Backtest results

How we test

MT5 Strategy Tester
The tester built into MetaTrader 5; get the file free and rerun it yourself
Every real tick
Gold (XAUUSD) from QuantDataManager, 'Every tick based on real ticks'
23 years
2003–2026, through gold's rises, falls and sideways years, not just a good stretch
$1,000 · 1%
Starting balance and Default.set's risk per trade, the same preset you get
24 pip
A realistic spread on every trade (0.24 USD), the average we measured on real Standard accounts
Kaze filter
Our own indicator that reads tick volume to check the market is active enough. On for every bot except Asahi

Net profit compounds over 23 years, so it looks big; read it with the profit factor and the drawdown. The figures on this page are for gold (XAUUSD); BTCUSD results are in the BTCUSD tab on each strategy page.

Compare every detail side by side

Backtest overview (XAUUSD) · 2003–2026

Download the monthly summary (CSV)

Growth

Balance ($1,000 start)

Log scale (each gridline is 10×) so every bot fits. These figures compound over 23 years; they aren't what to expect.MT5 Strategy Tester backtests · real tick data from QuantDataManager · XAUUSD · Default.set 1% · 24-pip spread · 2003–2026 · not a forecast

Hoshi grew the most: with 5 entries it trades most often, so its profit compounds fastest, in exchange for deeper drawdowns.
Kumo and Hayate climb in steps: the line jumps when gold runs hard.
Sora and Asahi are steadier. Sora does best when gold trends clearly; Asahi makes modest wins and is out within the day.

Yearly

Yearly result · Hoshi

Year-end balance against the previous year-end; 2003 and 2026 are part-years.MT5 Strategy Tester backtests · real tick data from QuantDataManager · XAUUSD · Default.set 1% · 24-pip spread · 2003–2026 · not a forecast

Hayate was positive in 19 of 24 years; its best was 2013 (+44.4%).
Hoshi was positive in 18 of 24 years, and topped +100% in several years when gold ran hard.
Kumo's best years tend to be big-trend years, such as 2016 (+57.8%) and 2013 (+53.4%).
Every bot has losing years; no bot made money every year.

Drawdown

Down from the peak · Hoshi (Worst -39.0%)

From month-end balances; the MT5 report's equity drawdown, measured during trades, is deeper.MT5 Strategy Tester backtests · real tick data from QuantDataManager · XAUUSD · Default.set 1% · 24-pip spread · 2003–2026 · not a forecast

Below 0 is time the balance spent under its previous peak, waiting for a new high.
Deeper than you can accept? Lower RiskPercent: the depth falls roughly in proportion.

Sets

Hoshi + Asahi + Hayate (Worst -20.9%)

The set splits the risk so the total equals one bot at 1% (e.g. 4 bots at 0.25% each). Thin lines are its bots alone.MT5 Strategy Tester backtests · real tick data from QuantDataManager · XAUUSD · Default.set 1% · 24-pip spread · 2003–2026 · not a forecast

Kumo + Sora + Asahi + Hayate: smoother than any of its bots. Its deepest month-end drawdown was about −14.9%, against −19.8% to −31.6% for its bots alone.
Hoshi + Asahi + Hayate: fast growth from Hoshi, with Asahi and Hayate smoothing out Hoshi's drawdowns.

Try your own numbers

Pick the markets and bots, set a risk per trade for each and enter your balance, then see how the backtests would have played out.

CLOUDPips Hoshi

Breakout + trend following · 5 entries in one bot

Hoshi v1.1 backtest with Default.set, XAUUSD 2003–2026, realistic 24-pip spread: results table and balance curve

Preset	Period (XAUUSD)	Net profit	PF	Equity DD
Default 1%	2003–2026	+87,709%	1.67	47.92%
Aggressive 2%	2003–2026	+5,288,160%	1.82	72.43%
Default 1%	2024–2026	+367%	1.74	20.85%
Aggressive 2%	2024–2026	+2,563%	1.76	31.78%

Default and Aggressive use the same settings at 1% and 2% risk per trade.

See this strategy's backtests Get the free backtest file

CLOUDPips Kumo

Breakout · rides the move with a trailing stop

Kumo v1.2 backtest with Default.set, XAUUSD 2003–2026, realistic 24-pip spread: results table and balance curve

Preset	Period (XAUUSD)	Net profit	PF	Equity DD
Default 1%	2003–2026	+749%	1.40	22.51%
Aggressive 2%	2003–2026	+6,472%	1.50	44.81%
Default 1%	2024–2026	+159%	2.43	9.21%
Aggressive 2%	2024–2026	+234%	2.02	15.18%

Default and Aggressive use the same settings at 1% and 2% risk per trade.

See this strategy's backtests Get the free backtest file

CLOUDPips Sora

Trend following · enters only with the higher-timeframe trend

Sora v1.6 backtest with Default.set, XAUUSD 2003–2026, realistic 24-pip spread: results table and balance curve

Preset	Period (XAUUSD)	Net profit	PF	Equity DD
Default 1%	2003–2026	+124%	1.13	34.27%
Aggressive 2%	2003–2026	+233%	1.19	63.09%
Default 1%	2024–2026	+96%	1.32	27.54%
Aggressive 2%	2024–2026	+280%	1.64	31.97%

Default and Aggressive use the same settings at 1% and 2% risk per trade.

See this strategy's backtests Get the free backtest file

CLOUDPips Asahi

Day trade · follows the Asian session's direction, no overnight positions

Asahi v1.0 backtest with Default.set, XAUUSD 2003–2026, realistic 24-pip spread: results table and balance curve

Preset	Period (XAUUSD)	Net profit	PF	Equity DD
Default 1%	2003–2026	+199%	1.22	24.94%
Aggressive 2%	2003–2026	+274%	1.17	28.33%
Default 1%	2024–2026	+129%	1.42	25.14%
Aggressive 2%	2024–2026	+132%	1.43	24.30%

Default and Aggressive use the same settings at 1% and 2% risk per trade.

See this strategy's backtests Get the free backtest file

CLOUDPips Hayate

Daily breakout · trades only the days gold breaks its range with the big trend

Hayate v1.0 backtest with Default.set, XAUUSD 2003–2026, realistic 24-pip spread: results table and balance curve

Preset	Period (XAUUSD)	Net profit	PF	Equity DD
Default 1%	2003–2026	+1,038%	1.31	22.72%
Aggressive 2%	2003–2026	+10,921%	1.30	45.92%
Default 1%	2024–2026	+120%	1.39	28.20%
Aggressive 2%	2024–2026	+133%	1.40	31.83%

Default and Aggressive use the same settings at 1% and 2% risk per trade.

See this strategy's backtests Get the free backtest file

How to read these numbers

Every test costs each trade a realistic 24-pip (0.24 USD) spread, the typical spread of a standard account. Gold trended clearly in 2024–2026; read it next to 2003–2026 to see both the good stretches and the hard ones. Past results don't guarantee future results.

Reading the backtest numbers How to backtest in MT5

Past performance does not guarantee future results. Backtests cannot fully reproduce live conditions such as wider spreads around news, slippage or lost connections.