When does gold move? Returns by session

Micro Gold (MGC) · 5-minute bars · 2016 to 2026 · Seasonality

Marginal: profit factor 1.01 out of sample, not clearly above break-even, t = 0.9.

Profit factor, out of sample: 1.01
Average trade, out of sample: +0.01R
Win rate, out of sample: 53%
Trades, 10 years: 2,778
Max drawdown: $12,053
Pass odds, median plan: 0.1%

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 gold session returns rules on Micro Gold (MGC) 5-minute bars from 2016-01-11 to 2026-10-30: 2,778 trades, with commission and a tick of slippage on every market and stop fill. In sample (to 2023-08-03) it made −0.02R per trade with a profit factor of 0.83; out of sample (from 2023-08-04) +0.01R per trade and 1.01 over 834 trades. 7 of the 16 parameter settings we tried were profitable out of sample. Sized at $250 of risk a trade, simulated 50K evaluations passed 0% to 0% of the time across 36 tracked plans.

1Hypothesis

Split gold's trading day into Asia, London, New York and the afternoon, and measure each session's average return. Then hold gold through London's morning only.

If gold's return concentrates in one session, holding only that session should keep the return with a fraction of the exposure. We fixed London, gold's home market, before testing.

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:

Entry: buy at entry_min (03:00 ET, as minutes after midnight: 180), every trading day.
Exit: sell at exit_min (08:20 ET: 500), the COMEX open.
Stop: one previous day's full-session range below the entry, for a defined risk.
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"""Gold's London hours: hold Micro Gold (MGC) from the London open to the COMEX open, long only.2 3The study measures gold's average return in each session of the day; the rule tests one window, fixed4before the test: London's.5 6Rules:7  1. Entry: buy at `entry_min` (03:00 ET, as minutes after midnight: 180), every trading day.8  2. Exit: sell at `exit_min` (08:20 ET: 500), the COMEX open.9  3. Stop: one previous day's full-session range below the entry, for a defined risk.10  4. Size: about `risk_usd` dollars of risk a trade (micro contracts, rounded down, at least 1 and at11     most `max_contracts`).12"""13import qy14 15 16class Bot(qy.Strategy):17    """Own gold through London's morning."""18 19    symbol = "MGC"20    timeframe = "5m"21    session = None22    params = {"entry_min": 180, "exit_min": 500, "risk_usd": 250, "max_contracts": 50}23    max_position = 5024    overnight = False25 26    def setup(self):27        self.h = self.l = None28        self.range = None29 30    def on_day_start(self, day):31        if self.h is not None:32            self.range = self.h - self.l33        self.h = self.l = None34 35    def _size(self, risk_points):36        per_contract = risk_points * self.point_value37        n = int(self.p.risk_usd // per_contract) if per_contract > 0 else 138        return max(1, min(self.p.max_contracts, n))39 40    def on_bar(self, bar):41        self.h = bar.high if self.h is None else max(self.h, bar.high)42        self.l = bar.low if self.l is None else min(self.l, bar.low)43        t = bar.dt.hour * 60 + bar.dt.minute + 544        if self.position.is_long and t == self.p.exit_min:45            self.close_position("session over")46            return47        if t == self.p.entry_min and self.position.is_flat and self.range and not self.orders:48            self.buy(self._size(self.range), stop_price=bar.close - self.range, tag="London hours long")

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	1,944	834	2,778
Net P&L	−$9,932	+$657	−$9,275
Profit factor	0.83	1.01	0.91
Average trade	−0.02R	+0.01R	−0.01R
95% interval of the average	−0.04 to −0.01	−0.01 to 0.03	−0.03 to 0.00
t-statistic of trade R	−2.82	0.95	−2.01
Win rate	49%	53%	51%
Average win / average loss	0.85	0.90	0.90
Max drawdown	$12,036	$5,104	$12,053
Sharpe ratio (daily, annualized)	−0.99	0.06	−0.40
Longest losing streak	17	8	17
Long / short profit factor	0.83 / n/a	1.01 / n/a	0.91 / n/a
Average holding time	320 min	319 min	319 min
Commission paid	$4,022	$1,284	$5,306

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 (1,944 trades)Out of sample (834)Averages

Averages: −0.02R in sample, +0.01R 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.

Figure 3. Gold's average log return in each session of its trading day, with 95% intervals

Asia 18:00-03:00London 03:00-08:20New York 08:20-13:30

Figure 4. Cumulative log return of gold in the Asia, London and New York sessions

5Robustness

We reran the strategy on a grid of 16 settings of its two main parameters, each on the full history with the same split. 0 were profitable in sample and 7 out of sample.

720

1.02

0.94

0.93

0.99

600

1.03

0.94

0.98

500

1.13

1.01

1.02

1.06

420

1.05

0.92

0.91

0.93

120

180

240

300

Entry (minutes after midnight ET) across, exit (minutes after midnight ET) up. Solid outline: the defaults we tested; dashed: the in-sample best.

Hover a cell for both sides of the split.

Entry (minutes after midnight ET) 120.0 to 300.0 across · Exit (minutes after midnight ET) 420.0 to 720.0 in depth · Profit factor, out of sample 0.91 to 1.13 as height and shade. Drag or use the arrow keys to turn it.

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

Figure 5. Profit factor across entry (minutes after midnight ET) and exit (minutes after midnight 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, entry (minutes after midnight ET) 240 and exit (minutes after midnight ET) 720, had profit factor 0.93 in sample but 0.93 out of sample, against 1.01 for the defaults.

In-sample yearOut-of-sample year95% interval

Figure 6. Average trade by calendar year with its 95% interval: a year-by-year walk through the history. Lighter bars are in-sample years.

Figure 7. 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 8. 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 9. 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 48.4% [46.2%, 50.6%]; 34% of years ended up; median ending P&L −$835, 5th to 95th percentile −$4,137 to +$2,583.

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
Apex Trader Funding 50K Intraday Trailing Drawdown Evaluation	0.4%	15%	84%	46
AquaFutures Beginner 50K Evaluation	0.4%	12%	88%	46
Bulenox 50K Qualification Account (Legacy, Option 1, trailing drawdown)	0.4%	8%	92%	46
Elite Trader Funding 50K 1-Step Evaluation	0.4%	15%	84%	46
Funded Futures Family 50K Evaluation (standard plan)	0.4%	14%	86%	46
FundedNext Futures Rapid Pro Challenge 50K	0.4%	14%	86%	46
Top One Futures Elite Challenge 50K	0.4%	14%	86%	46
Tradeify 50K GROWTH Evaluation	0.4%	13%	86%	46
Funded Futures Network STEADY 50K Evaluation	0.3%	14%	86%	44
Alpha Futures Standard 50K Evaluation	0.2%	14%	86%	42.5

Scale 0 to 10%.

Figure 10. 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

Session boundaries are fixed in New York time and drift against London time around daylight-saving changes.
Long only in a period when gold rose.
A backtest on the past, with fills simulated from one-minute bars; live fills, queues and outages differ.

8Conclusion

Marginal: profit factor 1.01 out of sample, not clearly above break-even, t = 0.9. Nothing on the grid helps. Its in-sample best, entry (minutes after midnight ET) 240 and exit (minutes after midnight ET) 720, had profit factor 0.93 in sample but 0.93 out of sample, against 1.01 for the defaults. The result is too close to break-even to tell apart from noise after costs.