ICFX Aurora Gold MT5
- Experts
- Versione: 2.0
- Aggiornato: 20 novembre 2026
- Attivazioni: 10
| The first 10 copies are sold at launch price $299. After every 10 copies sold, the price rises by $100 until it reaches the final price of $699. |
A multi-strategy gold portfolio, by Iceland Trader – ICFX
Aurora Gold is not a single strategy. It is a portfolio of 14 independent, rule-based gold strategies in one Expert Advisor. 11 are enabled by default, chosen together for the smoothest combined equity curve.
What makes it different
- It builds its own charts. Heikin-Ashi, smoothed Heikin-Ashi, Kagi, percentage Kagi, Mean Renko, efficiency-ratio Renko and Darvas charts are built inside the EA from MT5 price history. Each strategy sees gold through a different noise filter.
- No offline charts needed. Attach the EA to one regular XAUUSD M1 chart. No custom charts, extra indicators or tools are required.
- Built for live ticks. Renko charts start from the same brick state as the research backtest, every new brick is evaluated (also when several form in one minute), entries wait while the spread is abnormally wide, and temporary broker errors (requotes, price changes, off quotes, timeouts) are retried.
- Methodology diversification. Volatility breakout, retracement level, Kagi reversal, Kagi trend turn, channel breakout, compression breakout, pivot trend breakout, regression signal, moving-average group, range thrust and Kagi swing structure. Their trades rarely coincide, so they smooth each other's equity.
- Individually optimised. Each strategy has its own filters (seasonality, currency strength, trend slope, ADX regime, news days) and its own exits (split take-profits, break-even, trailing and chandelier stops, time stops).
- Volatility-adaptive. Stops and targets scale with volatility. For six strategies the stop is widened, or switched to a volatility (ATR) stop, when daily volatility jumps far above its normal level.
- On-chart information panel. Status, risk mode, balance, equity, open and closed profit, next high-impact USD news, and a table of all 14 strategies with trades, closed profit, profit per trade and open lots.
- No martingale. No grid. No averaging. Hard stop loss on every trade.
MT5 Strategy Tester result
ICFX Aurora Gold EA v2.00 · IC Markets Raw · XAUUSD M1 · 1-minute OHLC · 99% history quality · Jan 2024 – Nov 2026 · $100,000 initial deposit · Auto risk mode
| Metric | Default (11 strategies on) | All 14 strategies on |
|---|---|---|
| Net profit | +$275,447 (+275%) | +$377,176 (+377%) |
| Profit factor | 2.29 | 2.07 |
| Win rate | 71.9% (498 trades) | 70.1% (648 trades) |
| Long / short trades won | 73.8% / 66.9% | 72.4% / 64.1% |
| Maximum equity drawdown | 13.2% | 12.7% |
| Maximum balance drawdown | 8.8% | 9.1% |
| Recovery factor | 6.49 | 7.15 |
| Sharpe ratio | 3.12 | 3.11 |
| Equity linearity (LR correlation) | 0.98 | 0.97 |
| Maximum consecutive losses | 11 | 6 |
Historical backtest 2004–2026 (research with Claude)
Aurora Gold is the result of an extensive backtest project carried out with Claude (Anthropic's AI):
- Thousands of strategy and chart-type combinations were tested on 1-minute gold data from 2003 to 2026, with real spreads and commission.
- Candidates were ranked by equity-curve quality first, then maximum drawdown, then profit factor and win rate.
- Each selected strategy was fully optimised (about 60 filters, stop sizes, shifted entries, split take-profits, break-even, trailing and volatility-staged stops), and no filter was allowed to remove more than 40% of a strategy's trades.
- Seasonal tables were learned only from data before November 2023, so 2023–2026 is out-of-sample for them.
- Robustness checks included spread ×2/×3, extra slippage, news filters and Monte Carlo analysis.
Results in R (1R = the amount risked per trade):
| Metric | 11 strategies (default) | 12 (with RangeThrust) | All 14 |
|---|---|---|---|
| Trades | 2,317 | 2,941 | 3,276 |
| Win rate | 60.0% | 60.1% | 59.9% |
| Profit factor | 2.10 | 1.96 | 1.94 |
| Equity curve R² | 0.997 | 0.997 | 0.996 |
| Max drawdown | 11.9R | 14.0R | 15.8R |
| Monte Carlo 95% drawdown | 14.5R | 15.8R | 16.2R |
| Return per month | +2.95R | +3.42R | +3.81R |
| Profitable years | 23 of 23 | 23 of 23 | 23 of 23 |
| Profitable months | 71% | 72% | 71% |
| Profit factor 2023–2026 | 2.35 | 2.17 | 2.19 |
The 11-strategy default was profitable in 23 of 23 years, including gold's 2013–2015 bear market.
The strategies
2004–2026 backtest, each strategy alone
| # | Strategy | Chart | Method | Trades | Win % | PF | Default |
|---|---|---|---|---|---|---|---|
| S01 | VolBand | Heikin-Ashi H4 | Volatility band breakout | 437 | 57.9 | 1.90 | On |
| S02 | PrevWeek382 | Candles H4 | Previous-week 38.2% retracement | 199 | 57.3 | 2.39 | On |
| S03 | KagiML | Candles H4 | Kagi multi-level breakout | 131 | 55.7 | 2.41 | On |
| S04 | Channel55 | Darvas H1 | 55-bar channel breakout | 426 | 60.8 | 1.69 | On |
| S05 | KagiHS | Smoothed Heikin-Ashi H1 | Kagi three-peak reversal | 84 | 54.8 | 2.31 | On |
| S06 | KagiYang | Darvas H1 | Kagi trend turn | 111 | 55.0 | 2.58 | On |
| S07 | NR4 | Mean Renko | Compression breakout | 252 | 62.7 | 1.87 | On |
| S08 | KagiRS2 | Smoothed Heikin-Ashi H1 | Kagi shoulders (wide) | 77 | 64.9 | 2.14 | On |
| S09 | Pivot | Efficiency-ratio Renko | Pivot trend breakout | 304 | 63.8 | 2.18 | On |
| S10 | Bary | Candles H4 | Regression signal | 122 | 59.8 | 2.34 | On |
| S11 | RangeThrust | Heikin-Ashi H4 | Dual Thrust range breakout | 624 | 60.6 | 1.52 | Off (new) |
| S12 | KagiRS1 | Heikin-Ashi H4 | Kagi shoulders (tight) | 141 | 65.2 | 1.99 | Off |
| S13 | MAGroup | Candles H4 | Moving-average group | 174 | 62.6 | 2.20 | On |
| S14 | KagiPct | Percentage Kagi | Kagi rising shoulders | 194 | 52.6 | 1.66 | Off |
Strategies S11, S12 and S14 are profitable on their own. They are off by default because they increased the combined drawdown more than the return, or repeated the method of a default strategy. Each can be switched on individually.
Risk modes
Auto (default) sizes risk so that the portfolio's Monte Carlo drawdown matches the target (16% by default, adjustable).
| Mode | Target drawdown |
|---|---|
| Very conservative | 8% |
| Conservative | 12% |
| Auto / Moderate (default) | 16% |
| High moderate | 20% |
| Aggressive | 25% |
| Very aggressive | 30% |
| Super aggressive | 40% |
| Mega aggressive | 50% |
Also available: risk % of balance, risk % of equity, fixed lot. Targets come from the research backtest. They are not a guaranteed maximum.
Protection
- Daily loss limit, maximum drawdown block, maximum open strategies, maximum spread filter, abnormal-spread guard, maximum holding time.
- Optional high-impact news filter using the MT5 economic calendar (live trading only).
- Minimum-lot protection for small accounts: a trade is skipped if 0.01 lot would exceed the risk cap, and a separately optimised single-target exit is used when a position cannot be split.
Requirements
| Symbol | XAUUSD (gold) |
| Chart | One regular XAUUSD M1 chart. No offline or custom charts needed. |
| Account | Hedging. On a netting account the EA trades only one position at a time. |
| Deposit | $3,000 or more recommended. Minimum $800, where trades use the 0.01 minimum lot and risk per trade is higher. |
| Broker | Low-spread (ECN/Raw) account. A VPS is recommended. |
| Settings | Set "Max bars in chart" to Unlimited (Tools → Options → Charts). |
Backtesting in the MT5 Strategy Tester
The EA builds its charts from price history before the test start, as it does live. For a correct backtest:
- Download the broker's full XAUUSD history first (View → Symbols → Bars → XAUUSD, M1, from the earliest date → Request).
- Leave at least 3 years of history before the test start date. The H4 charts use about 4,000 bars (about 2.7 years) and the H1 charts about 8,000 bars.
- Set "Max bars in chart" to Unlimited before the test.
- Start dates from January 2015 onward use the embedded Renko starting states, so the Renko strategies match the research backtest. Earlier start dates still work, but the first Renko bricks can differ.
- Daily gold history is embedded in the EA, so the volatility-staged stops work even on brokers with short daily history.
- Use "1 minute OHLC" or "Every tick based on real ticks". The economic calendar is not available in the tester, so the news filter only works live.
Main inputs
- Risk mode and drawdown target
- Daily loss limit %
- Max drawdown %
- Max open strategies
- Max spread and spread guard
- News filter
- S01–S14 on/off and weight
- Information panel on/off and position
- Magic number base
