Backtest · XAUUSD 1-minute data · Oct 2021 to Sep 2026

The “Trading Challenges UMR” scalp, tested on five years of gold

I took every rule from Andre Rizky’s video, turned it into code, and ran it on 1.75 million one-minute gold bars with real spreads. It made 345 trades and finished at −8.1R. That is statistically indistinguishable from zero, and the H4-candle rule picked direction no better than a coin flip.

Verdict: no long-run edge

Net result−8.1R345 trades in 5 years

Per trade−0.02R95% range −0.16R to +0.11R

Win rate39%avg win +1.32R, avg loss −0.88R

Max drawdown25.3Rworst run: 10 losses in a row

Rp3.5M accountRp2.36Mpeaked at Rp5.49M in Apr 2023

R = profit or loss divided by the stop-loss distance. With his 80-pip gold stop at 0.01 lot, 1R ≈ Rp132,000 (≈4% of Rp3.5M).

Source

What the video says

The video (Trading Challenges UMR, modal 1 juta sampai 10 juta) shows a small account run by the Andre Rizky Investasi team, starting at about one month of minimum wage. The pitch is steady monthly income, not jackpots: fixed rules, losses calculated before every trade, and money reinvested.

He trades only the “expensive” markets: gold (XAUUSD), Bitcoin and indices, and rarely forex. In his trade history most losses were capped at 1–2.3% of the account. Many trades closed at +0.05% to +0.07% after he moved the stop into profit (“SL plus”), and a few winners were held to +3%, +5% and +8%.

Claimed on MyFXBook

Win rate: 66%
Gain: +73%
Max drawdown: −20%
Account: Rp3.5M → Rp6.1M
Markets: XAU, BTC, indices

His five rules (whiteboard)

1	Risk per trade at most 4%. Small accounts can’t risk 1%. Lot stays at 0.01.
2	Time frames top-down: M30 → M15 → M5.
3	Trade from 19:00 (Jakarta time, WIB).
4	One position (“layer 1”), never full margin.
5	Extra margin: deposit bonus ×2 plus 1:1000 leverage.

Loss per trade, calculated first (Rp3.5M account)

Market	Lot	SL	Loss
XAUUSD	0.01	80 pips	Rp140k · 4%
XAGUSD	0.01	25 pips	Rp220k · 6.3%
Forex	0.05	12 pips	Rp100k · 2.9%
Indices	0.05	100 pips	Rp100k · 2.9%

On gold, 1 pip = $0.10, so 80 pips is an $8 move.

The method

Step by step, from chart to risk

This is the scalping routine as he teaches it, in the order you would do it. The quotes are his own words.

Set up the account

One position at a time, never full margin. He adds a deposit bonus (×2) and 1:1000 leverage so a small account has room for margin.

“Layer 1, tanpa full margin.”One layer, no full margin (on screen).
Wait for the session

Only trade from 7 pm Jakarta time (12:00 UTC). That is the London afternoon, in the run-up to the New York open.

“Waktu trading wajib jam 7 malam.”Trading time must be 7 pm.
Read the H4 candle

Follow the colour of the H4 candle: red is selling pressure, so you look for sells; green means buys. Then mark the H4 critical area the move is heading for. That level becomes the target. In his gold example the H4 was red above a demand zone at 4,304, with about 270 pips of room. In the M30 version of the routine, if the higher time frame shows nothing interesting, you don’t go down to smaller charts.

“Candle H4-nya terbentuknya apa? Selling pressure merah.”What is the H4 candle forming? Red selling pressure.
Find the M15 continuation

On M15, look for the main trend plus a small move against it: down, a small rally, then down again. He draws the small rally as a counter-trend channel.

“Ketika kiri turun, ada naik kecil, maka dia lanjut turun.”When the left side falls and there’s a small rise, it continues down.
Wait for the M5 momentum break entry trigger

Don’t enter early. On M5, wait for a momentum candle that breaks the base of the small rally (the “rally-base-rally” fails, or the demand under it breaks). Enter on that break.

“Area level ini harus ditembus dulu.”This level has to be broken first.
Draw the stop-loss first

Before entering, place the SL and calculate what it costs in rupiah. On gold that is 80 pips at 0.01 lot, about Rp140k. If the loss is too big for the account, skip the trade or cut the lot (on forex, 0.05 down to 0.02).

“Sebelum lu trading, lu tarik SL dulu.”Before you trade, draw the SL first.
Enter with a fixed SL and the H4 target SL 80 · TP 270

Gold: SL 80 pips beyond the entry, TP at the H4 target, 270 pips in his example (about 1 : 3.4). His Bitcoin example used SL 25 and TP 100 (1 : 4).

“Aku SL di sini, yang tadi minus 80, TP-ku ikutin H4, plus 270.”My SL is here at minus 80; my TP follows the H4, plus 270.
Protect the trade with “SL plus”

Once the trade runs in your favour, move the stop into profit. Most of these close at a tiny gain; one or two that run far carry the month.

“Posisi diamankan dengan stop loss di-adjust di angka positif.”The position is secured by moving the stop to a positive number.
Grow the lot slowly

Keep 0.01 lot until the account reaches Rp5–6M, then raise it step by step.

“Baru naik lot size-nya ketika uangnya sudah bertambah ke 5 juta, ke 6 juta.”Only raise the lot once the money has grown to 5 or 6 million.

Implementation

How each rule became code

A backtest can only follow rules that are written down exactly. Where the video leaves room for judgment, I chose one base reading and tested the alternatives as variants (next sections).

Video rule	What the backtest does
H4 candle colour	Colour of the last closed H4 candle. Variant: the candle still forming at entry time.
M15 continuation	The last two M15 swing highs and swing lows are both falling (sell) or both rising (buy). The most recent swing is the start of the small counter-move, and price hasn’t broken the last swing high (sell) or low (buy). Swings are 2-bar fractals, known only 30 minutes after they form, so there is no look-ahead.
M5 momentum break	An M5 candle with a body of at least 50% of its range closes beyond the last confirmed M5 swing low (sell) or high (buy), with the previous close still inside. Entry at market on the next minute.
SL / TP	Fixed 80-pip SL ($8.00) and 270-pip TP ($27.00). Variants: TP at the H4 swing level (the critical area), 2R, 1R.
Time	Entries from 19:00 to 22:59 Jakarta time. Variants: 19–21, 19–24, all day.
Layer 1	No new trade while one is open.
SL plus	Off in the base. Tested with the SL moved to +2 pips once the trade is +0.5R or +1R in profit.
No SL or TP hit	Closed at 23:55 broker time, the end of the trading day. Variant: hold up to 24 hours.
Costs	Real per-minute spread from an MT5 gold export (about $0.35 on average). If SL and TP fall inside the same minute, the SL counts.

Base rules

Results

Cumulative R, trade by trade

Shaded area: out-of-sample period (Oct 2024 on)

The curve climbs to +14.8R by April 2023, then gives everything back. The first three years (in-sample) ended flat at +0.2R; the last two (out-of-sample) lost 8.3R.

How trades ended

Outcome	Trades	Share	Avg
Hit TP (+270 pips)	27	7.8%	+3.28R
Hit SL (−80 pips)	173	50.1%	−1.00R
Closed at day end107 green, 38 red	145	42.0%	+0.53R

Tuning period vs check period

Period	Trades	Win	R/trade	Total
Oct 2021 – Sep 2024in-sample	198	43%	+0.001	+0.2R
Oct 2024 – Sep 2026out-of-sample	147	33%	−0.057	−8.3R
All, spread ×2	345	37%	−0.066	−22.9R

Net R by year

Net gainNet loss

Show the numbers

Year	Trades	R/trade	Net R

Good and bad years alternate with no trend. 2021 covers only October to December, and 2026 only January to early September.

Robustness

Does the H4 candle pick direction?

To test the core idea, I replaced the H4 candle colour with a coin flip for each H4 bar and kept every other rule. I ran that 20 times with different random seeds. If reading the H4 candle carried information, the real rule should beat almost all of these runs.

R per trade: H4 candle rule vs coin flip

H4 candle colour (video rule)Coin-flip direction, 20 runs

The real rule (−0.023R) beat 14 of the 20 coin flips, which sits well inside the random spread (−0.20R to +0.05R). The coin flips averaged −0.049R, so the H4 colour may add about 0.03R per trade at most. That is smaller than the cost of doubling the spread.

Money

What it does to a Rp3.5M account

Equity at 0.01 lot (USD/IDR 16,500)

Line at Rp3.5M: starting balance

Following his sizing exactly (0.01 lot, raised only at Rp5–6M, never above 4% risk), the lot never moves past 0.01, because 4% of the peak balance was still less than two 80-pip stops. The account rose from Rp3.5M to Rp5.49M (+57%) by April 2023, then fell to Rp2.04M, a 63% drawdown, and ended at Rp2.36M.

That first year and a half looks a lot like the video’s +73%. A system with no edge can produce a run like that at 4% risk per trade, so a few months of MyFXBook history can’t separate skill from luck.

Diagnosis

Why it doesn’t hold up

The target is rarely reached. Only 7.8% of trades hit the 270-pip TP inside the session. Most of the profit comes from trades closed at day end, and that isn’t enough to pay for the 50% that hit the stop.
A fixed 80-pip stop shrinks as gold rises. $8 was 0.45% of the price in late 2021 and about 0.16% at $5,000 in 2026. In 2026, 51% of trades were stopped out within 15 minutes. Wider stops (120 and 200 pips) came out slightly positive in both periods (+0.03R to +0.06R per trade), but the 95% range still includes zero, and it isn’t his rule.
The direction filter is close to random. See the coin-flip test above.
The 19:00 rule adds nothing. Trading all day gave −0.013R per trade against −0.023R inside the window.
SL plus raises the win rate and lowers the profit. At +1R it turned −8.1R into −13.7R; at +0.5R, into −32.8R.
Forex is clearly worse. The same rules with his 12-pip stop on 9 major pairs: 1,791 trades, −0.20R per trade, −361R. Twelve pips is too close to spread and M5 noise.

Forex side test by pair

Pair	Trades	R/trade	Net R

Conclusion

Verdict: not applicable as a long-run system

Taken as written, the method produces no positive R over five years of gold, about −0.02R per trade after spread. That result holds across every reasonable reading of the rules, and its core filter does no better than chance. The results in the video can’t be reproduced from the rules it teaches. If he has an edge, it comes from judgment the video doesn’t spell out: which levels he counts as “interesting”, when he skips a setup, and when he closes by hand. It could also be a short, lucky sample on a single account.

The risk rule matters as much as the entries. At 4% per trade, a ten-loss streak (which happened here) costs about 34% of the account, and the 25R drawdown this test saw means a 60%+ loss. His rule of 4% is a constraint of the account size, not a safe risk level.

If you still want to trade it

Forward-test on demo for at least 100 trades, and log the H4 colour, M15 pattern, M5 trigger and every manual decision. That turns the judgment calls into rules you can backtest.
Risk 1–2% per trade. With gold’s 0.01 lot minimum and an 80-pip stop, 1% needs about Rp13M in the account.
Treat the bonus and 1:1000 leverage as margin tools only. They don’t change the expected result; they only make over-sizing easier.

Caveats

Limits of this test

The judgment parts (what counts as a continuation, which level is interesting) were fixed into definitions using 2-bar fractal swings. A skilled discretionary trader may filter better; this test can’t measure that.
Only gold was tested in full. There is no Bitcoin, index or silver data here; forex was a side test.
Spreads come from one broker’s MT5 export; USD/IDR is fixed at 16,500; swaps and commissions are not included.
345 trades leave a wide 95% range (−0.16R to +0.11R per trade). A small edge of +0.05R can’t be ruled out, but none is shown.
Partial closes and other manual exits aren’t modelled; the SL plus variants approximate them.

What the video says

His five rules (whiteboard)

Loss per trade, calculated first (Rp3.5M account)

Step by step, from chart to risk

Set up the account

Wait for the session

Read the H4 candle

Find the M15 continuation

Wait for the M5 momentum break entry trigger

Draw the stop-loss first

Enter with a fixed SL and the H4 target SL 80 · TP 270

Protect the trade with “SL plus”

Grow the lot slowly