Moving a gold limit order changes more than its price

The displayed price is only part of the order

A resting limit order has an instrument, side, price, quantity and an accepted position in the matching process. Changing one field can affect another property that is less visible on the trading screen. That is why moving an order back to its original price does not necessarily restore its original place.

CME’s current 1-Ounce Gold futures FAQ identifies its matching algorithm as 100% FIFO. CME’s general FIFO documentation describes timestamp-based priority and identifies price changes, increases in working quantity and account changes as modifications that lose timestamp priority. This article uses that documented FIFO setting; it does not assume every gold venue, spread or broker product follows the same allocation rules.

Source notes: [1] CME Group — FAQ: 1-Ounce Gold futures · [2] CME Group Client Systems — Globex Matching Algorithm Steps

A queue example with no live orders

Consider a deliberately simplified queue at one buy price. Order A has three contracts and arrived first. Your order has two and arrived second. Order B has four and arrived third. Assume plain displayed orders, pure FIFO and no other book changes. An incoming sell order for five contracts would fill A’s three and then your two.

Now change your working quantity from two to three before that sell order arrives. Under the stated loss-of-priority rule, your amended order goes behind B at that price. The same five-contract sell would fill A’s three and two of B’s four; your amended order receives none. These invented sequences isolate one mechanism. They do not predict a real fill or reconstruct an actual gold order book.

A request is not yet an accepted change

CME’s order-functionality documentation distinguishes a pending modification from an acknowledged order state and explains the risk of a fill occurring while a modification is in flight. It also describes an optional in-flight mitigation mechanism. A user should establish how the chosen broker or interface handles this behaviour rather than assume a clicked amendment instantly changes the exchange record.

My review would retain request time, acknowledgement time and any intervening fill. If a screen shows the requested quantity before acknowledgement, I would not use that display alone to reconstruct exposure. The authoritative event sequence matters because a decision made at the interface and its processing by the venue are separate events.

Source notes: [3] CME Group Client Systems — Order Functionalities

Improving the price can still be a deliberate choice

Losing time priority does not automatically make an amendment wrong. A trader may decide that the original price is no longer acceptable, that urgency has changed or that the intended size is different. The proper comparison is between the consequences of leaving the order and those of changing it under the updated objective.

I would write that objective in plain language before assessing the outcome. Was the goal to reduce exposure, obtain a faster fill, avoid an unattractive execution, or increase intended quantity? An amendment that produces no fill can be consistent with avoiding an unacceptable price. Judging every change solely by whether it filled rewards activity rather than the decision the order was supposed to implement.

Do not rebuild queue position from candles

A bar touching a limit price does not supply the complete order sequence needed for this audit. The simplified example required arrival order and quantities ahead; a candle does not contain those fields. Even a depth snapshot may be insufficient if relevant changes occurred between snapshots. Where the evidence is absent, the execution reconstruction should remain unresolved.

For a practice exercise, I would log the original and replacement identifiers, accepted price and quantity, modification reason, acknowledgements, fills and remaining quantity. Then compare the record with the rule actually applicable to the product. This keeps a signal-level observation separate from a claim that an individual account should have obtained a particular execution.

Measure the policy over more than a memorable fill

An amendment policy can be reviewed by recording every change over a defined interval, including no-fill outcomes and requests that were rejected or overtaken by fills. Useful questions include how often the rule changed size, how long acknowledgement took and whether the recorded reason matched the stated objective. A single frustrating queue outcome is not enough to establish a superior alternative.

No broker fill, fee, spread, slippage or P&L is measured in this article. The queue arithmetic is hypothetical, and no CY66 result is inferred from it. The practical improvement is narrower: treat order modifications as events with consequences and retain the evidence needed to distinguish an intended change from the state the venue actually accepted.