The WMR 4pm London closing benchmark provides a reference rate for valuation and other agreed uses. Orders can cluster around its calculation window, but a particular spread or price move is not guaranteed each weekday. Check the methodology and applicable holiday arrangements before treating the time as an execution signal.

The median of midpoints does not reproduce the medians of both sides

Do not interchange aggregation steps simply because both produce a “middle”. That shortcut can produce a different reference price.

Scroll sideways to see the full diagram

The median of midpoints does not reproduce the medians of both sidesThree fictional Bid/Offer pairs are 1.1000/1.1020, 1.1002/1.1003 and 1.1004/1.1008. Their midpoints are 1.10100, 1.10025 and 1.10060. The median midpoint is 1.10060. Separate medians of Bid 1.1002 and Offer 1.1008 average to 1.10050. The difference is one pip when a pip is 0.0001. This is an algebraic counterexample, not a WMR reconstruction.Aggregation order changes the resultThree fictional EUR/USD quotes; not a WMR calculation.RecordBidOfferMidpoint11.10001.10201.1010021.10021.10031.1002531.10041.10081.10060Median of three midpoints: 1.10060Mean of side medians: 1.10050Difference = 0.00010 = 1 pip in this model.
An independent counterexample with three quote pairs. The complete benchmark methodology is not applied.

Assumptions: Three fictional EUR/USD Bid/Offer pairs, synchronous within each record: 1.1000/1.1020, 1.1002/1.1003 and 1.1004/1.1008 USD per EUR. All are positive and Offer ≥ Bid. Each model midpoint is (Bid + Offer) / 2. A pip is 0.0001. Deliberately varied spreads serve only the arithmetic.

The midpoints are 1.10100, 1.10025 and 1.10060. Their sorted median is 1.10060. The median Bid is 1.1002 and the median Offer is 1.1008, whose mean is 1.10050. These operations differ here by 1 pip. This is neither slippage nor profit: no transaction was executed.

WMR methodology v30, section 4.4.2 describes independent side medians followed by calculation of Mid, together with validation, data selection and spread rules. Our three records omit those procedures and do not reconstruct WMR. They show why replacing the stated sequence with the median of midpoints would require a different definition.

What the London Fix actually is

Here, London Fix means WMR 4pm Closing Spot Rates, a currency benchmark with reference time 16:00 Europe/London. FTSE International Limited within LSEG is its administrator. Spot, forward and NDF services have distinct rules; identify the series, currency pair, quote direction and Bid, Offer or Mid before comparing it with an account price.

A benchmark can serve valuation, index calculation and contractual conversion. The relevant portfolio or contract determines which rate is used; not every participant requires the same benchmark. A reference rate is not automatically an executable price for an individual order.

An investor whose valuation uses a fixing benchmark may seek execution linked to that rate to manage tracking differences. Actual execution price, fees and the provider’s agreed method still matter. Benchmark-linked execution does not make every order non-speculative or remove risk.

Why some orders are linked to the benchmark

Our hypothetical fund needs to buy EUR 25,000 for USD. A specified WMR Mid is 1.1000 USD per EUR; we do not reconstruct its calculation. The broker quote when submitting the order is Bid 1.1001 / Ask 1.1003. Assume a full purchase fill at Ask 1.1004 and a USD 2 commission, with no other charges or currency conversion. The benchmark value is USD 27,500; the actual purchase is USD 27,510 before commission and USD 27,512 including it.

The purchase costs USD 10 more than the benchmark before commission, or USD 12 including it. Slippage against the displayed Ask is only (1.1004 − 1.1003) × 25,000 = USD 2.50, one pip at the assumed pip size of 0.0001. The entire deviation from WMR is not slippage. The quote’s two-pip spread is already reflected in the purchase price and should not be added again. This is one purchase-cost comparison; a completed trade’s P/L also needs its exit and remaining costs.

Concentrated benchmark-related orders can affect available liquidity and execution. Their direction, size and impact vary by date and currency. A month-end label alone does not establish a predictable volatility spike, its direction or the day’s strongest move.

The 2013 manipulation scandal and what followed

The FCA’s historical findings cover deficient controls from 1 January 2008 to 15 October 2013. Traders shared confidential client information and attempted to manipulate currency rates, including by colluding with other firms. The investigation included the 4pm WMR and 1:15pm ECB fixes. These are specific enforcement findings, not evidence that every current fixing move or stop activation is manipulation.

On 12 November 2014, the FCA announced fines totalling GBP 1,114,918,000 on five banks for failings in G10 spot FX controls. Separately, the FSB’s September 2014 final report recommended widening the then one-minute WMR window and improving execution transparency and controls. The FSB work was independent of enforcement investigations. A recommendation and a later implemented calculation rule should be identified separately.

WMR methodology v30 (January 2026) describes five-minute spot sampling around the reference time. For Trade Currencies, qualified trades take precedence subject to a minimum-data threshold; order data provide a fallback. Non-Trade Currencies use quote snapshots. Bid and Offer medians are calculated separately, with validation, spread and rounding rules; local-close currencies have additional treatment. This is neither a simple average nor a VWAP of all FX transactions. The methodology targets completion and publication within 15 minutes after the reference time, rather than promising the final rate at 16:00.

How the fix looks from a retail trader's seat

An individual account quote does not expose the full institutional order flow. Review the provider’s execution policy and the terms of any fixing-linked service; an ordinary retail order is not automatically a contract to execute at WMR. Fixing activity does not predetermine an order’s price or profitability.

For the 16:00 London reference, the five-minute window is 15:57:30–16:02:30 Europe/London. The reference is 16:00 UTC in British winter and 15:00 UTC in British summer. In 2026 it is 17:00 Europe/Warsaw in both seasons because the UK and Poland change clocks on the same dates. A broker server clock can use a different zone; neither the window’s end nor a chart label establishes when your program receives the final result.

Consider a separate hypothetical data-availability test: your program receives today’s final benchmark at 16:07 London time. A decision at 16:01 cannot use it; that would introduce information six minutes early. Store both the reference timestamp and the actual receipt timestamp. After receipt, the value may inform a new decision, but it does not make a past quote or a WMR-priced retail fill available. The invented 16:07 is not a service delivery promise.

Common misconceptions about the fix

Historical misconduct does not prove that every present move is manipulation. Equally, benchmark reforms do not establish that a particular move is legitimate. Price observations require separate evidence before attributing motive or conduct.

A reversal visible after the event is not a tested trading rule. Any fixing-window strategy needs specified entry and exit conditions, costs, execution assumptions and data availability. The sessions guide provides context, but cannot establish suitability for a majority of traders.

What to do tomorrow

  1. Mark 16:00 in the Europe/London time zone and the applicable calculation window in your calendar. Check the benchmark’s holiday arrangements and your own trading rules. A reminder supports awareness; it does not prevent an order or identify the day’s most unpredictable minute.
  2. Observe the complete 15:57:30–16:02:30 window for a dated sample and record Bid, Ask and price changes. Several days can illustrate observations, but cannot establish that every fix differs from surrounding hours in the same way.
  3. Review open positions and pending orders before the fixing window. If changing a stop, recalculate the cash risk and position size and check execution terms. Widening a stop alone increases the planned loss and does not guarantee protection against slippage.
  4. Mark month-end dates and test whether your method’s costs and execution differ on those dates. Apply a predefined risk limit and exclusion rule if your evidence supports it; do not assume month-end always produces the largest spike.