Skip to main content
ForexNews24

Backtests of trading systems

12 hypothesis studies. For 8 of them the signal was run on real EUR/USD quotes — the metrics are real and reproducible. The rest need data the portal does not have and are marked honestly.

A note on data: the portal has its own reproducible history only for EUR/USD. Hypotheses are checked on it, not on the instrument in the title. Hypotheses that need several instruments, intraday or tick data are not reproduced — a discussion of the limitations is given instead of a result.

RSI strategy on EUR/USD over 10 years

-4.3%
CAGR on data
7.3%
drawdown
Hypothesis: EUR/USD · 2015–2025

A recovery of RSI(14) from the oversold zone (<30) on EUR/USD is statistically more often followed by an upward price move over a 20–40 bar H1 horizon than a random entry.

MACD Crossover on GBP/USD, daily timeframe

-8.5%
CAGR on data
12.3%
drawdown
Hypothesis: GBP/USD · 2013–2025

A cross of the MACD line and the signal line on the daily GBP/USD timeframe leads to a trend continuation more often than a random walk in price would suggest.

Bollinger Squeeze on the BTC/USD crypto pair

+8.1%
CAGR on data
2.8%
drawdown
Hypothesis: BTC/USD · 2018–2025

A narrowing of the Bollinger Bands on BTC/USD (a Bollinger Squeeze) statistically precedes a price move of above-average amplitude over the following 10 bars.

Grid strategy on USD/JPY: the risk of a managed grid

+4%
CAGR on data
2.3%
drawdown
Hypothesis: USD/JPY · 2016–2025

A grid system with an ATR-linked step shows a more manageable drawdown on USD/JPY than a fixed-step grid under identical market conditions.

VWAP Fade on the S&P 500 CFD index

+11.3%
CAGR on data
3%
drawdown
Hypothesis: US500 · 2019–2025

A deviation of the US500 price from VWAP by more than 1σ statistically more often reverts to VWAP within the trading session than it continues in the same direction.

Ichimoku Cloud Break on gold (XAU/USD)

+1%
CAGR on data
4.3%
drawdown
Hypothesis: XAU/USD · 2014–2025

A price break above the Ichimoku Cloud on XAU/USD followed by a Tenkan/Kijun cross yields a positive expected value over a 10-day horizon.

ADX filter for trend entries on AUD/USD

-7.7%
CAGR on data
8.8%
drawdown
Hypothesis: AUD/USD · 2017–2025

Filtering trend entries by ADX(14) > 25 on AUD/USD reduces the share of false signals compared with the same system without the filter.

Triangular arbitrage on major pairs

not reproducible
Hypothesis: EUR/GBP/USD · 2020–2025

The divergence between the synthetic and real EUR/GBP cross rate exceeds transaction costs often enough for triangular arbitrage to be consistently profitable.

News volatility around the NFP

not reproducible
Hypothesis: EUR/USD · 2016–2025

EUR/USD volatility in the 30-minute window around the NFP release is statistically significantly higher than the average volatility in the historical data.

ATR Breakout on oil (WTI)

-0.2%
CAGR on data
6.8%
drawdown
Hypothesis: WTI · 2015–2025

An ATR range breakout on WTI confirmed by volume yields a positive expected value during periods of elevated oil-market volatility.

Mean reversion of the EUR/USD × GBP/USD pair

not reproducible
Hypothesis: EUR/USD, GBP/USD · 2018–2025

A deviation of the EUR/USD × GBP/USD spread from its historical mean, measured in z-score units, statistically reverts to the mean within 20 trading days.

Session Open Range on the NASDAQ-100 CFD

not reproducible
Hypothesis: US100 · 2019–2025

A breakout of the first-30-minute session range on US100 confirmed by volume yields a positive expected value within the trading day.

From research to application

In our Allocation product we implemented these algorithms with all the nuances covered across the portal.

Learn about Allocation