Oracle Ratchet Gold M15

RATCHET GOLD M15

Trend-continuation Expert Advisor for XAUUSD. No grid. No martingale. No averaging down.

THE CORE RULE

Every new long opens only ABOVE the previous long. Every new short opens only BELOW the previous short.

The EA never adds to a losing position and never tries to catch a reversal. It scales into a move only after the market has confirmed that the previous entry was right. That is the ratchet: a mechanism that turns in one direction only.

This single rule rejects roughly 87% of the raw breakout signals the EA detects. What survives is the part of the move the market has already paid for.

HOW IT TRADES

Entry. Price must break a Donchian channel on the close of an M15 bar, confirmed by an ATR based impulse filter. Three parallel subsystems watch channels of different length, so a position can be built across several stages of the same trend.

The anchor. Each executed entry raises the bar for the next one. A new position opens only once price has advanced beyond the previous entry by a configurable ATR margin. During flat periods the level decays slowly, so the EA cannot lock itself out permanently after entering at a market top.

Stop loss. Placed at the nearest swing extreme and never modified afterwards. No trailing, no breakeven, no partial closes. The risk you see at entry is the whole risk.

Exit. No fixed take profit. A position closes when price closes back through a slow moving average, which lets a winning trade run for as long as the trend lasts.

Scale independence. Volatility filter, minimum and maximum stop distance and the stop buffer are all expressed as a percentage of price rather than in fixed money. The same settings behave the same way at gold 1600 and at 5500.

WHAT YOU SEE ON THE CHART

A status panel shows the live anchor levels, the exact price at which the next entry unlocks, open positions, lot, ATR, spread with a warning when it is too wide, and how many raw breakouts were rejected and by which filter.

The trend EMA and the exit EMA are drawn across the visible history in separate colours. A regression trend channel and support and resistance built from real swing extremes complete the picture. Every trade becomes readable: why an entry happened, why a signal was skipped, why a position closed.

A signal diagnostics report is printed at the end of every run, listing how many breakouts were found and which filter rejected each one. Tuning becomes a measurement rather than guesswork.

POSITION SIZING

Three modes, selected by input.

Fixed lot. The simplest option and the default. Every trade uses the same volume regardless of account size.

Proportional. One base lot is traded for every configurable amount of balance, so the risk per trade stays constant whatever the account size. Balance excludes floating profit, which means the lot only grows on realised results and never shrinks after a drawdown. A hard lot cap limits how large a single position can become.

Step compounding. The lot increases by one step every time the balance gains a set percentage of the starting balance. Kept as an alternative; in testing the proportional mode produced a lower drawdown and a higher recovery factor.

HOW TO SWITCH MODES

All three modes live in the same file. The EA ships on the fixed lot so that nothing scales up until you decide it should.

Fixed lot, the default:

  InpBalancePerLot = 0

  InpUseStepCompound = false

  InpEntryAmount = the volume you want on every trade

Proportional, recommended from 1200 USD upward:

  InpBalancePerLot = 1200

  InpEntryAmount = 0.01

  InpMaxLotCap = 0.50

Reads as 0.01 lot for every 1200 of balance. A 6000 balance trades 0.05, a 24,000 balance trades 0.20. Lower the divisor for a larger lot and a deeper drawdown, raise it for the opposite. 600 is roughly the aggressive end of what this strategy tolerates.

Step compounding:

  InpBalancePerLot = 0

  InpUseStepCompound = true

  InpCompoundStepPct = 100

  InpCompoundLotAdd = 0.01

Adds 0.01 to the lot every time the balance doubles.

If both InpBalancePerLot and InpUseStepCompound are set, proportional mode takes priority. The status panel on the chart shows which mode is active: fixed, compound or prop 1:N.

BACKTEST, XAUUSD M15, 2022.01 to 2026.08, every tick, real spread

Fixed 0.01 lot, deposit 500 USD:

- net profit 4,602 USD, account grew 9.2 times

- profit factor 2.32

- recovery factor 3.01

- maximum equity drawdown 48%

Proportional sizing, one 0.01 lot per 1200 USD, lot capped at 0.50, deposit 25,000 USD:

- net profit 163,917 USD, account grew 7.6 times

- profit factor 2.06

- recovery factor 2.41

- Sharpe ratio 2.79

- maximum equity drawdown 39%

- 732 trades, 40.2% of them profitable

The proportional run carries the lowest drawdown of the three sizing modes because the risk per trade stays constant as the account grows.

HONEST RISK DISCLOSURE

This is a trend capture tool, not an all weather system. It earns during sustained directional moves in gold and does close to nothing the rest of the time. In testing, 2022 and 2023 were near flat, 2024 was negative, and the bulk of the profit arrived in 2025 and 2026.

Maximum equity drawdown reached 39% with proportional sizing and 48% with a fixed lot. This is not a tuning artefact, it is inherent to the approach: losses accumulate before the trend arrives and the profit comes with it. The only reliable way to reduce drawdown is to increase capital relative to lot size. Doubling the deposit halves the percentage drawdown.

Do not expect a smooth equity curve. If you cannot sit through twelve to eighteen months without meaningful progress, this EA is not for you.

A backtest is not a forward test. Run it on demo before committing capital.

REQUIREMENTS

Symbol XAUUSD, timeframe M15, hedging account, leverage 1:100 or higher.

Recommended deposit 500 to 1000 USD per 0.01 lot.

Maximum spread 60 points. VPS recommended.

For smaller accounts use a cent account with a proportionally scaled lot.

Which sizing mode fits which deposit. Below roughly 1200 USD use the fixed lot: proportional mode would simply return the minimum volume and behave identically. From 1200 USD upward proportional mode starts to scale, and it is the sensible choice from there on, because it keeps the risk per trade constant instead of letting a growing account drift into being under-traded.

A note on account size. At 500 USD the strategy is workable but tight: in testing the worst single month cost about 60% of a 500 USD account, and a start date landing directly in a choppy period could take it close to the broker minimum volume. At 1000 USD and above the same sequence of trades produces roughly half the percentage drawdown and leaves room to sit through a flat year. If you are starting near the lower end, expect wide swings and size down rather than up.

WHAT IS NOT INSIDE

No grid. No martingale. No averaging down. No trailing stop. No breakeven. No hidden recovery mode. No neural network claims. Every position carries a fixed stop loss from the moment it opens.

Past performance does not guarantee future results. Trading gold on margin carries substantial risk of loss. Trade only with capital you can afford to lose. This is a trading tool, not financial advice.