Gold settlement versus last trade: which price belongs in your journal?

Start with the purpose of the number

A gold trader can finish a session with two defensible closing references and still write an inaccurate recap. One field records the last observed transaction in a particular futures contract. Another records the exchange’s daily settlement for that contract. The error begins when both become an unqualified “gold close,” or when yesterday’s settlement is compared with today’s last trade without disclosing the change of basis.

The practical question is not which number looks more familiar. It is what the comparison is intended to measure. A settlement series supports a consistent settlement-to-settlement record. A timestamped last trade describes a transaction observed at that time. A broker account may use another contractual valuation convention. Those purposes can coexist, but each needs its own column.

The settlement is a defined calculation

CME’s December 2025 notice, effective for trade date 12 January 2026, specifies a 13:29–13:30 Eastern Time active-month gold settlement window. Its first tier uses the volume-weighted average price of eligible trades, rounded to the nearest tradable tick. The procedure also defines fallbacks when that calculation is unavailable, including checks involving the last trade and the closing bid and ask.

The same document gives separate procedures for other contract months, drawing on calendar-spread information and further fallbacks. Consequently, “the settlement” is not a universal last print copied across expiries. Read the procedure for the exact contract and preserve whether your record is preliminary or final. This article does not supply a current settlement quotation or claim that a particular fallback was used today.

Source notes: [1] CME Group — SER 9637: amendments to gold daily settlement procedure

A small example exposes the difference

Consider an entirely hypothetical calculation window containing one contract traded at 4,000 and three at 4,002. The volume-weighted average is (1 × 4,000 + 3 × 4,002) ÷ 4 = 4,001.5. If the final transaction was at 4,002, the last trade and the average differ by 0.5 per ounce. Neither calculation implies a data error. These numbers are arithmetic teaching inputs, not historical gold prices.

Now suppose a chart displays a later trade after that calculation window. Comparing it with the earlier settlement mixes both method and observation time. The difference cannot immediately be called a premium, an execution loss or evidence that the exchange price was wrong. First identify which part comes from the averaging method and which part may come from subsequent price movement.

Build a journal that preserves provenance

I would retain five fields beside each closing reference: full instrument identifier and expiry, price type, observation time with timezone, publisher, and retrieval time. A sixth field can record publication status or a later correction. Keep the raw value before applying display rounding. A note such as “December contract, final exchange settlement” is materially more informative than a screenshot labelled “XAU close.”

For daily change calculations, require matching price types and contract identity on both sides. If a futures roll changes the contract, flag the transition and explain the method rather than silently subtracting the two prices. If one observation is unavailable, retain a gap or create a separately labelled provisional comparison. Filling the gap with a convenient quote makes the spreadsheet complete while weakening its meaning.

Separate evaluation from execution

A signal recorded near the settlement window presents a different question: what reference price and timestamp were actually available when the signal was created? Using the day’s eventual settlement as the assumed entry introduces information that may not have existed at the decision time. An audit should preserve the signal’s original levels, then compare subsequent observations under a declared method.

Likewise, a closing reference is not evidence of a broker fill. A trader who submitted an order needs the execution record to establish price, quantity and time. The exchange settlement does not reveal the spread paid, partial fills, commissions or slippage. This distinction lets a journal examine market behaviour without accidentally converting a valuation convention into a claim about money earned.

Choose consistency before apparent precision

The useful output is a journal whose labels allow another reader to reconstruct every comparison. Record settlement when analysing a settlement series; record the transaction when studying a particular execution or observation. If both are relevant, show both and describe why. More decimal places cannot repair an unidentified instrument or a mismatched timestamp.

The separate Daily Gold Brief handles the current market narrative. For a Dubai buyer, even a correctly labelled futures settlement remains distinct from a dealer’s currency, purity, weight and premium calculation. The linked physical-market guide addresses that separate purchase. This article and its conceptual illustration provide a recording method, not a current price, trade recommendation or independently verified signal outcome.