Description
Notus trades gold (XAUUSD) around daily price levels. It places pending stop orders just before levels gold hasn't broken yet, so the order fills the moment price gets there, then manages the trade hard: an early breakeven, a target pulled closer when the trade goes against you, a quick exit when the breakout fails, and a swing-based trailing stop. About 37 trades a year, buys and sells, a stop loss on every trade, no martingale, no grid.
Backtest results
Real tick data from a raw-spread gold account, July 2021 to October 2026, real spread and commission, default settings: 198 trades, $10,000 grew to $17,292, profit factor 2.36, max drawdown $674 (6.7%). By year: 2021 (Jul-Dec) -$198, 2022 +$714, 2023 +$1,377, 2024 +$1,922, 2025 +$3,142, 2026 to October +$334.
On cTrader's own tick data (Spotware demo server, January 2025 to October 28, 2026) the same engine made +$2,802 from $10,000, profit factor 3.40, max balance drawdown 5.1%, over 56 trades. That run cut daily bars at the New York close; this version cuts them at 00:00 UTC, like the main test.
Why the management matters
I kept the moments Notus enters and replaced its direction with a coin flip, 100 times. With the management on, all 100 runs were profitable. With it off, the same coin flips lost on every real-tick run. The entry finds when gold is about to move; the management turns that into money. All of it ships switched on.
What you can set
- Lot mode: fixed lots (scaled with the gold price by default), percent risk per trade, lots per balance, or a drawdown budget; max lots; lots that never decrease after a drawdown
- Allow buy / allow sell
- US jobs report filter: minutes before and after, close open trades, cancel pending orders
- Spread filter and max spread
- Trading hours in UTC, with options to cancel pending orders or close trades outside them
- Label and order comment, so you can run more than one copy
The strategy settings themselves are fixed at the tested values. There is nothing to tune and no broker clock to enter: cTrader runs the cBot on UTC.
Requirements
- XAUUSD, attach to any timeframe (m15 is fine). Notus builds its own daily and hourly bars.
- Raw spread / ECN account recommended. A wider spread costs profit factor but the test stays well above 1 with up to $1.00 of extra cost per trade.
- Default lot: 0.1 per $2,000 of gold price on a $10,000 account (about 0.05 lot at $4,000 gold). Drawdown grows in step with the lot.
- A VPS, so Notus is running when a level breaks.
Backtesting it yourself
Use "Tick data from server". The "m1 bars" mode fills pending orders and stops at the extreme of each minute bar, and on this cBot it reports roughly a third of the tick-data profit. A free backtest-only trial is attached to this listing.
FAQ
Why no trade today? About 37 trades a year; some months have none.
Other symbols? Built and tested on gold only.
Prop firms? Stop loss on every trade and low drawdown at the default lot, but no built-in daily loss limit. Check your firm's rules.
More than one trade at once? Up to 5 when a level fills more than once, each with its own stop loss.
Requires no special permissions (AccessRights.None). No external connections.
Results are from historical backtests, not live trading. Past performance does not guarantee future results. Trading gold on margin carries a high level of risk.
