Gold Level Breaker

Every figure below comes from one backtest: XAUUSD, 121,924,747 real ticks, 2024.02.15 - 2026.09.26, history quality 100%.

What it trades
Yesterday's daily high and low are the two levels. Around each one sits a zone whose width is a percentage of price, so it adapts as the instrument moves - on gold that same setting meant $5.6 in 2024 and $11.2 in 2026 without being touched.

Three independent patterns run on their own magic numbers and can be switched off separately:

• S1 - breaks the level, then comes back to retest it (M15)
• S2 - retests the level first, then breaks away from it (M12)
• S3 - retests, pulls back, retests again (M6)

A series that moves against itself adds further entries at the same lot size. There is no doubling and no martingale. The series closes as a basket when the weighted average price reaches the take profit, or at the basket stop loss of 1.5%.

A day-range filter blocks everything when yesterday's range was below 25% or above 150% of the 4-day ATR - too quiet to break out, or already extended.

Measured results
XAUUSD M1, $100,000 account, 100% real ticks, 2024.02 - 2026.09:

• Conservative 0.25% - net $2,696, profit factor 2.95, max equity drawdown 0.59%
• Moderate 0.50% - net $6,014, profit factor 2.64, max equity drawdown 1.64%
• Aggressive 1.00% - net $13,294, profit factor 2.58, max equity drawdown 3.48%

Across the full sample: 541 trades, 79.7% of them profitable, longest losing streak 8, average position held 3 hours. Profit and drawdown scale together almost exactly, so the risk level changes the size of the result and not its shape.

Position sizing
The lot is solved from the strategy's own parameters rather than guessed: the number of entries that fit before the stop comes from the grid step, and those entries sit on average half the stop distance away.

When the broker's minimum lot would exceed the risk you asked for, the EA does not trade. It says so in the journal instead of quietly taking 30x the intended size. On a $1,000 account at 0.5% the honest lot is 0.0003 and the minimum is 0.01, so by default nothing is opened.

Account protection

• Daily loss stop - no new series for the rest of the day
• Total drawdown stop - no new series at all
• Hard cap on entries per series, and on one entry

These brakes never close a position that is already open. A basket is closed by its own take profit or stop loss.

On-chart panel
Balance, equity and floating P/L; the lot each pattern would use next and the real percentage of the account that lot risks; the ATR filter with yesterday's range as a percentage; and for every open series its leg count, average price, open P/L and the distance still to run to take profit. Yesterday's levels and their zones are drawn on the chart.

Requirements

• MetaTrader 5, hedging or netting account
• Any symbol and timeframe - every distance is a percentage of price, nothing is hardcoded for gold. The results above are XAUUSD only; test your own symbol.
• VPS recommended: the basket exit is evaluated by the EA, not left with the broker

What you should know before buying

• This is a grid system. Several entries can be open in one direction at once. The smooth equity curve in the screenshots is what grids look like while the deep case has not happened; in 32 months the worst single trade was $376 at the moderate level. That is not a promise that it is the worst possible one.
• Stop loss and take profit are managed by the EA, not placed with the broker. If the terminal is not running, an open basket is unprotected. Use a VPS.
• 32 months on one symbol is the whole evidence base. It covers one strong trend and one quiet period, not every regime that exists.
• Past results do not guarantee future performance.