The London breakout on gold

Micro Gold (MGC) · 5-minute bars · 2016 to 2026 · Breakout · Widely taught, from currency markets

Depends on the period: it lost in sample (profit factor 0.90) and made money out of sample (1.29).

Profit factor, out of sample: 1.29
Average trade, out of sample: +0.08R
Win rate, out of sample: 56%
Trades, 10 years: 1,696
Max drawdown: $13,651
Pass odds, median plan: 2.5%

Hypothetical results. These results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown.

SharePost on X LinkedIn Reddit

Abstract

We backtested the london breakout rules on Micro Gold (MGC) 5-minute bars from 2016-01-11 to 2026-10-30: 1,696 trades, with commission and a tick of slippage on every market and stop fill. In sample (to 2023-08-03) it made −0.05R per trade with a profit factor of 0.90; out of sample (from 2023-08-04) +0.08R per trade and 1.29 over 488 trades. 25 of the 25 parameter settings we tried were profitable out of sample. Sized at $250 of risk a trade, simulated 50K evaluations passed 0% to 6% of the time across 36 tracked plans.

1Hypothesis

Mark the Asia session's high and low on gold. At the London open, buy a break above the range or sell a break below it, with the stop at the other side and a target of one range.

If London's open brings the volume that ends Asia's quiet range, the first break of the range should run at least one range-width more often than it reverses.

2Data

Instrument: Micro Gold (MGC), traded on the GC price series (GC.v.0: the most traded contract, rolled on volume). CME Globex via Databento.
Period: 2016-01-11 to 2026-10-30: 2,791 sessions, 3,771,561 one-minute bars, built into 761,875 5-minute bars around the clock.
Costs: $0.75 commission a contract each way and 1 tick ($1.00) of slippage on every market and stop fill. Limit orders fill when price touches them; market orders fill at the next bar's open and stops at their price, each with one tick of slippage; when a bar touches both the stop and the target, the stop counts.
Position size: About $250 of risk a trade: micro contracts, rounded down, at least one. R is a trade's net P&L divided by its risk at entry.
Out of sample: The last 30% of weekdays: 2023-08-04 on (846 days), after 1,974 in-sample days. The rules were fixed before either side was run.

3Method

Our reading of the publicly described rules; the creator may trade it differently. The rules, exactly as the code runs them:

The Asia range is the high and low of 20:00 to 02:00 ET (the 5-minute bars starting 20:00 through 01:55).
At 03:00 ET, the London open, a buy stop goes in one tick above the Asia high and a sell stop one tick below the Asia low. When one fills, the other is cancelled. If price is already outside the range at 03:00, there is no trade that day.
Stop: the other side of the Asia range.
Target: target_r (1) times the risk, about one range width.
Unfilled orders expire at 08:00 ET; an open trade is closed at 11:00 ET. One trade a day.
Size: about risk_usd dollars of risk a trade (micro contracts, rounded down, at least 1 and at most max_contracts).

bot.py, the file the engine ran

1"""London breakout on Micro Gold (MGC) 5-minute bars, New York time.2 3Our reading of the commonly taught rules; traders who teach it may trade it differently.4 5Rules:6  1. The Asia range is the high and low of 20:00 to 02:00 ET (the 5-minute bars starting 20:007     through 01:55).8  2. At 03:00 ET, the London open, a buy stop goes in one tick above the Asia high and a sell stop one9     tick below the Asia low. When one fills, the other is cancelled. If price is already outside the10     range at 03:00, there is no trade that day.11  3. Stop: the other side of the Asia range.12  4. Target: `target_r` (1) times the risk, about one range width.13  5. Unfilled orders expire at 08:00 ET; an open trade is closed at 11:00 ET. One trade a day.14  6. Size: about `risk_usd` dollars of risk a trade (micro contracts, rounded down, at least 1 and at15     most `max_contracts`).16"""17import qy18 19 20class Bot(qy.Strategy):21    """Break of the Asia session's range after the London open."""22 23    symbol = "MGC"24    timeframe = "5m"25    session = None26    params = {"asia_start": 20, "asia_end": 2, "target_r": 1.0, "risk_usd": 250, "max_contracts": 50}27    max_position = 5028    entry_window = ("03:00", "08:00")29    flatten_time = "11:00"30    overnight = False31 32    def setup(self):33        self.hi = self.lo = None34        self.placed = False35 36    def _size(self, risk_points):37        per_contract = risk_points * self.point_value38        n = int(self.p.risk_usd // per_contract) if per_contract > 0 else 139        return max(1, min(self.p.max_contracts, n))40 41    def on_day_start(self, day):42        self.hi = self.lo = None43        self.placed = False44 45    def on_fill(self, fill):46        if fill.kind == "entry":47            self.cancel_all()48 49    def _in_asia(self, hour):50        a, b = self.p.asia_start, self.p.asia_end51        return a <= hour or hour < b if a > b else a <= hour < b52 53    def on_bar(self, bar):54        start = bar.dt.hour * 60 + bar.dt.minute55        if self._in_asia(bar.dt.hour):56            self.hi = bar.high if self.hi is None else max(self.hi, bar.high)57            self.lo = bar.low if self.lo is None else min(self.lo, bar.low)58            return59        # The bar that closes at 03:00 ET: place both orders once.60        if start != 175 or self.placed or self.hi is None or self.hi <= self.lo:61            return62        self.placed = True63        if not self.lo < bar.close < self.hi:64            return65        tick = self.tick_size66        buy_at, sell_at = self.hi + tick, self.lo - tick67        width_pct = (self.hi - self.lo) / bar.close * 10068        risk = buy_at - self.lo69        qty = self._size(risk)70        self.buy(qty, stop=buy_at, stop_price=self.lo, target_r=self.p.target_r, tag=f"London long | f={width_pct:.4f}")71        self.sell(qty, stop=sell_at, stop_price=self.hi, target_r=self.p.target_r, tag=f"London short | f={width_pct:.4f}")

4Results

Table 1. In sample (2016-01-11 to 2023-08-03) against out of sample (2023-08-04 to 2026-10-30), after costs.
Statistic	In sample	Out of sample	All
Trades	1,208	488	1,696
Net P&L	−$12,015	+$12,862	+$847
Profit factor	0.90	1.29	1.01
Average trade	−0.05R	+0.08R	−0.01R
95% interval of the average	−0.10 to 0.00	0.01 to 0.15	−0.05 to 0.03
t-statistic of trade R	−1.81	2.21	−0.44
Win rate	48%	56%	50%
Average win / average loss	0.97	1.03	1.00
Max drawdown	$13,651	$4,231	$13,651
Sharpe ratio (daily, annualized)	−0.80	1.63	0.04
Longest losing streak	7	6	7
Long / short profit factor	0.82 / 0.98	1.10 / 1.50	0.90 / 1.12
Average holding time	280 min	301 min	286 min
Commission paid	$6,780	$1,250	$8,030

Cumulative P&LDrawdown from the highOut of sample from 2023-08-04

Drag across the chart to zoom into a period, double-click to reset.

Figure 1. Cumulative P&L after costs and its drawdown from the running high, 2016-01-11 to 2026-10-30, at about $250 of risk a trade. The shaded part is out of sample.

In sample (1,208 trades)Out of sample (488)Averages

Averages: −0.05R in sample, +0.08R out of sample.

Figure 2. The distribution of trade results in R, as a share of each side's trades. Bars are in sample; the line is out of sample.

Winning tradesLosing tradesOLS fitDensity contoursy = x

1,600 of 1,696 points drawn; the fit and the means use all of them. Drag a box to zoom, double-click to reset.

Figure 3. The London morning's range against the Asia range, one point per trade, log x-axis, log y-axis; the dashed line is the OLS fit (r = 0.47, p < 0.001, n = 1,696; Spearman ρ = 0.43). Out of sample alone: r = 0.54, p < 0.001. The dotted line is y = x.

5Robustness

We reran the strategy on a grid of 25 settings of its two main parameters, each on the full history with the same split. 0 were profitable in sample and 25 out of sample.

1.00

1.14

1.25

1.27

1.16

1.17

1.19

1.25

1.20

1.31

1.29

1.28

1.25

1.30

1.27

1.25

1.41

1.31

1.33

1.32

0.50

1.5

Target (R) across, start of the Asia range (hour ET) up. Solid outline: the defaults we tested; dashed: the in-sample best.

Hover a cell for both sides of the split.

Target (R) 0.50 to 3.00 across · Start of the Asia range (hour ET) 18.0 to 22.0 in depth · Profit factor, out of sample 1.00 to 1.41 as height and shade. Drag or use the arrow keys to turn it.

Drag to rotate; ctrl or ⌘ and scroll to zoom.

Figure 4. Profit factor across target (R) and start of the Asia range (hour ET), as a map and as a surface of the same grid. Values above 4 are drawn at 4.

What would change it

Nothing on the grid helps. Its in-sample best, target (R) 1.5 and start of the Asia range (hour ET) 19, had profit factor 0.95 in sample but 1.27 out of sample, against 1.29 for the defaults.

In-sample yearOut-of-sample year95% interval

Figure 5. Average trade by calendar year with its 95% interval: a year-by-year walk through the history. Lighter bars are in-sample years.

Figure 6. Average trade in three volatility regimes: terciles of the 20-session average of the daily range (% of the open), up to the day before the trade (full session). The cut-offs use the whole period, so this describes, it doesn't forecast.

Figure 7. Average trade by weekday and by the hour of entry (New York time), with 95% intervals.

Median25th to 75th percentile5th and 95thRuin: down $2,000

Figure 8. 2,000 simulated years of 252 sessions, drawn session by session from the whole backtest (days without trades included), with 48 paths drawn; red paths lost $2,000 at some point. Risk of ruin 44.1% [41.9%, 46.3%]; 51% of years ended up; median ending P&L +$85, 5th to 95th percentile −$4,333 to +$4,327.

6Prop-firm implications

The backtest's sessions run through the rules of 36 tracked 50K evaluations: 2,000 runs of up to 60 sessions each, with the drawdown and daily limits checked against each trade's intraday low. Intervals are Wilson 95% intervals of the simulation. A simulation of the past, not the chance that anyone passes.

Plan	Pass	Fail	Out of time	Median sessions
FundedNext Futures Flex Challenge 50K	5.9%	41%	53%	45
Alpha Futures Standard 50K Evaluation	2.5%	20%	78%	50
Alpha Futures Zero 50K Evaluation	2.5%	20%	78%	50
Apex Trader Funding 50K EOD Evaluation	2.5%	20%	78%	50
Apex Trader Funding 50K Intraday Trailing Drawdown Evaluation	2.5%	23%	74%	50
AquaFutures Beginner 50K Evaluation	2.5%	10%	87%	50
Bulenox 50K Qualification Account (Legacy, Option 1, trailing drawdown)	2.5%	11%	87%	50
Bulenox 50K Qualification Account (Legacy, Option 2, EOD drawdown)	2.5%	9%	89%	50
Elite Trader Funding 50K 1-Step Evaluation	2.5%	24%	74%	50
Elite Trader Funding 50K End of Day Drawdown Evaluation	2.5%	20%	77%	50

Scale 0 to 10%.

Figure 9. Simulated pass odds by plan, best first, with each plan's fail and out-of-time shares and the median sessions to a pass.

7Limitations

One market (MGC); the strategy is usually taught on currencies.
The Asia window and the 03:00 start are fixed in New York time, so they move against London time around daylight-saving changes.
Stop entries fill with a tick of slippage; fast breaks can slip more.
A backtest on the past, with fills simulated from one-minute bars; live fills, queues and outages differ.

8Conclusion

Depends on the period: it lost in sample (profit factor 0.90) and made money out of sample (1.29). Nothing on the grid helps. Its in-sample best, target (R) 1.5 and start of the Asia range (hour ET) 19, had profit factor 0.95 in sample but 1.27 out of sample, against 1.29 for the defaults. A rule that loses for years and then wins says more about the market changing than about the rule: the out-of-sample gain needs data after 2026-10-30 before it means anything.