Prop Firm EA Drawdown: Check the Loss Path and Test Report

Prop firm EA drawdown should be assessed from the account’s loss path and the exact programme rules, not from a single attractive backtest number. A trading strategy, its risk controls and permission to use the software are separate checks. This guide helps you read a report and construct a repeatable review; it does not report a tested EA, a measured pass rate or guaranteed compliance.

Understanding Prop Firm Drawdown Rules

Record provider, product, phase, account cohort, initial simulated capital, daily reset/time zone, balance/equity basis, open P/L and charges, plus the static or trailing overall formula. Do not infer these from a brand name. A nominal funded-account label may describe simulated capital rather than cash owned by the participant.

For current FTMO CFD 2-Step Challenge, Verification and subsequent 2-Step Account, the daily floor is the balance at 00:00 CE(S)T minus 5% of initial simulated capital; the total floor is static at 90% of initial simulated capital. The official objectives distinguish 1-Step and Instant. A market-opening time, Sunday rollover or assumed UTC midnight is not a substitute for the specified reset.

Read the EA report without changing the metric

MetaTrader’s tester report distinguishes balance from equity and absolute, maximum monetary and maximum relative drawdown. Balance at closed-trade timestamps can miss a worse open-position equity path. Absolute drawdown concerns the fall below the initial deposit; it is not a general name for every peak-to-trough loss.

Fictional balance sequence with no deposits or withdrawals: $10,000 → $11,000 → $10,450 → $12,000 → $11,430. The first decline is $550/$11,000 = 5%; the second is $570/$12,000 = 4.75%. Maximum monetary drawdown is $570 while maximum relative drawdown is 5%, from a different decline. Absolute drawdown below the initial deposit is zero in this sequence. These are observations of this artificial path, not future loss caps.

Profit factor compares gross winning amounts with the absolute gross losing amounts; it is not profit per unit of position risk. MQL5 STAT_RECOVERY_FACTOR is net profit divided by maximum balance drawdown, not days to recover. Fictional net profit $500 and maximum balance drawdown $250 give recovery factor 2. Recovery duration needs timestamps. Neither a high factor nor a smooth graph establishes programme eligibility.

Replay the daily floor along the actual path

This equity is below the daily floor even though the decline from $10,500 is only $510/$10,500 ≈ 4.857%. “Maximum historical drawdown below 5%” therefore would not prove that this daily rule passed. Keep the original event order, reset boundaries, floating exposure and charges. Do not subtract losses again when equity already includes them. A later recovery does not erase an earlier breach; the next day does not automatically cure it or create a waiver.

Verify position size and trade direction

Use the actual contract and account currency. In a fictional contract where one lot moves $10 per pip, a 50-pip stop represents $500 per lot before extra costs. With a chosen $500 budget and $20 cost reserve, the price-risk allocation is $480: $480/(50 × $10) = 0.96 lots if a 0.01 step is supported. This is estimated risk, not a maximum possible gap loss. A stop distance alone does not determine loss; quantity matters.

For a separate fictional short entered at 1.1000 and closed at 1.1080, with a 0.0001 pip and the same $10/lot/pip at 1.5 lots, price loss is 80 × $10 × 1.5 = $1,200 before costs. The price rose against the short; closing that short requires a buy transaction. A downward move would be favourable before costs under these assumptions.

Keep a reproducible test and incident record

1. Save EA version/hash, input file, terminal build, account mode, instruments, interval, data mode, contract specifications and cost/execution assumptions. Separate tuning data from an unseen evaluation interval.
2. Replay all applicable floors with timestamped equity, charges, open positions and resets. Record first violations instead of evaluating only the final balance.
3. Test wider spreads, changed execution assumptions, gaps, missing data, disconnection, restart and rejected/partial closes. A historical simulation does not demonstrate broker-side failure handling.
4. Record requests, response codes, confirmed deals and remaining exposure. A close request is not a fill. A local buffer is a chosen assumption, not proof that a contractual limit cannot be crossed.
5. Document the trigger for pausing and the conditions for resuming. Neither blindly trusting an EA nor changing it after every loss constitutes validation.

No universal win rate, Sharpe score, modelling-quality badge, fixed trade count or number of years certifies a pass. Position increases after losses and accumulated grid exposure need explicit cash-risk analysis; do not infer permission or prohibition merely from a strategy label. News/session permissions must match the exact programme and actual execution.

Choosing an EA: documentation and permission before promises

The JPTC EA Hub is customer-operated MetaTrader software. Check the specific file’s documented risk functions, platform and permission; this article does not establish automatic support for every provider or a tested configuration. The separate JPTC Algo is hosted by JPTC, with no customer installation or own VPS. The company’s broker-statement figures are not an independent audit, your account tracker or a test of another EA.

Passing a loss check does not alone satisfy profit targets, other eligibility conditions or payout requirements. Primary rules and metric definitions checked on 16 October 2026. All numerical paths and trades above are fictional worksheets.