Thirty Four Basket MR Buy
- 专家
- 版本: 4.23
- 激活: 5
When an established long-term uptrend suffers a panic sell-off, the EA buys a basket of assets and takes profit once the panic repairs. Average holding time is a few days. No manual intervention is required.
Overview
Cross-sectional diversification across a 19 to 34 symbol basket: extreme moves in any single symbol are diluted by the normal behaviour of the others, and different asset classes follow different paths, which keeps portfolio level drawdown contained. More symbols also means more trading opportunities.
Every trade carries built-in take-profit and stop-loss logic.
The rules have been tested over 6, 9, 20 and 36 year horizons. Across those windows the equity curve grows steadily, drawdowns stay shallow and recovery is fast. Over the 8 year live-account basis backtest the profit factor holds above 1.6, the payoff ratio is close to 1:1 and the win rate is around 70 percent.
The system uses no future functions, no repainting and no overfitting.
Setup
Attach the EA to the daily chart of any single symbol in the basket, for example XAUUSD or US500, and set your preferred maximum drawdown: 10 percent for low risk, 20 to 30 percent for medium, 40 percent for high. All sizing and exit logic is handled internally.
Symbols
34 symbols across six asset classes, all traded on the daily timeframe.
- Indices — US500, USTEC, US30, DE30, UK100, JP225, FR40
- US equities — META, V, MSFT, AMT, GILD, CSCO, MA, AMD, NFLX, CHTR, VRTX, TSM, ORCL, TSLA, ADBE
- Precious metals — XAUUSD, XPTUSD
- Energy — UKOIL
- Base metals — XCUUSD
- Crypto — BTCUSD, ETHUSD, SOLUSD, XRPUSD, ADAUSD, DOGEUSD, LTCUSD, LINKUSD
Backtest Setting
Backtest Results
Core performance
Basis: native Exness daily bars, 2019.01.23 to 2026.10.22, 1,431 trades, with each symbol’s measured spread deducted trade by trade. Starting equity 10,000 USD.
- Win rate — 69.3 percent over 1,431 trades
- Profit factor — 1.65
- Average trades per month — 15.6
- Profitable months — 67 out of 92 (73 percent)
[[ IMAGE 1 — chart-equity-curve.png ]]
Three notional sizing tiers, 20, 30 and 40 percent of account equity per trade. Starting from 10,000 USD the 20 percent tier ends at 59,730 USD, the 30 percent tier at 137,564 USD and the 40 percent tier at 303,766 USD. Logarithmic vertical axis.
Year by year, 20 percent notional
- 2019 — 73 trades · +1,600 USD · PF 2.64 · win rate 74.0 percent · max drawdown in year 3.6 percent
- 2020 — 96 trades · +2,009 USD · PF 1.78 · win rate 71.9 percent · max drawdown 13.7 percent
- 2021 — 230 trades · +3,120 USD · PF 1.42 · win rate 73.5 percent · max drawdown 17.2 percent
- 2022 — 80 trades · −1,503 USD · PF 0.70 · win rate 55.0 percent · max drawdown 13.7 percent
- 2023 — 265 trades · +3,349 USD · PF 1.37 · win rate 69.8 percent · max drawdown 12.6 percent
- 2024 — 287 trades · +6,493 USD · PF 1.58 · win rate 65.5 percent · max drawdown 11.6 percent
- 2025 — 230 trades · +18,213 USD · PF 2.45 · win rate 72.2 percent · max drawdown 8.9 percent
- 2026 — 170 trades · +16,449 USD · PF 2.35 · win rate 68.2 percent · max drawdown 6.2 percent
Seven profitable years out of eight, one losing year. 2022 was the global stocks-and-bonds sell-off: profit factor 0.70, win rate 55.0 percent, and a loss of 1,503 USD on the 20 percent tier. Every other year is positive.
[[ IMAGE 2 — chart-annual-pl.png ]]
Annual profit and loss distribution. 2022 is the only losing year. Returns scale up from 2024 onward, tied to higher market volatility and a larger number of oversold setups.
Monthly breakdown, 20 percent notional
94 calendar months in total: 67 profitable, 25 losing, and 2 months with no closed trades (2019.02 and 2020.04).
- Profitable months — 67 out of 94 (71.3 percent)
- Average monthly return — +2.03 percent
- Best month — +15.4 percent (2025.08)
- Worst month — −13.3 percent (2020.03)
- Longest winning streak — 10 months
- Longest losing streak — 3 months
[[ IMAGE 3 — chart-monthly-distribution.png ]]
The vast majority of months land in the 0 to 5 percent range. Losing months cluster in 2020.03, across 2022 and in mid 2021. Note the symmetry between August 2025 at +15.4 percent and March 2020 at minus 13.3 percent: month level swings alone span more than a threefold range.
Five risk tiers, 5 to 40 percent
Identical signals and identical 1,431 trades, only the share of equity put into each trade changes. The larger the size, the steeper the compounding curve and the proportionally larger the drawdown. This section is the exchange rate between return and risk.
- 5 percent — final equity 15,860 USD · total return +59 percent · annualised 6.1 percent · max drawdown minus 4.1 percent · worst month minus 3.5 percent
- 10 percent — final equity 24,916 USD · total return +149 percent · annualised 12.5 percent · max drawdown minus 8.2 percent · worst month minus 6.8 percent
- 20 percent — final equity 59,730 USD · total return +497 percent · annualised 25.9 percent · max drawdown minus 16.4 percent · worst month minus 13.3 percent
- 30 percent — final equity 137,564 USD · total return +1276 percent · annualised 40.2 percent · max drawdown minus 24.6 percent · worst month minus 19.6 percent
- 40 percent — final equity 303,766 USD · total return +2938 percent · annualised 55.3 percent · max drawdown minus 32.7 percent · worst month minus 25.5 percent
[[ IMAGE 4 — chart-risk-tiers.png ]]
The five curves are identical in shape and fan out proportionally. The 5 percent tier turns 10,000 USD into 15,860 USD over 7.7 years, the 40 percent tier into 303,766 USD, paid for by a maximum drawdown widening from minus 4.1 to minus 32.7 percent. Logarithmic vertical axis.
Annual returns by tier
- 5 percent — 2019 +3.9 · 2020 +4.2 · 2021 +5.8 · 2022 minus 2.2 · 2023 +5.3 · 2024 +7.9 · 2025 +14.9 · 2026 +8.5
- 10 percent — 2019 +7.8 · 2020 +8.5 · 2021 +11.6 · 2022 minus 4.5 · 2023 +10.8 · 2024 +16.4 · 2025 +31.8 · 2026 +17.6
- 20 percent — 2019 +16.0 · 2020 +17.3 · 2021 +22.9 · 2022 minus 9.0 · 2023 +22.0 · 2024 +35.0 · 2025 +72.7 · 2026 +38.0
- 30 percent — 2019 +24.5 · 2020 +26.4 · 2021 +33.6 · 2022 minus 13.6 · 2023 +33.6 · 2024 +55.9 · 2025 +125.0 · 2026 +61.5
- 40 percent — 2019 +33.2 · 2020 +35.8 · 2021 +42.9 · 2022 minus 18.1 · 2023 +45.6 · 2024 +79.4 · 2025 +191.7 · 2026 +88.5
Values in percent, by year-end equity. Seven profitable years out of eight, with 2022 the only losing year. All five tiers move in the same direction, and the depth of a losing year scales with position size as well.
Top 10 symbols by profit contribution
Below are the ten best performing symbols in the basket, ranked by their contribution to total return. Under the same rules the distribution is never even, yet every symbol earns its place. That is the point of pool diversification: some symbols drive profit, some spread risk, some preserve capital in a crisis, and the whole never depends on any single one. When the macro environment turns against us, part of the basket avoids or hedges the risk; when it turns favourable, part of it pushes for more profit. Playing to each symbol’s different character also raises overall trade frequency.
- AMDm — 49 trades · win rate 77.6 percent · PF 2.10 · average per trade +1.78 · cumulative +87.0 · contribution 9.3 percent
- BTCUSDm — 63 trades · win rate 73.0 percent · PF 1.92 · average per trade +1.30 · cumulative +81.9 · contribution 8.8 percent
- METAm — 42 trades · win rate 69.0 percent · PF 5.29 · average per trade +1.66 · cumulative +69.7 · contribution 7.5 percent
- LTCUSDm — 42 trades · win rate 66.7 percent · PF 1.57 · average per trade +1.65 · cumulative +69.2 · contribution 7.4 percent
- JP225m — 64 trades · win rate 81.2 percent · PF 4.08 · average per trade +0.93 · cumulative +59.5 · contribution 6.4 percent
- VRTXm — 43 trades · win rate 76.7 percent · PF 4.07 · average per trade +1.35 · cumulative +57.8 · contribution 6.2 percent
- ETHUSDm — 56 trades · win rate 69.6 percent · PF 1.46 · average per trade +0.98 · cumulative +54.9 · contribution 5.9 percent
- XRPUSDm — 47 trades · win rate 68.1 percent · PF 1.39 · average per trade +1.08 · cumulative +50.7 · contribution 5.4 percent
- USTECm — 77 trades · win rate 76.6 percent · PF 2.36 · average per trade +0.60 · cumulative +46.4 · contribution 5.0 percent
- TSMm — 33 trades · win rate 72.7 percent · PF 2.59 · average per trade +1.31 · cumulative +43.4 · contribution 4.7 percent
MT5 Strategy Tester record, 17 symbols, 2021 to 2026
This is raw output from the MetaTrader 5 Strategy Tester, not a simulation: real spreads, real swaps, 1:100 leverage, tick by tick matching. The run enabled only 17 symbols at the 20 percent notional tier, from 2021.01 to 2026.10.
[[ IMAGE 5 — tester-report.png ]]
Result: 441 trades, profit factor 1.72, win rate 64.85 percent, net profit 15,331 USD on a 10,000 USD deposit, maximum equity drawdown 25.87 percent, Sharpe ratio 1.09, history quality 98 percent.
Basis comparison at the same 20 percent notional tier: from 2021 the 34 symbol basis returns +338.9 percent with a minus 16.4 percent maximum drawdown; the same rules with only 17 symbols return +153.3 percent with a minus 25.9 percent drawdown. Halving the basket halves diversification and immediately doubles the drawdown. Neither figure is wrong, they simply reflect a different number of symbols sharing the risk.
Ultra long horizon validation, 1990 to 2026
To check whether the rules only work over the last eight years, a 19 symbol basket was run back to 2000 and a core basket back to 1990: 2,960 trades in total.
[[ IMAGE 6 — chart-long-horizon.png ]]
- 2 percent notional — 2,960 trades · final equity 14,266 USD · total return 43 percent · annualised 1.3 percent · max drawdown minus 3.5 percent
- 5 percent notional — final equity 24,139 USD · total return 141 percent · annualised 3.3 percent · max drawdown minus 8.6 percent
- 10 percent notional — final equity 56,905 USD · total return 469 percent · annualised 6.7 percent · max drawdown minus 16.8 percent
- 30 percent notional — final equity 1,382,007 USD · total return 13720 percent · annualised 20.2 percent · max drawdown minus 45.1 percent
Monte Carlo stress test
The 1,431 real trade results are resampled with replacement 1,000 times, to see what the same strategy delivers under different luck.
[[ IMAGE 7 — chart-monte-carlo.png ]]
1,000 resampled paths at the 30 percent notional tier. Dark blue is the median path, the light band is the 5th to 95th percentile range.
- 20 percent — median final 60,769 USD · 5th percentile 33,884 USD · 95th percentile 110,299 USD · median drawdown 13.9 percent · worst drawdown 41.9 percent · ruin probability 0 percent
- 30 percent — median final 143,366 USD · 5th percentile 59,114 USD · 95th percentile 351,387 USD · median drawdown 20.5 percent · worst drawdown 56.6 percent · ruin probability 0 percent
- 40 percent — median final 326,427 USD · 5th percentile 99,185 USD · 95th percentile 1,085,340 USD · median drawdown 26.9 percent · worst drawdown 68.0 percent · ruin probability 0 percent
Across 1,000 resamples none of the three tiers blew up; the worst path at the 40 percent notional tier drew down 68 percent. This table shows the distribution of the same strategy under different luck. It is not a forecast.
Who it suits
The EA suits users with idle capital. It is best used as one allocation inside a broader portfolio, for those aiming at 30 to 50 percent annualised over the long run. Its edge is steady compounding across multiple market cycles.
Long-term stability is not measured by a few days or weeks of testing, but by compounding in months and years. In the 2021 to 2026 report, 73 percent of the 92 months were profitable. That is how this strategy performs: a payoff ratio close to 1:1, a win rate near 70 percent, and 15 to 25 trades per month.
Across several decades of testing, the rules passed through the 2000 dot-com bubble, the 2008 financial crisis and the 2020 pandemic without breaking, recovering quickly from shallow drawdowns and returning to a higher annualised rate.
Questions and answers
My MT5 backtest does not match the results in this description.
Symbols, historical data, spreads and swap charges differ between brokers, so long horizon MT5 results are never exactly identical across users, although they generally converge from 2025 onward. This is not overfitting and there is no future function involved. Most brokers simply do not carry complete symbols and complete history before 2025. If your backtest still looks poor, send me the report through the MQL5 message system: I will compare it against mine, which is built on a more complete symbol set and dataset, and tell you exactly why they differ.
My broker does not offer all the symbols. Can I still use the EA?
Yes. The number of symbols determines trade frequency: a single symbol only produces a handful of setups per year, while more symbols means more opportunities. The symbol pool also determines how risk is spread, since it covers US indices, other indices, US equities, industrial metals, precious metals, crypto and energy. The basket should cover as many different pools as possible to spread risk, and include as many symbols as possible to generate profit.
Support and risk note
Please use the product comments section or the MQL5 message system for all questions about settings, set files, activations and backtest reports. I answer there directly.
Backtest results are historical and do not guarantee future performance. Position sizing is selectable from 5 to 40 percent of equity per trade, and drawdown scales with it: at the 40 percent tier the historical maximum drawdown reached 32.7 percent, and the worst Monte Carlo path reached 68 percent. Trade only with capital you can afford to lose.
