signal

41 of 60 firm rule sets (68.3%) carry a recheck signal that the last-cited source sentence could not be re-located (reasons: cited sentence no longer found: 21, page moved: 14, page gone: 4, site moved: 2). This is a source-stability signal, not a confirmed rule-value change — compare block 9’s 11 confirmed value changes across 10 firm rule sets. A disconfirmation prompts a re-read; it does not by itself mean the rule moved.

Cite this › method → · “41 of 60 firm rule sets (68.3%) carry a recheck signal that the last-cited source sentence could not be re-located (reasons: cited sentence no longer found: 21, page moved: 14, page gone: 4, site moved: 2). This is a source-stability signal, not a confirmed rule-value change — compare block 9’s 11 confirmed value changes across 10 firm rule sets. A disconfirmation prompts a re-read; it does not by itself mean the rule moved.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-10

How computed: Read live from rules/freshness.json (dated 2026-10-15): a firm rule set is counted when its most recent recheck could not re-locate the previously-cited source sentence (last_disconfirmed set). Source dated: 2026-10-15.

11. Corpus size

60 firm rule sets across 247 account-size configurations. Coverage date 2026-07-16 (every rule set re-read at least this recently); most recent re-read 2026-09-27.

Cite this › method → · “60 firm rule sets across 247 account-size configurations. Coverage date 2026-07-16 (every rule set re-read at least this recently); most recent re-read 2026-09-27.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-11

How computed: Read from tools/corpus-sync.js computeCorpus(): firm-rule-set and account-size-configuration counts (not unique operators), plus the coverage-date guarantee (the oldest per-firm re-read anchor) and the most recent re-read. Coverage date: 2026-07-16.

12. Gaps between recorded recheck entries

Median 30 days between recorded recheck entries on the same config (n=352 gap observations from 203 configs); range 1–74 days. Entries may be scoped rechecks and do not establish whole-plan verification cadence.

Cite this › method → · “Median 30 days between recorded recheck entries on the same config (n=352 gap observations from 203 configs); range 1–74 days. Entries may be scoped rechecks and do not establish whole-plan verification cadence.” — computed with the PropSurvival strategy-survival engine corpus (seeded, reproducible) · propsurvival.com/research/prop-firm-rules-statistics#block-12

How computed: Gaps in days between consecutive dated entries in each config’s own rechecks[] history, pooled across the 203 of 247 configs with 2+ dated rechecks. Coverage date: 2026-07-16.

See a firm’s rules against your numbers Same engine, your own limits

PropSurvival models published evaluation rules. It does not rank firms, publishes no pass rate for any named firm, and is not affiliated with any firm. Dataset licence: CC BY 4.0. Dataset: JSON · plain text: Markdown. Method: research methodology. Firm-by-firm detail: rules census · change history: rule volatility.