Institutional tier
How do FX dealers and the 4pm fix actually work?
Foreign exchange trades over the counter, with no central exchange. Dealers quote two-way prices to clients, earn the spread, and hedge the inventory they are left holding. The 4pm London fix is a benchmark rate calculated from trades in a short window around 16:00 London time, and large client orders tied to it can move price visibly.
The short version
Smart money concepts tell a story about large players engineering moves to collect retail stop orders. Before accepting or rejecting that story, it helps to know how the market is actually organised. Currencies do not trade on one exchange with one order book. They trade over the counter, across many venues and bilateral relationships, and the largest share of the business passes through dealers: banks and, increasingly, non-bank liquidity providers.
A dealer makes a market. It shows a price at which it will buy and a slightly higher price at which it will sell, and it deals with whoever asks. The gap between the two is the spread, and the spread is the main source of income. After a client trades, the dealer holds a position it did not choose. That position is inventory, and inventory is risk. Much of what a dealing desk does all day is manage that risk: offsetting one client against another, shading its quotes to attract the flow it needs, and hedging the remainder in the wider market.
The clients are corporations paying foreign suppliers, asset managers buying foreign shares and bonds, hedge funds, central banks, and retail aggregators that bundle the orders of many small accounts. Retail trading is a small corner of this. A retail platform is usually the counterparty to its own clients, or passes their net position to a liquidity provider. The individual stop orders sitting on that platform are not published to the interbank market. There is no shared map of retail stops for a large bank to consult.
The principle
Order flow is real and it moves prices. Who was behind a move, and why, cannot be read from a retail candle chart: it is an inference.
Step by step
- FX is over the counter: many venues, no central exchange, no single order book.
- Dealers quote two-way prices, earn the spread, and are left holding inventory they must manage and hedge.
- Liquidity, in the professional sense, is the ability to trade size without moving the price much. It is not a pool of retail stops.
- Benchmark orders, such as those tied to the 4pm London fix, concentrate real buying or selling into a short, known window.
- Sharp moves through obvious highs and lows have several plain explanations. The smart money story is one candidate, not the default.
A market with no centre
A share trades on an exchange with a published order book. A currency pair does not. Foreign exchange is an over the counter market: a web of dealers, electronic platforms and direct relationships, linked by credit lines. Prices on different venues stay close because traders arbitrage the differences away, not because a central book enforces one price.
This has a practical consequence for chart readers. The candles on a retail platform are built from the price feed of that platform alone. Two platforms can show slightly different highs and lows for the same minute. There is no official volume figure for spot foreign exchange, and the "volume" on a retail chart is normally a count of price updates.
The scale is large. The Bank for International Settlements survey for April 2025 measured about 9.6 trillion US dollars of turnover a day, later revised to 9.5 trillion. Most of it is not spot trading at all: FX swaps were about 42%, spot 31%, outright forwards 19% and options about 7%. The US dollar was on one side of about 89% of trades.
What a dealer does all day
A dealer is paid for immediacy. A client who needs to trade now gets a price now, and pays the spread for it. The dealer then owns the other side of that trade until it can pass the risk on.
A desk that has just sold euros to a client is short euros. If the next client wants to sell euros, the two offset and the dealer keeps both spreads. If not, the dealer may nudge its quotes, bidding a little higher to attract sellers, or hedge directly with another dealer or on an electronic venue. Holding the position and hoping is also possible, within risk limits, but it is not the business model.
- Two-way price: a bid to buy and an offer to sell, shown together.
- Spread: the gap between them, and the main revenue.
- Inventory: the net position left by client trades.
- Hedging: trading elsewhere to reduce that position.
- Skewing: shifting quotes to attract flow that reduces inventory.
Who the clients are
Most currency trading is a by-product of something else. A manufacturer converts export revenue. An asset manager buys foreign bonds and hedges the currency. A central bank manages reserves. A hedge fund takes a view on interest rates. Each has a reason to trade that has nothing to do with a retail chart pattern.
Retail traders reach this market through a platform that either takes the other side of their trades or routes the net exposure to a liquidity provider. In neither case does the wider market see individual retail orders. The idea that a large institution studies where small accounts have placed their stops, and steers price there, does not match how the plumbing works.
Two meanings of "liquidity"
On a dealing desk, a liquid market is one where a large amount can be traded quickly with little effect on price. Liquidity is deep in the major pairs during the London and New York overlap and thin late in the New York afternoon, around holidays and in the first hours of the Asian day.
In smart money vocabulary, liquidity means resting orders, especially stop orders above highs and below lows. The two meanings meet in one place. When price reaches a cluster of stops, those stops become market orders. If the market is thin at that moment, they move price further and trigger the next cluster. That chain is a stop cascade. It can produce a long wick through an obvious level and a quick recovery, with no one directing it.
How the 4pm fix works, and why its window was widened
The benchmark widely known as the WM/Reuters 4pm London fix publishes reference rates for currency pairs. For the actively traded pairs the rate is calculated from actual trades, sampled once a second during a window centred on 16:00 London time. The administrator takes the median of the traded bids and the median of the traded offers, and the midpoint between the two is the fix.
Funds that track an index need to value foreign holdings at the same rate the index uses, so they ask dealers to transact at the fix. A fixing order is that instruction: buy or sell this amount at the benchmark rate, whatever it turns out to be. The dealer accepts the order before the rate is known and carries the risk in between.
Because many clients use the same benchmark, orders arrive from all directions and the dealer nets them. Whatever is left over has to be bought or sold in the market ahead of and through the window. On days when the net is large and one-sided across the market, such as some month ends, price can move noticeably into 16:00 London and partly retrace afterwards.
Until 15 February 2015 the calculation window was one minute long: thirty seconds either side of 16:00. With a window that short, most fixing trades were packed into about a minute, and the official review concluded that a wider window would be harder to influence.
On 30 September 2014 the Financial Stability Board published recommendations on foreign exchange benchmarks, including widening the window. From 15 February 2015 the window became five minutes: two and a half minutes either side of the hour. A wider window means the same order has to be worked over a longer period, and any single burst of trading counts for less.
What regulators found
In November 2014 and May 2015 authorities in the United Kingdom and the United States penalised several banks over their spot foreign exchange trading. The findings included attempts to manipulate benchmark rates such as the 4pm fix, and traders sharing confidential client order information in electronic chat rooms.
- UK Financial Conduct Authority, 12 November 2014: GBP 1,114,918,000 in total across five banks, for failing to control business practices in their spot foreign exchange trading.
- UK Financial Conduct Authority, May 2015: a sixth bank, GBP 284,432,000.
- US Commodity Futures Trading Commission, 12 November 2014: over USD 1.4 billion across five banks, for attempted manipulation of benchmark rates.
- US Department of Justice, 20 May 2015: criminal fines of more than USD 2.5 billion on four banks that agreed to plead guilty to conspiring to manipulate the price of euros against dollars.
- The FX Global Code, published in May 2017 and updated in December 2024, now sets out conduct principles for market participants, including principles on pre-hedging.
What this means for the SMC story
The enforcement cases show that order flow matters and that dealers with large orders can and did affect a benchmark. They do not show a market organised around collecting retail stops. The conduct was about benchmark orders from institutional clients. The UK regulator also found that traders at times tried to trigger clients' stop-loss orders. Those were orders the banks' own clients had left with them, which the traders could see, not stops sitting on retail platforms.
When a chart shows a spike through a high that reverses, the smart money label is a liquidity sweep. Plainer candidates for the cause include:
- Benchmark flows: fixing orders executed around a known window.
- Option hedging: dealers adjusting hedges as price nears a strike.
- Thin liquidity: few resting orders, so a normal trade moves price a long way.
- Stop cascades: clustered stops from all kinds of traders triggering each other.
- News: a data release or headline repriced in seconds.
Play it from the other side
The Institutional tier of ForexLens puts you in the seat this page describes. The pages below describe how each game plays.
- The Dealer's Book: quote two-way prices to clients, shade them to manage your inventory, and hedge only what is beyond your limit.
- The Fix Window: net the benchmark orders, work what is left across the 4pm window and see the footprint you leave.
- Order Flow Truth: one price run shown as candles, an order book and a trade tape. Decide which of four readings the evidence supports.
Worked example
Here is a hypothetical afternoon on a dealing desk. Every figure in this example is invented for illustration. In the hour before 16:00 London time, clients send fixing orders: instructions to trade at whatever the benchmark rate turns out to be. One asset manager needs to buy 300 million euros against dollars. A pension fund needs to sell 180 million. A corporation needs to buy 40 million.
The dealer nets them. Buys of 340 million against sells of 180 million leave 160 million euros to buy. The dealer has promised each client the fix rate, whatever it is. If the dealer does nothing and the euro rises into the fix, it must deliver euros at a price below what they now cost. So it has to buy roughly 160 million euros in the market, and it has to do so around the window, because buying much earlier or later exposes it to a different price from the one it has promised.
If it spreads the buying evenly across the calculation window, its average price should sit close to the benchmark. If it bunches the buying into a few seconds, or if several dealers hold orders in the same direction that day, the buying pushes the price up through the window. On a five minute retail chart this prints as a sudden rally at 16:00 London, perhaps through a recent high, followed by a drift back once the orders are done.
A chart reader using smart money vocabulary would label it a sweep of buy-side liquidity and a reversal. The label describes the shape correctly. The cause, in this example, was a pension fund, an asset manager and a corporation valuing their books, and a dealer hedging a promise.
Takeaway: The same candle can be a "liquidity sweep" on the chart and an ordinary benchmark hedge on the desk.
Common mistakes
- Thinking banks can see your stop. Retail stop orders sit on the retail platform. They are not broadcast to the interbank market, and no dealer has a chart of them.
- Treating "smart money" as one actor. The market is thousands of participants with different goals: hedging, investing, paying invoices, speculating. Most are not thinking about retail traders at all.
- Confusing two meanings of liquidity. To a dealer, liquidity is how much can be traded without moving the price. In SMC, it means resting stop orders. The same word, two different things.
- Reading the fix scandal as proof of SMC. The misconduct that the authorities punished was about benchmark orders, confidential client information and, in some cases, stop orders that clients had left with the bank itself. It was not a scheme to raid stops on retail platforms.
- Assuming every spike is engineered. Thin liquidity, option hedging, news and stop cascades all produce sharp wicks through obvious levels with nobody planning them.
What is and is not established
What is established: the structure of the market described here is documented by central banks and regulators. The Bank for International Settlements surveys turnover every three years. Benchmark fixes exist, large orders are tied to them, and price behaviour around the fix window has been studied by regulators. Misconduct around the fix was investigated and punished. None of that is in dispute.
What is also established, in a general way: research on a dealing bank's order book found stop-loss orders clustered near round numbers, and found that when price reaches such a cluster the triggered orders can extend the move for a while. That is a real microstructure effect. Obvious highs and lows are less studied, but they are natural places for stops as well. It gives the SMC idea of liquidity above highs and below lows a factual footing.
What is not established: that a retail trader can tell from a candle chart who caused a move or what they intended, or that reading moves as engineered raids produces profit. The chart shows the footprint, not the foot. The honest position is that the pattern is observable, the explanation is uncertain, and several ordinary explanations fit at least as well. The page on whether SMC works takes that question head on.
Questions people ask
What is the 4pm London fix?
It is a benchmark exchange rate, published for many currency pairs, calculated from trades and quotes in a short window around 16:00 London time. Asset managers and index funds use it to value foreign holdings, so they often ask dealers to trade at that rate. Since 15 February 2015 the window has been five minutes wide, two and a half minutes either side of the hour. Before that it was one minute.
Do banks hunt retail stop losses?
There is no good evidence that interbank dealers target the stops of retail traders, and they have no view of orders sitting on retail platforms. Price does often spike through obvious highs and lows, because stops from all kinds of participants cluster there and thin liquidity exaggerates moves. A retail platform that is the counterparty to its clients has a separate conflict of interest, which is a matter of regulation and disclosure rather than of chart patterns.
Who are the market makers in forex?
Market makers, or dealers, are firms that continuously quote a buying price and a selling price and stand ready to trade at them. In foreign exchange they are large banks and, more and more, non-bank electronic liquidity providers. They earn the spread and manage the inventory that client trades leave them with. The ForexLens game The Dealer's Book lets you play that role with simulated client flow.
How big is the forex market?
The Bank for International Settlements triennial survey for April 2025 put turnover at about 9.6 trillion US dollars a day, a figure later revised to 9.5 trillion. FX swaps made up about 42% of that, spot trades 31%, outright forwards 19% and options about 7%. The US dollar was on one side of about 89% of all trades. Spot trading by retail accounts is a small part of the total.
Does the fix explain moves that SMC calls a liquidity sweep?
Sometimes it plausibly does. Benchmark orders put real, one-directional buying or selling into a known window, and the move often fades afterwards, which on a chart looks like a raid and reversal. That does not mean every sweep is a fix effect, or that the fix can be traded for profit. It means a plainer explanation exists and should be considered. The game Order Flow Truth practises exactly that comparison.
Practise it
Reading about a rule is not the same as applying it. These games generate a fresh chart each round, ask you to find it first, and grade you against the same rule.
The Dealer's Book
Play the dealer. Quote two-way prices to clients, shade them to manage your inventory, and hedge what you cannot hold.
The Fix Window
Benchmark orders arrive for the 4pm fix. Net them, work the rest around the window, and see the footprint a chart later calls a stop hunt.
Order Flow Truth
One move, three views: candles, order book and tape. Decide whether the smart money story or a plainer one fits the evidence.
