The Asia Opening Range
on Gold

One trade a day, Monday to Thursday, decided in the fifteen minutes after midnight UTC. This is the complete rule set, why each rule is there, and every trade it produced in 2026.

Risk per trade

Every percentage here is R × risk. R values, win rate and profit factor do not move. Past 1.96% the worst run on record breaks the maximum loss; past 2.74% one bad trade breaches the daily limit alone.

A run of 6 losses costs 15.3%, past the 12% maximum loss. That run is in this data.

Total return

+109%

+43.6 R banked at 2.5%

Win rate

52.0%

39 wins · 36 losses

Per trade

+0.582 R

±0.165 standard error

Profit factor

2.40

+74.7 R won vs 31.1 R lost

Average win

+1.92 R

35 reached the target

Average loss

-0.86 R

12 of 36 halved by the stop move

Challenge pass rate

96.8%

one phase, +12% target

Trades to pass

about 14 trading days

Worst drawdown

15.4%

the limit is 12%

Longest win run

wins in a row

Longest losing run

15.3% of the account

Trades taken

from 147 eligible sessions

01The rules

Everything is mechanical. No discretion, no judgement calls, and nothing that needs you awake at a screen.

One session end to end, on two straight scales: 10 px a minute across, 50 px an R up. Drawn with the fill at the range high; in practice it lands a few points beyond and the target is measured from the actual fill.

Mark the rangeTake the highest high and lowest low of the fifteen M1 candles from 00:00 to 00:14 UTC. That box is the range.
Split the box in half and see where the 00:14 candle closedClosed in the top half → you may only take an upward break. Closed in the bottom half → only a downward break. Breaks the other way are skipped — section 02.
Watch for a break until 00:29The first M1 candle to close outside the box triggers the trade — above the high goes long, below the low goes short. Wicks poking through don't count. Nothing by 00:29 means no trade.
Enter at marketOn the next candle, in the direction of the break.
Stop at the range midpointHalfway between the high and the low. That distance is your 1R.
Target at 2RTwice the distance from entry to stop, measured from the actual fill.
At +0.5R, move the stop to −0.5ROnce the trade is that far toward target, pull the stop so the worst case is a part loss instead of a full one. Once only — it never moves again.
Flat after 90 minutesIf neither level is hit, close at market. 95% of trades finish on their own before this.
Monday to Thursday, one trade a dayFridays off. One position, then done regardless of the outcome.
Risk 2.5% of the starting balanceFixed, not compounding. Position size = 2.5% ÷ the distance from entry to stop.

Settings, exactly as they appear in the EA
Input	Value	Meaning
`InpTimeZone`	TZ_UTC	session clock
`InpStartHour / Minute`	0 / 0	00:00 UTC
`InpRangeMinutes`	15	range length, so the box closes 00:14
`InpSignalTF`	M1	confirmation candle
`InpEntryMode`	MARKET_ON_CLOSE	market, the moment the signal candle closes
`InpNoEntryAfterMin`	15	stop looking at 00:29
`InpMinClosePos`	0.50	trade only the half it closed in
`InpTradeLongs / Shorts`	true / true	either direction
`InpSLPercentOfRange`	50	stop at the midpoint
`InpRR`	2	target
`InpStopMoveAtR / ToR`	0.5 / −0.5	the stop move
`InpMaxHoldMinutes`	90	time cap
`InpTradeFri`	false	Fridays off
`InpRiskPercent`	2.5	risk per trade

02The half-of-the-range rule

One filter, and it is the only one the edge needs. Draw a line across the middle of the range box and ask which half the 00:14 candle closed in. That half is the only direction you are allowed to trade today.

The whole rule

Closed in the top half → longs only. If price breaks down instead, you sit out.
Closed in the bottom half → shorts only. If price breaks up instead, you sit out.

The half the range closed in is the half price was already leaning into. Breaking that way continues the move; breaking the other way means price crossed the whole box first.

Reading it live, in three seconds

At 00:15, look at the box and find its middle.
Did the last candle close above the middle? Only a break of the high counts today.
Below the middle? Only a break of the low counts.
Price breaks the wrong side → no trade, close the platform.

The one thing people get wrong

The candle's colour doesn't matter. A red candle sitting near the top of the box still counts as the top half — it drifted down a little but price is still up there. You are reading where it is, not which way it moved.

Every 2026 Monday–Thursday break, split by which half the range closed in. Totals, not averages, except EV. Total % is the account move at 2.5% risk per trade. Profit factor is R won divided by R lost — 2.00 means twice as much won as lost.
Which half broke	Trades	Wins	Win rate	Profit factor	EV per trade	Total R	Total %	Verdict
Broke the half it closed in	76	39	51.3%	2.33	+0.560	+42.6 R	+106.5%	TRADE
Broke the opposite half	18	4	22.2%	0.78	-0.126	-2.3 R	-5.8%	SKIP
Every break, no filter	94	43	45.7%	1.95	+0.429	+40.3 R	+100.8%

Why this filter is worth having

Trading every break returns +40.3 R from 94 trades at a 45.7% win rate. Trading only the correct half returns +43.6 R from 75 — virtually the same money from 18 fewer trades, at a 52.0 win rate. The 18 skipped trades contributed -2.3 R between them: all of the drawdown, none of the profit. Fewer trades at a higher win rate is exactly what a pass-rate account wants.

03The stop move

Once per trade, and only in one direction.

To scale — 1R is the same height everywhere.

12 of the 36 losses were halved.
Average loss -0.86 R, not −1.00 R.
Saved +4.9 R — +12.2% of the account.

How the 75 trades ended
Exit	Trades	Share	EV each	Total R	Total %
Stopped out	36	48.0%	-0.86 R	-31.1 R	-77.8%
Target hit (+2R)	35	46.7%	+2.04 R	+71.3 R	+178.2%
90-minute cap	4	5.3%	+0.84 R	+3.3 R	+8.2%

Nearly even on count — 35 targets, 36 stops. The edge is size: a win pays 2R, a loss costs 0.86R.

04Why each rule exists

Every number was measured, not chosen. Here is the reasoning in plain terms.

Rule	Why this and not something else
00:00 UTC	A quiet handover between New York closing and Asia opening — not an auction. Ranges built inside an opening auction get violated in both directions and carry no information.
15-minute range	Shorter is better. A 30- or 60-minute range gives a wider box, a wider stop, a smaller position and a worse result. The information is in the first fifteen minutes.
M1 confirmation	The fastest candle catches the earliest break, and the earliest break is the best one — 39% of all breaks happen within five minutes of the range closing.
Entries stop at 00:29	A range that takes more than fifteen minutes to give way is drifting, not breaking. Including later breaks cuts expectancy by more than half.
Stop at the midpoint	Tighter stops sit inside normal noise and get taken out before the move resolves. Wider stops shrink the position so much that the 2R target becomes unreachable inside the hour. The midpoint is the balance point.
2R target	Roughly five times the expectancy of a 1R target. Pushing to 3R lowers the win rate faster than it raises the payoff.
Stop to −0.5R, never to breakeven	A breakeven stop sits exactly where price retests after a breakout — it stops you for nothing on trades that would have worked. Half a risk unit back is far enough away to survive the retest.
Trigger at +0.5R	Moving earlier tightens the stop while the trade is still inside noise. Moving later leaves more on the table when it reverses.
90-minute cap	95% of trades finish on their own before it. The cap only touches the tail, and it keeps you out of the market rather than holding through the London open. At 60 minutes it was cutting three trades off within a hair of their target.
Monday to Thursday	Friday's Asia session runs into weekend position-squaring — flow driven by people flattening rather than committing. It was the only losing day of the week, at a 20.8% win rate.
One trade a day	Keeps a bad day to a single loss and the daily exposure well inside any prop-firm daily limit.

05Results

2026, MetaTrader 5, real tick data, spread and commission included. Risk is a fixed 2.5% of the starting balance on every trade — no compounding, so each R is worth a constant two percentage points.

Cumulative return across 2026. Every step is one trade.

By quarter — trading days available, trades taken, and totals
Quarter	Trading days	Trades	W / L	Win rate	Profit factor	EV per trade	Total R	Total %
Q1	50	26	16 / 10	61.5%	3.76	+0.828	+21.5 R	+53.8%
Q2	52	28	13 / 15	46.4%	2.02	+0.454	+12.7 R	+31.8%
Q3	45	21	10 / 11	47.6%	1.87	+0.445	+9.4 R	+23.5%

By month
Month	Trading days	Trades	W / L	Sequence	Win rate	Profit factor	EV per trade	Total R	Total %
Jan	16	6	5 / 1	W-W-W-W-L-W	83.3%	9.13	+1.394	+8.4 R	+21.0%
Feb	16	9	4 / 5	L-L-W-L-W-W-L-L-W	44.4%	1.77	+0.397	+3.6 R	+9.0%
Mar	18	11	7 / 4	L-L-L-W-W-L-W-W-W-W-W	63.6%	5.47	+0.874	+9.6 R	+24.0%
Apr	18	10	5 / 5	W-L-L-W-W-L-L-W-L-W	50.0%	2.22	+0.565	+5.6 R	+14.0%
May	16	8	4 / 4	L-L-L-W-W-W-L-W	50.0%	2.17	+0.455	+3.6 R	+9.0%
Jun	18	10	4 / 6	L-L-L-W-L-L-W-W-W-L	40.0%	1.73	+0.343	+3.4 R	+8.5%
Jul	18	7	5 / 2	L-L-W-W-W-W-W	71.4%	4.82	+1.149	+8.0 R	+20.0%
Aug	17	10	5 / 5	W-L-W-W-W-L-L-W-L-L	50.0%	2.16	+0.538	+5.4 R	+13.5%
Sep	10	4	0 / 4	L-L-L-L	0.0%	0.00	-1.018	-4.1 R	-10.2%

Every quarter positive

Q1 +21.5 R, Q2 +12.7 R, Q3 +9.4 R.

06Week by week

Actual trading days per week — Monday to Thursday, so four in a full week and fewer around holidays. Greyed rows are weeks where no session met the entry conditions.

Every week of 2026. Trades taken out of the days available.
Week	Trading days	Trades	W / L	Sequence	Win rate	Profit factor	EV per trade	Total R	Total %
05 Jan – 08 Jan	4	1	1 / 0	W	100%	–	+2.037	+2.0 R	+5.0%
12 Jan – 15 Jan	4	3	3 / 0	W-W-W	100%	–	+1.787	+5.4 R	+13.5%
19 Jan – 22 Jan	4	2	1 / 1	L-W	50%	1.94	+0.482	+1.0 R	+2.5%
26 Jan – 29 Jan	4	0	–	–	–	–	–	–	–
02 Feb – 05 Feb	4	3	1 / 2	L-L-W	33%	1.28	+0.147	+0.4 R	+1.0%
09 Feb – 12 Feb	4	1	0 / 1	L	0%	0.00	-1.034	-1.0 R	-2.5%
16 Feb – 19 Feb	4	2	2 / 0	W-W	100%	–	+2.101	+4.2 R	+10.5%
23 Feb – 26 Feb	4	3	1 / 2	L-L-W	33%	0.98	-0.013	-0.0 R	-0.0%
02 Mar – 05 Mar	4	3	0 / 3	L-L-L	0%	0.00	-0.544	-1.6 R	-4.0%
09 Mar – 12 Mar	4	2	2 / 0	W-W	100%	–	+1.605	+3.2 R	+8.0%
16 Mar – 19 Mar	4	1	0 / 1	L	0%	0.00	-0.520	-0.5 R	-1.2%
23 Mar – 26 Mar	4	4	4 / 0	W-W-W-W	100%	–	+1.534	+6.1 R	+15.2%
30 Mar – 02 Apr	4	3	2 / 1	W-W-L	67%	4.39	+1.151	+3.5 R	+8.8%
06 Apr – 09 Apr	4	2	1 / 1	L-W	50%	1.90	+0.474	+0.9 R	+2.2%
13 Apr – 16 Apr	4	3	1 / 2	W-L-L	33%	1.29	+0.150	+0.4 R	+1.0%
20 Apr – 23 Apr	4	2	1 / 1	W-L	50%	2.19	+0.595	+1.2 R	+3.0%
27 Apr – 30 Apr	4	1	1 / 0	W	100%	–	+2.026	+2.0 R	+5.0%
04 May – 07 May	4	0	–	–	–	–	–	–	–
11 May – 14 May	4	3	0 / 3	L-L-L	0%	0.00	-0.680	-2.0 R	-5.0%
18 May – 21 May	4	4	3 / 1	W-W-W-L	75%	5.64	+1.246	+5.0 R	+12.5%
25 May – 28 May	4	1	1 / 0	W	100%	–	+0.698	+0.7 R	+1.8%
01 Jun – 04 Jun	4	2	0 / 2	L-L	0%	0.00	-0.514	-1.0 R	-2.5%
08 Jun – 11 Jun	4	2	1 / 1	L-W	50%	1.87	+0.466	+0.9 R	+2.2%
15 Jun – 18 Jun	4	3	1 / 2	L-L-W	33%	1.27	+0.141	+0.4 R	+1.0%
22 Jun – 25 Jun	4	2	2 / 0	W-W	100%	–	+2.066	+4.1 R	+10.2%
29 Jun – 02 Jul	4	2	0 / 2	L-L	0%	0.00	-1.018	-2.0 R	-5.0%
06 Jul – 09 Jul	4	1	0 / 1	L	0%	0.00	-1.094	-1.1 R	-2.8%
13 Jul – 16 Jul	4	0	–	–	–	–	–	–	–
20 Jul – 23 Jul	4	3	3 / 0	W-W-W	100%	–	+2.040	+6.1 R	+15.2%
27 Jul – 30 Jul	4	2	2 / 0	W-W	100%	–	+2.014	+4.0 R	+10.0%
03 Aug – 06 Aug	4	1	1 / 0	W	100%	–	+1.995	+2.0 R	+5.0%
10 Aug – 13 Aug	4	3	2 / 1	L-W-W	67%	7.46	+1.156	+3.5 R	+8.8%
17 Aug – 20 Aug	4	3	1 / 2	W-L-L	33%	0.98	-0.013	-0.0 R	-0.0%
24 Aug – 27 Aug	4	2	1 / 1	W-L	50%	1.95	+0.489	+1.0 R	+2.5%
31 Aug – 03 Sep	4	3	0 / 3	L-L-L	0%	0.00	-1.013	-3.0 R	-7.5%
07 Sep – 10 Sep	4	2	0 / 2	L-L	0%	0.00	-1.027	-2.1 R	-5.2%
14 Sep – 17 Sep	3	0	–	–	–	–	–	–	–

The pattern worth noticing: you take a trade on roughly half the days available. 75 trades from 147 eligible sessions. Most weeks produce two or three trades, and a handful produce none at all — those are weeks where the half-of-the-range rule or the entry window ruled every session out. Sitting out is part of the edge, not a failure of it.

07Risk and sizing

A funded-account challenge is a race between a profit target and a drawdown limit. That makes streaks matter more than averages.

Simulated on the 2026 outcome distribution. Target +12%, 12% maximum loss, 3% daily, no minimum days.
Risk per trade	Pass	Trades needed	~Days	Total return	Worst drawdown	6 losses cost	Worst single loss
0.5%	100.0%	40	78	+22%	3.1%	3.1%	-0.55%
0.75%	100.0%	26	51	+33%	4.6%	4.6%	-0.82%
1%	100.0%	19	37	+44%	6.2%	6.1%	-1.09%
1.25%	99.8%	15	29	+54%	7.7%	7.7%	-1.37%
1.5%	99.5%	13	25	+65%	9.2%	9.2%	-1.64%
1.75%	99.0%	11	22	+76%	10.8%	10.7%	-1.91%
2%	98.4%	9	18	+87%	12.3%	12.2%	-2.19%
2.25%	97.4%	8	16	+98%	13.8%	13.8%	-2.46%
2.5%	96.8%	7	14	+109%	15.4%	15.3%	-2.74%

The highlighted row is the risk set at the top; the field there accepts any value and snaps to a simulated level. Pass, trades and days come from walking the real outcomes against the barriers, not from scaling.

Losing runs that actually happened in 2026. The cost is the worst real sequence of that length, not a run of average losses — the stop move halves some losses, so averaging them understates the run that matters.
Consecutive losses	Times	Worst it actually cost	At the risk above
1 in a row	6×	-1.09 R	2.7% of the account
2 in a row	6×	-2.10 R	5.3% of the account
3 in a row	4×	-3.13 R	7.8% of the account
6 in a row	1×	-6.12 R	15.3% of the account

The one risk that matters

A 6-loss run happened once, and at 2.5% risk it costs 15% — past the 12% limit.

Realised drawdown reached 15.4% because a winner landed before that run started from a peak. The same sequence hitting a fresh account is the failure case, and it is the only one worth planning for — nearly every simulated failure is one long losing run rather than a slow bleed.

Position sizing, worked

Account $10,000, risk 2.5% → $250 per trade, every trade.
Range is 1,000 points ($10.00) → stop sits 500 points ($5.00) away.
Gold moves $1 per point per lot, so $5.00 of stop = $500 per lot.
Size = $250 ÷ $500 = 0.50 lots.

A wider range means a wider stop and therefore a smaller position — the dollar risk stays at $200 regardless. That is what makes every result comparable in R.

08The range filter

Filter on at 0.14%: 136 of 274 sessions traded, +0.260 R a trade instead of -0.019, 39.7% won instead of 29.9%.

Every session in the record, sorted by how wide its box was and split into four equal groups.

The same box, year by year, measured against gold's price.

Range width against result, all 274 sessions on file
Group	Range width	Sessions	Win rate	EV per trade	Total R	Total %
Narrowest 25%	0.031–0.082%	68	14.7%	-0.474 R	-32.2 R	-80.5%
Second 25%	0.082–0.137%	68	25.0%	-0.138 R	-9.4 R	-23.5%
Third 25%	0.137–0.237%	68	33.8%	+0.118 R	+8.0 R	+20.0%
Widest 25%	0.241–2.167%	70	45.7%	+0.405 R	+28.3 R	+70.8%

09Monte Carlo simulation

Draw a trade at random from the 75 real results, add it, repeat until the account hits +12% or −12%. That is one attempt. Eighty of them are drawn below; the numbers come from 40 000.

Passed — reached +12% Failed — hit −12%

Passed

96.8%

Failed

3.2%

Median trades to pass

Attempts simulated

40,000

10Every trade

All 75 trades of 2026, in order. Each chart shows the range, the entry, the stop, the target, where the stop moved to, and the exit. Click any one to open it full size.

win — target reached loss — stopped out ▲▼ entry · ✕ exit · shaded box = the 15-minute range

Result

Day

Month