Source result: a single industry-wide prop-firm pass rate is not verified. In the official sources reviewed, Topstep publishes 2025 metrics with several explicit denominators; comparable current rates for FTMO, Apex, The5ers, and FundedNext were not established. No comparable official figure was found for those firms, so community estimates are not substituted.
“Pass rate” can mean the percentage of challenge starts that advance, unique people who ever advance, people at a funded level who receive a payout, or accounts that remain active. These are different populations. Combining them produces a precise-looking number that answers no stable question.
Pass Rate Overview by Firm
| Firm / program family | Official metric found | Period and denominator | Safe interpretation |
|---|---|---|---|
| Topstep Trading Combine | 16.8% advanced | All Trading Combines initiated in 2025 | Attempt-level advancement, not unique traders |
| Topstep participants | 51.8% advanced at least once | Distinct 2025 participants who entered at least one Combine | Person-level “ever advanced” result |
| Topstep Funded Level | 33.3% received a payout | Distinct 2025 participants at Funded Level | Payout-recipient rate within that level |
| Topstep Express to Live | 0.71% called to Live | Distinct 2025 Express Funded participants | Live-call-up rate, not payout rate |
| FTMO evaluation | No comparable official rate found | No denominator-complete figure in the official pages reviewed | Do not use 10–12% as official |
| Apex evaluation | No comparable official rate found | No current official figure was established | Do not use 12–18% as fact |
| The5ers / FundedNext | No comparable official rate found | No current official figure was established | Program and phase still need exact scope |
Topstep's disclosure also warns that its programs use simulated environments and that historical performance does not predict future results. Read the definitions on the source page before quoting any percentage.
Open Topstep's current risk disclosure.
The program comparison below uses the same country scope, nominal size, and first evaluation phase. It supplies current rule context, not pass-rate estimates. Without compatible official denominators, ranking firms by pass rate is not defensible.
The Widely Cited 8–17% Estimate: Useful Signal, Unverified Denominator
The March 2026 version of this TSB guide published a cross-firm estimate that 8 to 17 out of 100 evaluation starts pass the challenge. External industry pages still cite that range. The original compilation described a blend of firm disclosures, community surveys, third-party trackers, and anonymized TSB evaluation tracking, but no frozen export, deduplication rule, program mix, time window, or reproducible calculation remains in the project.
Evidence label: treat 8–17% as a legacy industry-reported estimate, not a current official cross-firm rate and not a breakdown of TSB's 600K+ cumulative imported trades. It is useful for downside planning: model a low-pass environment and cap total attempt spend. It is not suitable for ranking firms whose denominators differ.
The same legacy compilation estimated that roughly 70% of recorded failures involved maximum-loss or daily-loss limits. That failure-reason cohort is no longer reproducible, so the percentage remains an editorial research hypothesis—not a measured current TSB fact. The operational insight survives: loss-limit mechanics deserve a dedicated replay before purchase, while strategy quality, target failure, prohibited conduct, inactivity, data gaps, and voluntary abandonment must remain separate possible paths.
The externally cited 100-trade test is also preserved as a method rather than a magic threshold. If you have 100 eligible trades under one stable strategy version, replay all 100 through the exact program rules and report how many were included, excluded, breached daily loss, breached maximum loss, or reached the target. If only 63 trades meet the filters, call it a 63-trade eligible cohort; do not imply that the other 37 passed or failed.
Why Traders Fail: What the Rules Can Prove
The reviewed evidence does not support a percentage breakdown of failure reasons. Maximum loss, daily loss, profit target, minimum days, time limits, inactivity, and prohibited conduct are mechanical failure paths; knowing that they exist does not show how frequently each causes failure.
For the exact program, create a rule-to-data map:
| Rule path | Data required | Pre-purchase test |
|---|---|---|
| Daily loss | Definition of day, balance basis, reset time, included costs | Replay every historical trading day |
| Maximum loss | Static or trailing method and intraday/EOD basis | Rebuild account equity at required frequency |
| Profit target | Target and eligible-profit definition | Measure target-reaching time without increasing risk |
| Minimum days / time window | Qualifying-day and expiry definitions | Apply actual trading frequency |
| Restrictions | News, holding, copying, automation, instrument rules | Compare each strategy behavior with current terms |
A failed replay identifies a specific incompatibility. It does not prove that the same path is the most common reason other traders fail; the legacy ~70% estimate above remains a hypothesis until its cohort can be reproduced.
Phase-by-Phase Breakdown: Keep Cohorts Intact
For a two-phase evaluation, define:
p1= Phase 1 passes ÷ Phase 1 starts;p2= Phase 2 passes ÷ Phase 2 starts from the same cohort;- combined pass share =
p1 × p2, only whenp2is conditional on passing Phase 1 and cohort definitions align.
Example only: if 40 of 100 starts pass Phase 1 and 20 of those 40 pass Phase 2, p1 = 40%, p2 = 50%, and the combined result is 20%. Do not mix a community Phase 1 survey with a firm's all-time funded count.
For a one-phase program, advancement still may not mean payout. Report the funnel separately: starts → advanced accounts or people → funded-level participants → payout recipients → live accounts, where applicable.
What You Can Test Before an Evaluation
No reproducible TSB dataset supports claims that passers risk 0.5–1%, take 3.2 trades a day, or use 60–80% of the time window. Replace that invented profile with personal stress tests:
- Replay the exact daily and maximum loss definitions against timestamped trades.
- Run adverse-order simulations: shuffle trade order or use rolling historical windows without changing the outcomes.
- Test smaller fixed risk fractions and record target completion versus breach frequency.
- Include spreads, commissions, slippage assumptions, and timezone boundaries.
- Freeze the strategy and rule version before an out-of-sample replay.
These tests estimate compatibility with a program. They do not manufacture a personal pass probability from a small or selected sample.
How Many Attempts to Pass?
No sourced median number of attempts is available here. Track actual purchases and outcomes per person and per program. Never infer a person-level median from attempt-level pass rates.
If each attempt had a constant independent success probability q, a mathematical model would give:
probability of at least one success in n attempts = 1 − (1 − q)nReal attempts are not necessarily independent: traders learn, switch programs, change risk, stop participating, or repeat the same error. Use the formula as a sensitivity exercise, not a forecast or reason to buy more attempts.
The net prop-firm cash ledger is simpler and observable: cumulative attempt cost is the sum of every evaluation, subscription, reset, activation, data, and platform payment. Use current exact-program prices from official checkout rather than a stale firm-level table.
After Passing: Funding, Payout, and Live Are Different
“Funded” can refer to a simulated reward-eligible level or a live brokerage account. FTMO's official challenge page describes its environment as simulated. Topstep separately reports Express Funded and Live Funded pathways. Neither label should be converted into owned capital.
Topstep's 33.3% payout statistic uses individual Funded Level participants as its denominator. Its 0.71% Live call-up metric uses individual Express Funded participants. They do not show that a stated percentage of all challenge starters was paid or remained funded for 90 or 180 days.
The old 40–50% 90-day survival assertion and causal loss-aversion explanation were removed because the cited evidence did not establish either.
The Real Cost of Failing
Use cash facts rather than “estimated cost to get funded”:
net cash outcome = payouts received − all attempts − resets − activation − data/platform/transfer costs − taxes dueA failed attempt has zero payout in the ledger. A passed evaluation does not erase prior costs. A promised fee reimbursement enters only when its exact program condition is met and the value is actually received or credited.
Use the like-for-like program comparison, then save the checkout and rule version you actually accepted. Prices and restrictions can change, so recalculate the cash outcome before every purchase.
How to Improve Your Decision Process
- Define the denominator: decide whether you need attempt advancement, person advancement, payout receipt, or survival.
- Choose one exact program: region, size, stage, platform, and rule version must match.
- Replay before paying: apply loss rules and qualifying-day logic to untouched trade data.
- Set a cash stop: cap total attempt spending before the first purchase.
- Keep a cohort log: one row per attempt with dates, program ID, outcome, failure path, and cost.
- Review changes: if current written rules conflict with the plan, stop and reassess before paying.
Your Next Step
Use the prop-firm rules checklist to capture the exact rule version and the prop-firm calculator for scenario testing. Label an unknown rate as No comparable official rate found.
The evidence-backed conclusion is narrower than the old article but more useful: Topstep's published 2025 rates can be quoted only with their denominators; the other named firm rates remain unknown in the official evidence reviewed. Your purchase decision should come from exact rules, personal replay results, and an affordable zero-payout case—not a synthetic industry average.
