Let us concede the premise before dismantling the framing. Aggregate order flow in gold does, on specific sessions, carry a signature that is consistent with new participation rather than positioned rotation — the pattern is real and documented in BIS triennial survey commentary on Asian session gold turnover. What is not consistent is the retail analyst habit of reading that signature from a single chart, on a single feed, on a single day, and posting it as a call. The Tokyo–Singapore–Hong Kong liquidity window has structural reasons for producing prints that look like new buyers when they are, in aggregate observation across many sessions, something else entirely.
The Concession: Fresh Buying Does Leave a Signature
There is a pattern worth taking seriously. When new participation enters gold in size, it tends to arrive as directional buying that does not immediately hedge itself out — the print asymmetry lasts across two or three fixings before positioned traders absorb it. Aggregate turnover data in the BIS triennial framework treats this as a distinguishable regime from rebalancing flow, which typically self-cancels within the same session window.
The signature is not visible in price alone. Price moves on both new participation and short covering, and the two look identical on a candle chart. What separates them is the residual: does the position stay after the print, or does it get flattened by the close. New buyers, by definition, hold. Rotation flow, by definition, unwinds.
This is why the honest version of the analyst claim requires more than a chart. It requires COMEX Commitment of Traders positioning updates, ETF creation-redemption balances, and — for the Asian session specifically — the LBMA morning and afternoon fix differentials. None of these are available in real time. All of them lag the intraday chart the analyst is posting from.
The Tokyo fix window is 15:00 JST daily. The London morning fix is 10:30 GMT. The gap between them is roughly seven hours during standard time.
So the concession is narrow: yes, new gold buyers do enter markets on identifiable sessions, and yes, this leaves a data signature. The claim being made in most retail posts is a different claim entirely — one that the intraday feed cannot support.
The Asian Session Framing Problem
The Asian session for gold runs roughly from Sydney open through the Hong Kong afternoon, with the dense liquidity window concentrated between Tokyo open and the Shanghai Gold Exchange afternoon session. This is a structurally different market from the London or New York sessions. Turnover is lower. Spreads are wider. The participant mix is more concentrated in physical intermediaries and central bank–adjacent flow than in speculative positioning.
This matters because low-turnover sessions produce a specific chart artifact. A single moderate order — one that would be invisible in London hours — generates a print that looks disproportionately meaningful on the chart. Retail analysts trained on the visual shape of orders in dense liquidity misread these Asian prints as institutional entry when, in aggregate observation across many sessions, they are more often small opportunistic flow into a thin book.
The JFSA regulatory framework governing FX-linked and commodity-linked retail products in Japan tightened after the 2005 FX law and again after the 2010 retail leverage cap moved to 25:1. One documented consequence is that Japanese retail participation in gold-linked products migrated toward regulated CFDs on domestic platforms rather than into the offshore market. This changed the composition of what a foreign analyst sees when looking at "Asian session gold flow" — the Japanese retail component that once traded visibly through international brokers now trades on a book most non-Japanese analysts do not see.
*The Tokyo gold market maintains a fix at 15:00 JST. Most retail chart feeds outside Japan render this in local time without flagging it.*
The MAS Singapore wholesale market framework, formalized progressively from 2008 onward, positioned Singapore as a physical gold hub with a different flow signature — vaulted metal transactions and Kilobar trades that clear physically rather than on futures. When aggregate order flow spikes during Singapore hours, the meaningful question is whether the print is physical or paper. The retail feed does not answer that question. It cannot.
The Hong Kong linked exchange rate, in place since 1983, sits underneath all of this. Gold priced in USD, with the HKD pegged inside a narrow band against USD, means the Hong Kong flow signature is closer to a pure USD gold position than the Tokyo flow signature — where JPY-USD moves overlay any pure gold read. An analyst who does not distinguish these is reading two different markets as one market.
The Volume-Print Illusion in Retail Feeds
The pattern here is consistent enough to name. Retail platforms display "volume" on gold charts, but the number displayed is broker-internal aggregated volume — sometimes tick volume, sometimes contract volume on a specific feed, almost never true market-wide turnover. Two traders using two different brokers, watching what they believe is the same chart, are looking at incompatible volume numbers.
The volume bar on a retail gold feed is a broker artifact, not a market fact — and the distance between the two is where the "new buyers" call quietly dies.
This is not an edge case. Compare the platform lists across major APAC-accessible brokers: AvaTrade quotes on MT4, MT5, AvaOptions and AvaTradeGO; FBS runs FBS Trader alongside MT4 and MT5; HF Markets runs its HFM App and the MT stack; FXTM runs FXTM Trader and the MT stack; Exness quotes MT4, MT5, its Mobile app and WebTerminal. Each of these platforms sources its price and volume from a different aggregation, often through a different liquidity provider stack. A "spike in volume" on one is not the same event as a "spike in volume" on another.
The regulatory picture reinforces this. Exness lists regulation across FCA, CySEC, FSCA, CBCS, CMA Kenya, FSA, FSC BVI, FSC Mauritius and JSC Jordan — nine jurisdictions, each with different execution and reporting requirements. FXTM lists FCA, FSCA and FSC. HF Markets lists FCA, CySEC, FSCA, DFSA and FSA. FBS lists ASIC, CySEC and FSCA. The tier-1 anchor for each varies. When a retail trader sees "1000 lots traded in the last minute" on their feed, that number reflects the trader's specific broker's book, not the aggregated global gold market — and the reporting standard behind that number varies by which regulatory entity is the primary supervisor of that book.
The analyst posting a chart with the caption "new buyers coming in" is, in most cases, showing a volume spike on one specific broker's feed and extrapolating a market-wide claim. The extrapolation does not survive contact with any of the following: LBMA aggregate turnover data (published quarterly with a lag), COMEX daily volume by contract series, or Shanghai Gold Exchange session data.
None of this makes the analyst dishonest. It makes the framing structurally underspecified.
Where the Aggregate Signal Actually Comes From
The primary documents that would support a "new buyers" claim on a specific day are these, and the reader should know none of them are available intraday: LBMA quarterly turnover summaries, BIS triennial central bank survey commentary on gold market activity, COMEX Commitment of Traders reports (published Fridays for Tuesday positioning), and World Gold Council quarterly demand trends. Each of these has a different definition of "new participation" and each has a different lag.
The BIS triennial framework treats new participation as sustained net directional flow across multiple reporting periods. The COT framework treats new participation as changes in non-commercial long positioning week-over-week. The World Gold Council framework treats new participation as ETF net creations plus documented central bank purchases. These three definitions overlap partially and disagree in the margins — and the margins are exactly where retail analyst calls tend to live.
There is a specific contradiction worth unwinding. BIS triennial commentary on gold turnover in Asian sessions has, across recent editions, described elevated turnover as increasingly driven by intermediation rather than end-user participation — the same volume, but a different attribution. Meanwhile, World Gold Council quarterly demand trends have described sustained central bank net purchasing across the same period as evidence of durable new participation. Both are looking at overlapping data. Both reach defensible conclusions. Both are operative.
The reconciliation, when done carefully, is that intermediation turnover has grown faster than end-user turnover, so the ratio has shifted even as end-user demand has grown in absolute terms. The retail analyst version of this — "new buyers" — collapses the distinction and picks the more exciting reading. The primary documents do not.
Central bank buying, when it happens, tends not to show up in the intraday chart at all. The transactions are structured as forward contracts with bullion banks and delivered against physical vaulted metal — the price impact is smoothed across days or weeks by design. If an analyst is claiming to see central bank entry on a five-minute chart, the claim is not supported by how these transactions actually clear.
*The World Gold Council data on central bank purchasing carries a one-quarter lag. The BIS turnover commentary carries a three-year cycle.*
The honest read on aggregate signal is that it is available, but it is not fast. Speed and accuracy are traded off. The retail framing wants both. The primary documents deliver one at a time.
So What Do You Actually Do
Stop treating a volume print on a retail chart as evidence of new participation. It is not evidence. It is a broker feed artifact filtered through one aggregation stack, and it does not survive comparison to any of the primary sources that would validate the claim.
If you want to track actual new gold participation, the toolkit is different and slower. Weekly COT reports for speculative positioning changes. Quarterly LBMA turnover summaries for aggregate market activity. Quarterly World Gold Council demand trends for the end-user attribution. For Asian session specifics, Shanghai Gold Exchange published data and — when available — Tokyo Commodity Exchange reporting. These are the sources that let you distinguish intermediation flow from directional participation. The lag is real. The lag is also the price of an accurate read.
For traders operating in the region through APAC-accessible desks — Saxo Bank APAC, Interactive Brokers Asia, OANDA Asia, IG Group Asia — the more useful discipline is to separate what the platform shows you from what the market is doing. The platform is a window into your broker's book. The market is a distributed system across LBMA, COMEX, Shanghai Gold Exchange, and the physical vaulting infrastructure in Singapore, Hong Kong and Zurich. The two overlap. They are not the same. Every trading decision benefits from remembering which one you are actually looking at when you make the call.
Whether the aggregate signal of new participation in gold across Asian sessions in 2026 reflects durable end-user demand or an accelerating shift in intermediation share is a question the primary documents themselves have not fully settled. The BIS and World Gold Council frameworks point in overlapping but non-identical directions, and no retail feed collapses that ambiguity. If your read on a specific session resolves it, the burden is on you to say which primary source your read is grounded in — and which lag you accepted to get there.