Grid EA on Prop Firms: Permission, Basket Risk and Testing

A grid EA places or manages entries across price levels. A mean-reversion grid may buy as price falls and depend on a recovery to exit. Other grids follow breakouts. Neither a fixed lot nor a profitable backtest establishes a bounded basket loss or permission to use that strategy on a prop programme.

Grid entries and martingale are different

Write down the entry direction, spacing, number of levels, maximum lots and exit for every leg. Price spacing alone does not define risk. In a mean-reversion grid, a continuing trend can fill additional losing positions before a reversal occurs. Closing profitable legs can leave losing inventory behind; the realised balance can therefore look better than the equity.

A grid and martingale are different concepts. A grid describes entry levels, while martingale increases size following losses. They can be combined. Five equal 0.01-lot positions total 0.05 lot; sizes of 0.01, 0.02, 0.04, 0.08 and 0.16 total 0.31 lot. Fixed sizing avoids that doubling sequence but does not stop aggregate exposure or losses from growing.

Check actual programme permission first

Official pages checked on 16 October 2026: FundedNext’s CFD prohibited-strategies page explicitly prohibits grid trading. Removing martingale or adding a loss stop does not override that restriction. EA permission and a Grid mode in software are separate from permission for the strategy.

FTMO’s CFD strategy FAQ conditionally permits algorithmic trading subject to its forbidden practices and risk-management rules. Those include overexposure and increasing cumulative risk in a symbol or correlated symbols. This is not blanket grid approval. The FAQ also states that a stop-loss is not mandatory; it does not promise a code review that approves each EA before activation.

Check your exact programme, phase, platform, strategy and contractual permissions. Do not transfer a CFD rule to a futures programme or assume that a third-party preset is approved by a named firm.

Calculate the entire grid at one common exit price

Illustrative USD-account calculation only: EUR/USD contract size 100,000 EUR per standard lot and pip size 0.0001 USD. Assume all five buy orders have filled, each at 0.05 lot, with no intervening exits. Spread, commission, swap and slippage are excluded.

The basket’s distances to 1.0800 sum to 800 pips. For an explicitly chosen $500 scenario budget and assumed $20 total cost reserve, equal volume per level is ($500 − $20) ÷ (800 × $10 per pip per standard lot) = 0.06 lot. The five legs then lose $480 before the assumed costs, totalling $500. This is a scenario calculation, not a recommended size or maximum possible loss. A worse exit or extra filled levels exceeds it. Use the actual contract, account currency, conversion and volume step; MT5’s OrderCalcProfit can estimate individual-leg profit or loss in account currency.

Why a configured loss stop can still fail

Apply the programme’s actual equity or balance reference, reset time and static or trailing floor. For the current FTMO CFD 2-Step example, the daily loss amount is 5% of initial simulated capital and the overall static amount is 10%; the daily equity limit is recalculated at 00:00 CE(S)T. Remaining room can already be consumed by losses, costs and other open positions.

An EA threshold only requests action when its code runs. It cannot guarantee the exit price during gaps, wide spreads, rejected orders or a disconnected terminal. Pending orders can fill while a basket is being closed. Blocking new entries does not cancel existing pending orders or close current positions unless the implementation does those things.

A wider grid is not automatically safer: it changes fill count, distances, holding time and loss per leg. A larger take-profit or volatility filter also does not create a demonstrated profitable edge.

Test the basket path and implementation

Freeze file version, inputs, contracts and rules. Track every entry, pending order, fill, exit, cost and equity observation. Compare the result with the intended no-recovery path, not only with a final profitable balance. Record the deepest floating loss, maximum lots, margin and duration of underwater inventory.

Exercise sustained trends, a gap across multiple levels, one profitable leg closing while losses remain, rejected basket closure, partial fills, restarts and daily-reset boundaries. Check whether a restart reconstructs the existing basket or mistakenly opens it again. Compare fixed-size, capped and no-grid variants with the same data and cost assumptions.

MT5’s tester supports selectable tick models and execution-delay settings. Record the actual mode and modelling limits. Keep parameter selection separate from held-out evaluation. No universal 500-bar minimum, one-year test, 55% win rate, Sharpe threshold or ATR setting proves a strategy will pass. This article reports no measured passing rate or customer case study.

Separate JPTC software from evidence

The downloadable JPTC EA Hub runs in the customer’s own MT4 or MT5 terminal and offers Grid among its workflows. A feature or preset is not a provider approval. Risk settings depend on the exact build and inputs; losses can exceed configured thresholds. The separately hosted JPTC Algo is a different service.

JPTC’s results page reports figures from its broker statement. It is not an independent audit or proof that an EA Hub grid passes an evaluation. Compare an explicit strategy specification, permission, complete test records and limitations before deciding whether a particular grid merits further testing.

Sources checked 16 October 2026: MT5: profit calculation · MT5: tester · FTMO: strategy FAQ · MT5: margin · FTMO: forbidden practices · FTMO: objectives · FundedNext CFD: allowed · FundedNext: prohibited strategies · JPTC EA Hub.

FAQ: Grid EA on Prop Firms