Part 3: The week and the news · Chapter 7 of 14 · 15 min

The weekend and Sunday evening

Across 42 weekends the price of gold moved between the Friday close and the Sunday reopening by a median 10.0 $ (90th percentile 44.0 $, maximum 91.9 $), with a random direction (50% up). The gap does not anticipate Monday (65% ± 16 points) and the "fill" is true only because gold moves about 100 $ every day. The Sunday reopening has a first five-minute candle 5.0 times the median.

How gold stops and how it restarts

The gold market stops once a week. For COMEX futures the close is Friday at 17:00 New York time (23:00 in Rome) and the reopening is Sunday at 18:00 New York time (00:00 Monday in Rome). During those forty-eight hours futures do not trade and almost all brokers stop quoting, while the news keeps coming. If you have a CFD account, your chart stops on Friday and restarts on Sunday with a jump between the last Friday price and the first Sunday price.

If news arrives over the weekend (a conflict, a statement from a central bank, an event that moves the dollar), the Sunday opening price already contains all of it. Whoever had a position open finds a price different from the one on which the stop was placed.

A note on the data. Ours come from the Binance perpetual, which quotes on Saturday and Sunday too. For it the "gap" is not a jump: it is the move of the prices across the forty-eight closed hours, the same move your broker shows you as a gap at the reopening. The sizes are similar, not identical: your broker's reopening price depends on its liquidity providers. When you read "gap" here, read: Friday 17:00 New York close against Sunday 18:00 New York open of the perpetual.

How far gold moves over the weekend

We took all the weekends in the sample: 42, from the Friday close to the Sunday reopening (between 46 and 50 hours of closed market). For each we computed the difference between the opening price of the first Sunday candle and the close of the last Friday one.

MeasureValue
Weekends measured42
Mean signed move+4.3 ± 8.3 $
Weekends up21 (50%, ± 15 points)
Median absolute move10.0 $
Mean absolute move18.6 $
75th percentile25.5 $
90th percentile44.0 $
Maximum91.9 $ (Sunday 1 March 2026)
Median range of the weekend (high-low)27.6 $

Three readings. The direction is random: half of the weekends go up, half go down, and the mean move (+4.3 $) sits inside the error (± 8.3 $). The weight is small compared with a normal week: over the same length of time with the market open (49 hours, starting between Monday and Wednesday) the price moves by a median 56.0 $ in absolute value, so the weekend weighs 0.18 times as much (10.0 $ against 56.0 $). The maximum, however, is large: 91.9 $ at the start of March, almost a whole day of movement.

Figure 1. The widest weekend in the sample (27 February - 2 March 2026, UTC hours): the closed market (grey zone), with candles smaller than on weekdays, and the Sunday reopening at +91.9 $ from the Friday close. Below, the range of each candle.Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 27 Feb 2026 08:00 UTC to 2 Mar 2026 07:00 UTC.
Read the chart in words

H1 gold chart with the weekend highlighted in grey, the Friday close line and the Sunday reopening jump of 91.9 dollars.

WhenWhatPrice
27 Feb 2026 22:00 UTC → 1 Mar 2026 22:00 UTCmarket closed
27 Feb 2026 18:00 UTCFriday close 5276.85,276.83
1 Mar 2026 23:00 UTCreopening: +91.9 $5,368.77

Times in UTC.

Figure 2. A typical weekend (5-8 June 2026, UTC hours): the reopening is 10.0 $ below the Friday close, exactly the median of the 42 weekends.Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 5 Jun 2026 07:00 UTC to 8 Jun 2026 06:00 UTC.
Read the chart in words

H1 gold chart with the June 2026 weekend highlighted in grey and the reopening gap of minus 10 dollars.

WhenWhatPrice
5 Jun 2026 21:00 UTC → 7 Jun 2026 21:00 UTCmarket closed
5 Jun 2026 17:00 UTCFriday close 4332.14,332.10
7 Jun 2026 22:00 UTCreopening: −10.0 $4,322.07

Times in UTC.

The real risk: the gap against you

Whoever holds a position over the weekend does not risk the stop they chose: they risk the gap. If you have a stop at 20 $ and the price reopens 50 $ against you, the order is filled at the first available quote and the loss is 50 $ per ounce, not 20 $. With a good broker the stop protects you from continuous movement, not from jumps.

How often did the price move at least a given amount (in either direction)?

Gap of at leastWeekends out of 42
10 $21 (50%)
20 $15 (36%)
30 $8 (19%)
50 $4 (10%)
100 $0

For an open position only the opposite direction counts, so, if the direction is random as the data suggest, about half of these cases. With a random direction, holding a position with a 20 $ stop over the weekend means facing a jump past the stop in about one case out of six.

A worked example. You have a 10,000 $ account and accept losing 1% per trade, that is 100 $. With a 20 $ stop the position would be 5 ounces. If you reason on the 90th-percentile gap (44 $), the position that keeps the loss at 100 $ in the worst of the common cases is about 2 ounces. The difference is the price of prudence: less than half of the position, for the hours when you cannot act. Or you close on Friday.

The choice, in the end, is among three things: close before the close, reduce the size, or accept the risk with the size computed on the gap. There is no single right answer: it depends on what a jump costs you (in dollars, not as a percentage of the stop) and on what you think might come out over the weekend. If you have no good reason to stay, closing is the choice with the fewest surprises.

Does the gap say anything about Monday?

Many traders think so, in two ways: "the price will come back and fill the gap" and "a gap in one direction continues". We measured both.

The direction. In the 34 weekends with a gap of at least 5 $, the first trading day (Sunday 18:00 - Monday 17:00 New York) goes in the same direction in 22 cases (65%, ± 16 points). The mean return of the Monday is −9.1 ± 21.0 $. With 34 cases the difference from 50% cannot be told apart from chance. The direction of the weekend compared with that of Friday also matches in 21 cases out of 34 (62%, ± 16): same conclusion. There is no usable information.

The fill. The "gap fill" is measured like this: after a gap, does the price come back to the Friday closing level? With a gap of at least 10 $ (21 cases) it does so by the end of Monday in 10 cases (48%) and within five days in 17 (81%). With a gap of at least 20 $ (15 cases): 7 (47%) and 13 (87%). At first sight it is a strong result: eight times out of ten the gap is closed within a week.

There is a trap that shows up in many tests. We went to see what happens on the other side: how often the price moves away by the same amount, that is the same distance but in the opposite direction. With the gap of at least 10 $, by Monday the price does it in 16 cases (76%), within five days in 20 (95%). With at least 20 $: 11 cases (73%) and 14 (93%). The price does not "come back" more than it "runs away". It moves 100 $ a day, and a level 10 or 20 $ away gets touched from both sides almost always. The fill is not a property of the gap: it is a consequence of volatility. To say a gap fills you need to show that it fills more often than the price touches a level at equal distance on the other side. Here that does not happen (indeed, in this sample of 21 cases the opposite occurs, but there are too few to say so strongly).

Sunday evening: the reopening

The noisiest moment of the weekend is its end: the Sunday reopening at 18:00 New York time (00:00 Monday in Rome, 23:00 in the three spring weeks when New York is already on summer time and Europe is not).

The first five-minute candle is on average 5.0 times the median, against 1.9 times at the same time Monday to Thursday. Then it calms quickly: at 18:05 it is 2.5 times, at 18:30 1.4 times, at 19:00 1.7 times. Over a whole hour, the median range of the first hour on Sunday is 29.4 $ against 13.3 $ Monday to Thursday at the same hour (2.2 times). In the following hours Sunday stays above normal days until 22:00 (20.1, 25.2, 28.1 and 21.2 $ against 13.5, 19.0, 23.0 and 16.7 $), and at 23:00 it drops to 12.7 $, the calmest hour of the evening.

New York hour18:0019:0020:0021:0022:0023:00
Median H1 range, Sunday29.4 $20.1 $25.2 $28.1 $21.2 $12.7 $
Monday-Thursday, same hours13.3 $13.5 $19.0 $23.0 $16.7 $13.1 $

The livelier hours, 20:00 and 21:00 New York time, are the Tokyo and Shanghai opens (chapter 3): on Sunday evening the market restarts with the weekend's prices still to absorb and, right after, with Asia coming in.

Figure 3. The reopening of Sunday 7 June 2026 (UTC hours). Below, the median H1 range by New York hour on Sunday evening (bars) and Monday to Thursday (line).Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 7 Jun 2026 22:00 UTC to 8 Jun 2026 20:00 UTC.
Read the chart in words

H1 gold chart from the Sunday evening reopening with, below, the comparison between the median range of Sunday and that of weekdays.

WhenWhatPrice
7 Jun 2026 22:00 UTCSunday reopening4,337.50

Times in UTC.

Monday in Asia

Sunday evening and Monday's Asian morning count for a bit more than usual: the share of the trading day's activity (the sum of the hourly ranges) falling between 18:00 and midnight New York time is 32% on Monday (43 days) and 25% on the other days (168). The Asian range (00:00-07:00 UTC) of Monday has a median of 57.4 $, against 49.6 $ for the other days. The permutation test gives p = 0.095: an in-between value, which allows neither yes nor no. By day: Monday 57.4 $, Tuesday 52.5 $, Wednesday 49.9 $, Thursday 51.4 $, Friday 43.0 $. Friday is the lowest. The last hour of Friday (16:00-17:00 New York) has a median range of 10.4 $, against 9.7 $ Monday to Thursday: there is no rush before the close.

Figure 4. Exercise: four H1 candles around a weekend (21-25 August 2026), marked with the letters A-D.Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 21 Aug 2026 01:00 UTC to 25 Aug 2026 04:00 UTC.
Read the chart in words

H1 gold chart from 21 to 25 August 2026 with four candles marked with the letters A, B, C and D.

WhenWhatPrice
21 Aug 2026 19:00 UTCPoint A4,629.04
22 Aug 2026 15:00 UTCPoint B4,609.76
23 Aug 2026 22:00 UTCPoint C4,627.38
24 Aug 2026 13:00 UTCPoint D4,684.81

Times in UTC.

What to do with it

  1. Before holding a position over the weekend, compute the gap, not the stop. The median gap is 10 $, and one time in ten it exceeds 44 $. Size the position on the 90th percentile or close.
  2. Do not use the gap as a signal. Monday's direction (65% ± 16 points) and its relation to Friday (62% ± 16) cannot be told apart from chance.
  3. Do not count on the "gap fill". The price returns to the Friday level, but it also moves away by the same amount on the other side. A near level gets touched almost always.
  4. Do not enter in the first minutes after the reopening. The first five-minute candle is 5.0 times the median and after half an hour it is 1.4. Quotes are wide and market orders can be filled badly. Wait at least ten minutes.
  5. Check your broker's hours. The reopening can happen at 23:00 or 00:00 Rome time, depending on the week and the broker.

Procedure in four steps

  1. On Friday afternoon decide: close, reduce or hold. If you hold, size the position on the 90th-percentile gap (about 44 $).
  2. Do not look at the gap as a signal for Monday: the data do not support it.
  3. At the reopening wait at least ten minutes before entering, and check the spread before any order.
  4. If a strategy says "buy the gap", check how often the price moves away by the same amount: that is the missing comparison.

Exercises

Chapter exercises

1. On the chart the letters A-D are on four H1 candles around a weekend (August 2026). Which letter is the Sunday evening reopening candle, the first quote after the weekend?

Chart exercise. Data: gold, XAUUSDT perpetual future (real candles) · 1 hour (H1) · from 21 Aug 2026 01:00 UTC to 25 Aug 2026 04:00 UTC.
Read the chart in words
WhenWhatPrice
21 Aug 2026 19:00 UTCPoint A4,629.04
22 Aug 2026 15:00 UTCPoint B4,609.76
23 Aug 2026 22:00 UTCPoint C4,627.38
24 Aug 2026 13:00 UTCPoint D4,684.81

Times in UTC.

2. In 15 weekends out of 42 the price moved at least 20 $ between the Friday close and the Sunday reopening. What does it mean for someone holding a position with a 20 $ stop?

3. After a gap of at least 10 $ the price returns to the Friday close within 5 days in 81% of cases (17 of 21), but moves away by the same amount on the other side in 95% (20 of 21). What do you conclude?

4. The first five-minute candle after the Sunday reopening is worth 5.0 times the median, and at 18:30 it is 1.4 times. What do you do with a market order at the reopening?

Have a question or want to share your exercise?

Post in the community, or join the free signals room on Telegram.

Educational content, not financial advice. Trading involves risk.