One market, ten days: how a gold scalper earned a $5,459 payout on a $100K account

Day one was red. That's the point.

The first session on this account lasted six minutes and lost $403. Six trades on gold, all closed within moments of each other, all red. And then — nothing. No revenge trades, no doubling down, no switching to another pair to “win it back.” The terminal went quiet until the next trading day.

That gap in the data is the most important trade this trader never took. Over the next eight sessions the account added between $744 and $1,586 every single day, using the same setup, the same market, and — until the final week — the same position size.

Eight consecutive green sessions weren't built on a better strategy. They were built on refusing to trade a worse one.

The playbook, from the raw trade log

We reconstructed the approach purely from execution data — entries, exits, sizing, and timing. Four patterns repeat so consistently they amount to a system:

1. One market. Never left it.

All 98 trades were XAUUSD. No index dabbling on slow days, no crypto weekends. Specialists compound their read of one market; generalists restart from zero on five.

2. Baskets, not bets

Entries come in clusters of 3–8 small positions built over a few minutes, then closed as one at the same second. Same idea, averaged entry, single decision to exit.

3. Session discipline

47% of all trades sit inside the London morning (06:00–09:00 UTC). No entries in dead hours, ever. He traded a window, not a mood.

4. Size followed proof

Week one: 0.2 lots flat. Only after seven green sessions did size step up — 0.4, then 0.5, briefly 1.0. Sizing tracked evidence, not excitement.

When the trades actually happened (entry hour, UTC)

Anatomy of the best session

August 18 was the account's biggest day: +$1,556. The morning shows the whole method in miniature — a small long basket that went nowhere, closed flat without drama, then a short basket built over 23 minutes when the real move showed up:

Notice the $1.50 trade. The first short was closed almost flat the moment it stopped behaving — and re-entered two minutes later when the setup confirmed. Scratching a trade isn't a loss; it's rent paid for a better entry.

The giveback — because real data has one

On August 25, after the payout was already approved, a five-position long basket lost $2,330 in sixteen minutes — the largest red cluster on the account, wiping out that morning's gains and then some. The session ended −$1,085.

We publish that number for a reason. Every equity curve in this series is complete — winners, losers, and the days the trader would rather forget. Two things stand out even in the worst basket: position size stayed at the standard 0.4 lots (no tilt-sizing), and the loss was cut in minutes, not hours. The account remains active and funded.

What's transferable

• Specialize. 98 trades, one instrument. Depth in one market beat breadth across five.
• Trade a window, not a day. Nearly half of all activity sat in the London morning. Outside the window, flat.
• Scale entries, not convictions. Baskets of small positions with one exit decision kept any single entry from mattering too much.
• Let size lag results. Sizing stepped up only after a week of green sessions — and never during a drawdown.
• Stop fast on red. Both losing days ended within minutes of the first losses. No recovery trading appears anywhere in the log.

Common questions