Worked example · winner
Soft CPI long, entered on the held pullback
CPI headline 2.9% vs 3.2% consensus, core also below forecast. Real-yield implication is unambiguously gold-positive. No FOMC or other tier-one event scheduled for the remainder of the day.
- Release context
- CPI at 13:30 UTC, soft on both headline and core, no second event that session
- Phase 1 (0–2 min)
- 13:30–13:32 UTC, gold spiked from 4,486 to 4,509, spread peaked at $1.15
- Phase 2 confirmation
- 13:35 and 13:40 UTC 5-minute candles both closed higher, second beyond the first
- Entry
- 13:48 UTC, pullback held at 4,501.20, long entered at 4,503.40
- Stop ($/oz)
- 4,497.10, beyond the pullback low plus a $1.00 slippage allowance — $6.30 risk per ounce
- Lot size
- £200 ÷ ($6.30 × $100/lot) ≈ 0.32 lots
- Spread paid at entry
- $0.38, close to normal, confirming phase 1 had passed
- Slippage on entry
- $0.20, within the allowance already built into the stop
- Target 1
- Prior day high 4,516.00, filled 14:20 UTC, half closed, stop moved to entry
- Target 2 / exit
- Trailed on 15-minute structure, closed remainder at 4,529.80 at 16:35 UTC
- R outcome
- Approximately +2.7R blended across the two exits
The checklist passed on every item: a tier-one release, two closed candles confirming direction, spread back near normal, a held pullback rather than a chase, and a macro reading that agreed with the price action — soft inflation implies lower real yields implies a higher gold price.
The stop of $6.30 per ounce was only possible because the entry waited for phase 2 to resolve. A spike entry at 13:31 UTC would have needed a stop beyond the entire $23 phase-1 range, three to four times as wide, for the same directional thesis.
No second scheduled event existed that session, so the position was allowed to run rather than being forced flat, and the trailing exit captured a large share of a real-yield repricing that took over three hours to fully play out.
What it teaches: The size of the stop and the quality of the trade were both functions of patience, not of predicting the number correctly.