MEASURED · 68,119 ENTRIES
How likely is a $3, $5 or $10 gold stop to be hit within an hour?
More likely than not, unless the stop is $10 or wider. On a year of one-minute gold (XAUUSD) bars, a stop $3 from an arbitrary entry was hit within the hour on 73.5% of entries, a $5 stop on 59.0% and a $10 stop on 33.8%. The New York morning, 8 a.m. to noon ET (7 to 11 a.m. CT), was hardest on stops: 81.3%, 70.0% and 47.0%. The New York afternoon was the easiest.
The quarter mattered more than the session. The same $5 stop was hit 32.0% of the time in Aug–Sep 2025 and 72.2% in Jan–Mar 2026. A fixed dollar stop does not have a fixed hit rate.
How it was measured
Every five minutes on the clock, from August 1, 2025 to August 4, 2026, we opened two imaginary trades at the open of the one-minute bar: one long, one short. No signal and no filter. Each got a stop $3, $5 or $10 away and an hour to live. A long’s stop counted as hit if any bar’s low in the next 60 minutes reached it, a short’s if any bar’s high did. That gave 68,119 entries, each opened both ways, on 261 trading days.
Entries from 4 p.m. ET were left out, because their hour runs into the 5 p.m. ET daily break. So were 115 entries whose hour had fewer than 55 one-minute bars. Most fell on US holidays when the feed stopped early, Thanksgiving and Christmas Eve among them; the rest were short gaps in the data and the last entries before the window ends. A closed market is not a stop that held. MT5 bars are bid prices, so the main figures are price only. The spread comes back in later.
Sessions are fixed on the US Eastern clock, as in our study of gold’s New York session range: Asia 6 p.m. to 3 a.m. ET, London 3 a.m. to 8 a.m., New York morning 8 a.m. to noon, New York afternoon noon to 4 p.m..
The answer by session
| Session, ET (CT) | GMT, US winter | $3 stop | $5 stop | $10 stop | Hit 1 in 4 at |
|---|---|---|---|---|---|
| Asia6 p.m. to 3 a.m. (5 p.m. to 2 a.m.) | 23:00–08:00 | 72.0% | 57.2% | 32.5% | $12.30 |
| London3 a.m. to 8 a.m. (2 a.m. to 7 a.m.) | 08:00–13:00 | 73.4% | 58.0% | 29.7% | $11.37 |
| New York morning8 a.m. to noon (7 a.m. to 11 a.m.) | 13:00–17:00 | 81.3% | 70.0% | 47.0% | $17.76 |
| New York afternoonnoon to 4 p.m. (11 a.m. to 3 p.m.) | 17:00–21:00 | 68.8% | 53.0% | 28.6% | $11.07 |
| All four | 73.5% | 59.0% | 33.8% | $12.79 |
Long and short entries averaged. GMT is one hour earlier while the US is on daylight time. Hit 1 in 4 at: the move against the entry that ordinary movement reached within the hour on a quarter of entries. A stop that wide was hit about one time in four.
The New York morning is the outlier. A $10 stop, hit on 28.6% to 32.5% of entries in the other three sessions, was hit on 47.0% between 8 a.m. and noon ET. Asia and London were close to each other at every size, and the afternoon was the gentlest of the four.
The last column is the practical one. To get the odds of being stopped by ordinary movement down to one in four, a stop had to be $17.76 away in the New York morning, against $11.07 in the afternoon. One in ten took $29.77 and $19.66. Longs and shorts were stopped at almost the same rate over the year (34.5% and 33.1% at $10), so the table averages them.
How fast the stop goes
| Stop | 5 min | 15 min | 30 min | 60 min | Median minute |
|---|---|---|---|---|---|
| $3 | 30.8% | 51.5% | 63.5% | 73.5% | 7 |
| $5 | 15.1% | 33.0% | 46.2% | 59.0% | 13 |
| $10 | 3.8% | 12.0% | 21.4% | 33.8% | 23 |
All sessions. Share of all entries stopped by then; median minute among the stops that were hit.
Tight stops go early. Of the $3 stops hit within the hour, half were gone by minute 7; 30.8% of all $3 entries were stopped inside five minutes, and 44.2% in the New York morning. A $10 stop lasted longer, a median 23 minutes when it was hit, but in the New York morning 20.8% of entries had lost it within a quarter of an hour.
Hour by hour
| Entry hour, ET (CT) | $3 stop | $5 stop | $10 stop |
|---|---|---|---|
| 6 p.m. (5 p.m.) | 67.2% | 51.0% | 28.4% |
| 7 p.m. (6 p.m.) | 73.9% | 60.4% | 37.9% |
| 8 p.m. (7 p.m.) | 77.9% | 66.1% | 41.5% |
| 9 p.m. (8 p.m.) | 75.9% | 61.5% | 37.6% |
| 10 p.m. (9 p.m.) | 71.0% | 55.0% | 27.9% |
| 11 p.m. (10 p.m.) | 66.4% | 49.6% | 23.5% |
| midnight (11 p.m.) | 67.9% | 52.9% | 29.6% |
| 1 a.m. (midnight) | 73.6% | 57.9% | 33.3% |
| 2 a.m. (1 a.m.) | 74.1% | 59.9% | 32.1% |
| 3 a.m. (2 a.m.) | 73.7% | 60.0% | 31.9% |
| 4 a.m. (3 a.m.) | 71.4% | 55.3% | 26.9% |
| 5 a.m. (4 a.m.) | 71.7% | 55.3% | 25.1% |
| 6 a.m. (5 a.m.) | 72.0% | 55.5% | 27.4% |
| 7 a.m. (6 a.m.) | 78.1% | 63.8% | 37.2% |
| 8 a.m. (7 a.m.) | 82.1% | 71.5% | 48.0% |
| 9 a.m. (8 a.m.) | 84.8% | 75.8% | 55.7% |
| 10 a.m. (9 a.m.) | 81.9% | 70.5% | 47.3% |
| 11 a.m. (10 a.m.) | 76.2% | 62.1% | 36.9% |
| noon (11 a.m.) | 72.6% | 57.1% | 30.7% |
| 1 p.m. (noon) | 69.9% | 54.6% | 30.5% |
| 2 p.m. (1 p.m.) | 67.5% | 52.2% | 29.2% |
| 3 p.m. (2 p.m.) | 65.2% | 48.0% | 24.0% |
The hardest hour was the one starting at 9 a.m. ET, when a $5 stop went on 75.8% of entries and a $10 stop on 55.7%. The easiest were 3 p.m. ET (48.0% at $5) and 11 p.m. ET (49.6%). Asia is not one quiet block: the 8 p.m. ET hour hit a $5 stop on 66.1% of entries, the most of any hour outside New York hours.
The quarter mattered more than the session
| Quarter | Days | Median price | $3 stop | $5 stop | $10 stop | $5, NY morning |
|---|---|---|---|---|---|---|
| Aug–Sep 2025 | 43 | $3,474 | 52.2% | 32.0% | 9.5% | 49.7% |
| Oct–Dec 2025 | 65 | $4,141 | 73.7% | 58.4% | 31.2% | 70.3% |
| Jan–Mar 2026 | 63 | $4,963 | 82.7% | 72.2% | 50.7% | 80.5% |
| Apr–Jun 2026 | 64 | $4,555 | 78.5% | 65.5% | 38.3% | 73.8% |
| Jul–Aug 2026 | 26 | $4,064 | 73.0% | 56.5% | 28.2% | 67.5% |
The last row runs to August 4, 2026, 26 trading days.
The ranking held: the New York morning had the highest $5 rate in every quarter. The level did not. A $5 stop went from 32.0% to 72.2% across quarters, while the whole year’s gap between the hardest and easiest session was 17.0 points.
Price alone does not explain it. From Aug–Sep 2025 to Jan–Mar 2026 the median hour’s move against an entry went from $3.18 to $10.22, 3.2 times as far, while the median price rose 43%. That is the argument for sizing a stop to volatility rather than to a dollar amount, which our ATR stop study tested against a fixed stop at the same average distance.
What the spread adds
A real trade pays the spread. A long fills at the ask and is stopped by the bid; a short fills at the bid and is stopped by the ask. Either way the bid only has to travel the stop minus the spread. This account’s median spread was $0.16 in every session. Measured from the fill, the $3 hit rate rose from 73.5% to 75.1% and the $10 rate from 33.8% to 34.6%. MT5 keeps one spread value per one-minute bar, so a single tick can be wider than the value used here. The hourly spread study shows the hours when this account’s spread does widen, and the half hour after the 6 p.m. ET reopen is measured minute by minute in gold’s spread at the daily break.
What to do with this if you run a bot
Treat $3 and $5 as noise-sized stops. In every session, ordinary movement reached $3 on at least 68.8% of arbitrary entries within the hour, and $5 on at least 53.0%. A bot with a stop that tight is stopped more often than not unless its entries are much better than arbitrary.
Give the New York morning more room, or less size. Matching the afternoon’s one-in-four odds between 8 a.m. and noon ET took $17.76 instead of $11.07. At the same dollar risk per trade, the wider stop means a smaller lot.
Re-measure, or scale to volatility. A $5 stop that was hit 32.0% of the time in one quarter was hit 72.2% of the time two quarters later. A fixed dollar stop set in a calm month is a different stop in a fast one.
Check the units and the clock. Whatever EA you run, the Tech Kick Trading Bot included, check whether its stop input is in dollars, points or ATR, and which clock its session times use. This feed’s server clock is UTC, so 8 a.m. ET is 12:00 on the server in US summer and 13:00 in winter. Trading from the US? Read whether US traders can use MT5 trading bots first: the instrument measured here is a spot gold CFD.
Method and limits
356,223 one-minute XAUUSDm bars from an Exness MetaTrader 5 feed, August 1, 2025 to August 4, 2026. The bar clock is UTC: the daily break moved by an hour in UTC when US clocks changed, so it stayed at 5 to 6 p.m. ET. ET, CT and GMT come from the IANA time zone database.
Entries overlap, one every five minutes and each held an hour, so neighbors share most of their bars and are not independent. We give no confidence interval from the entry count; the quarter table shows how far the answer moved instead. Entries are arbitrary by design. A strategy that enters after a fast move or at a news release will see its stops hit more often than this, and one that enters in quiet conditions less. Only the touch of the stop is measured, not whether a target would have been reached first.
Our hourly study ran a five-minute version of this test (long entries only, all hours pooled) and found 72.7%, 58.3% and 34.2% for $3, $5 and $10. This one-minute version lands within a point of it. One symbol, one broker, one year. Every other figure comes from one script, research-gold-stop-hit-by-session.py, run on the same bar archive as our other studies.
Common questions
What is a good stop loss size for gold in dollars?
There is no fixed number, because gold's movement changed so much within one year. From August 2025 to August 2026, ordinary movement reached $12.79 against an arbitrary entry within the hour on a quarter of entries, and $22.34 on a tenth. In the New York morning those distances were $17.76 and $29.77. Size the stop to current volatility, then size the lot to the stop.
Why does my gold stop loss get hit so quickly?
Because a few dollars sits inside gold's ordinary movement. Of the $3 stops that were hit within the hour, half went within 7 minutes, and 30.8% of all $3 entries were stopped inside five minutes. Between 8 a.m. and noon ET it was 44.2%. No signal was involved: these were arbitrary entries.
Which session is easiest on a tight stop?
On this year of data, the New York afternoon: a $5 stop was hit within the hour on 53.0% of entries from noon to 4 p.m. ET (11 a.m. to 3 p.m. CT), against 70.0% in the morning. The quietest single hours were 3 p.m. and 11 p.m. ET. Asia was not one quiet block: its 8 p.m. ET hour was the busiest outside New York hours.
Are long or short stops hit more often on gold?
Almost equally over the full year: a $10 stop was hit on 34.5% of long entries and 33.1% of short ones. Single quarters leaned one way or the other; in Aug–Sep 2025 it was 8.2% and 10.7%. The tables average the two.