The short answer
A run of down days does not, on its own, make gold more likely to bounce. We looked at every streak of falling daily closes in COMEX gold futures from 30 August 2000 to 19 November 2026, 6,582 closes in all. After four, five or six lower closes in a row, the next day, the next week and the next month were up about as often as on any normal day. Streaks also kept going more often than most people think: after four straight falls, a fifth came in just over half the cases.
The one clear difference came from the bigger trend. Streaks that happened while gold was above its 200-day moving average tended to recover over the next week. Streaks below that average more often kept falling. The samples are small, so this is context, not a signal.
That matters on a week like this one. Spot gold fell from $3,887.6 on Friday 13 November to $3,627.9 on Thursday 19 November on our own spot feed, four lower closes in a row, and dropped to about $3,520 in Asian trade on Friday. “It has fallen four days, it has to bounce” is exactly the thought this study tests.
How we counted the streaks
We kept the method simple so you can check it yourself:
- Data: daily closes of COMEX gold futures (GC=F), the same series used across this site, from 30 August 2000 to 19 November 2026.
- Down day: a close lower than the previous close. An unchanged close ends a streak.
- Streak of N: the day on which the count of lower closes in a row first reaches N. A 7-day streak therefore counts once as a 5-day streak, once as a 6-day streak and once as a 7-day streak, at the day each length was reached.
- What happened next: the change from that close to the close 1, 5 and 20 trading days later (about a day, a week and a month).
- Normal day: the same numbers measured from every close in the data, so you can compare a streak with doing nothing special at all.
We used futures closes because they give 26 years of one clean series. Spot XAUUSD, the price most retail platforms show, tracks futures closely but settles at a different time, which can change a streak by a day (see the FAQ above and our guide to spot gold vs gold futures).
What happened after 2 to 8 down days in a row
| Streak reached | Times | Next day up | Median next day | 5 days later up | Median 5 days | 20 days later up | Median 20 days |
|---|---|---|---|---|---|---|---|
| Any day (baseline) | 6,581 | 52.5% | +0.05% | 55.5% | +0.29% | 56.3% | +0.74% |
| 2 down days | 748 | 54.8% | +0.12% | 52.2% | +0.10% | 58.3% | +0.83% |
| 3 down days | 334 | 57.8% | +0.13% | 54.5% | +0.29% | 56.9% | +0.97% |
| 4 down days | 140 | 48.6% | −0.02% | 54.3% | +0.15% | 59.3% | +0.99% |
| 5 down days | 71 | 56.3% | +0.11% | 53.5% | +0.05% | 56.3% | +1.09% |
| 6 down days | 31 | 41.9% | −0.22% | 54.8% | +0.14% | 64.5% | +2.13% |
| 7 down days | 18 | 44.4% | −0.09% | 38.9% | −0.21% | 66.7% | +1.80% |
| 8 down days | 10 | 40.0% | −0.14% | 80.0% | +0.76% | 80.0% | +1.86% |
How to read the table: after gold had closed lower five days in a row, which happened 71 times, the next close was higher 56.3% of the time. On an ordinary day the next close was higher 52.5% of the time. A gap of about four points on 71 cases is the kind of difference that chance alone produces often.
Three things stand out:
- The next day was a coin flip. For streaks of 2 to 5 days, the chance of an up day stayed within a few points of normal. For 6, 7 and 8 days it was actually lower than normal, though those samples are tiny (31, 18 and 10 cases).
- The next week was no better than normal. Five sessions after a 4- or 5-day streak, gold was up 53-54% of the time, against 55.5% for any day.
- The 20-day numbers look better after long streaks, with medians of about +1.8% to +2.1% after 6 to 8 down days. But with 10 to 31 cases, a handful of big rebounds in 2006, 2008 and 2011 drive those numbers. A trader on leverage also has to survive the first few days to see a 20-day result.
Sample size matters. With 71 cases, one result moving from “down” to “up” changes the percentage by about 1.4 points. With 10 cases it changes it by 10 points. Read the rows for 7 and 8 days as stories, not statistics.
Streaks last longer than you think
The other half of the question is: once gold has fallen several days, how often does it fall again? We counted how often a streak that reached one length went on to the next.
| After this many down days | Times | Next day also down | Share |
|---|---|---|---|
| 1 | 1,716 | 748 | 43.6% |
| 2 | 748 | 334 | 44.7% |
| 3 | 334 | 140 | 41.9% |
| 4 | 140 | 71 | 50.7% |
| 5 | 71 | 31 | 43.7% |
| 6 | 31 | 18 | 58.1% |
| 7 | 18 | 10 | 55.6% |
There is no point at which a streak becomes “due” to end. After four lower closes the chance of a fifth was about one in two, and after six, a seventh came more often than not. The longest run in the data was 11 straight lower closes in September 2000, and gold had a 10-session streak in July 2015.
This is the same trap as the gambler’s fallacy, which our losing streak guide covers for trading results: past outcomes do not create a debt that the next outcome has to repay. Markets have memory in their drivers, not in the colour of the last candle.
Did gold at least stop falling soon?
A bounce trader needs not just an up close at some point but no deep further drop first. So we also checked the lowest close in the 10 sessions after the streak.
| Starting point | Times | A lower close came within 10 sessions | Median lowest close | Worst 10% of cases |
|---|---|---|---|---|
| Any day | 6,571 | 78.3% | −1.16% | n/a |
| After 3 down days | 333 | 79.9% | −1.23% | −4.43% or worse |
| After 4 down days | 140 | 76.4% | −1.09% | −4.08% or worse |
| After 5 down days | 71 | 74.6% | −1.36% | −4.88% or worse |
| After 6 down days | 31 | 77.4% | −1.23% | −5.09% or worse |
In about three cases out of four, gold printed a lower close within the next two weeks after a 5-day streak, the same as on any day. In the worst tenth of cases it closed almost 5% lower. At $3,600 an ounce, 5% is about $180, which is 180 dollars per 0.01 lot on a typical XAUUSD contract. The gap between “it usually recovers eventually” and “my stop survives the next two weeks” is where accounts get hurt.
The one thing that mattered: the bigger trend
We split the streaks by whether gold closed above or below its 200-day simple moving average on the day the streak reached its length. The 200-day average is a common way to describe the long-term trend (our moving averages lesson explains it).
| Streak and trend | Times | 5 days later up | Median 5 days | 20 days later up | Median 20 days |
|---|---|---|---|---|---|
| Any day above the 200-day | 4,590 | 55.7% | +0.31% | 56.1% | +0.70% |
| Any day below the 200-day | 1,793 | 55.3% | +0.31% | 57.9% | +1.03% |
| 5 down days, above the 200-day | 30 | 70.0% | +1.05% | 66.7% | +2.19% |
| 5 down days, below the 200-day | 34 | 41.2% | −0.64% | 52.9% | +1.10% |
| 4 down days, above the 200-day | 73 | 56.2% | +0.75% | 63.0% | +1.30% |
| 4 down days, below the 200-day | 58 | 51.7% | +0.02% | 56.9% | +1.01% |
(The first 199 days of data have no 200-day average, so a few early streaks are not in this split.)
A 5-day pullback inside an uptrend was followed by a higher close a week later in 21 of 30 cases. The same streak in a downtrend was followed by a higher close a week later in only 14 of 34. That fits a simple idea: in an uptrend a streak is often a pause, in a downtrend it is often the trend itself. With 30 and 34 cases, though, a few results either way would move these numbers a lot, so use this to set expectations, not as a trading rule. Our lesson on mean reversion on gold covers when fading a move makes sense and when it does not.
We also checked the size of the fall. Of the 71 five-day streaks, 14 had lost 5% or more over the five days. After those, gold was up five sessions later in only 7 of the 14 cases, and the median 20-day change was −0.13%. Bigger falls were not followed by bigger bounces in this data.
Two streaks from 2026, side by side
Gold had two 5-day losing streaks in 2026 before this month. They looked alike and ended very differently.
| March 2026 | June 2026 | |
|---|---|---|
| Close before the streak | $5,008.2 (17 March) | $4,505.0 (4 June) |
| Close on the fifth down day | $4,402.0 (24 March) | $4,114.0 (11 June) |
| Fall over the streak | −12.1% | −8.7% |
| Next day | $4,552.3, +3.4% | $4,238.8, +3.0% |
| 5 sessions after the fifth down day | $4,647.6 (31 March) | $4,245.9 (18 June) |
| 20 sessions after | $4,753.0 | $4,005.7, below the streak low |
Both streaks ended the next day with a 3% jump, the kind of move that makes “buy after five red days” feel clever. A month later, March’s buyer was well ahead, while June’s buyer was sitting below the price at which the streak ended. Nothing in the count of red candles told you which one you were in.
What this means for your trading
Do not size up because a streak feels “stretched”
Long streaks often come with wide daily ranges. On 19 November 2026 the 20-day average daily range of COMEX gold was about $93.9, and Thursday’s own range was $138.9 (from $3,759.5 down to $3,620.6) on our live data. Here is what that does to position size on a hypothetical $5,000 account risking 1% ($50) per trade, where 0.01 lot moves about $1 for every $1 in gold:
| Situation | Stop distance | Loss at 0.01 lot | Loss at 0.02 lot | Fits 1% risk? |
|---|---|---|---|---|
| Quiet week, tight stop | $20 | $20 | $40 | Yes, up to 0.02 lot |
| Volatile week, stop at about half the day’s range | $45 | $45 | $90 | Only 0.01 lot |
| Volatile week, stop beyond the day’s range | $95 | $95 | $190 | No, even 0.01 lot is 1.9% |
The account sizes and stops are examples; the ranges are real. On a $1,000 account, 1% is $10, so even the quiet-week $20 stop at 0.01 lot would be 2% risk. The point: a fast fall shrinks the position you can safely hold, which is the opposite of what the “it must bounce” feeling suggests. Our position size calculator does this sum for any account, and the volatility tool shows how wide recent days have been.
Averaging down on a streak is the classic account killer. Adding to a losing buy after each red day doubles your exposure exactly when the move is going against you. Our guide to averaging down shows the maths.
Use the streak as a question, not an answer
When you notice four or five red candles, ask three things instead of counting:
- What is driving it? This week it was a near-certain December Fed hike, a dollar near a 19-month high and forced selling during a tech sell-off (see the daily news). If the driver is still there tomorrow, the streak has no reason to stop.
- Where is gold against its bigger trend? Above a rising 200-day average, a streak was often a pullback in our data. Below it, the streak was more often the trend.
- What would prove you wrong, and how far away is that? If the honest answer is “$100 away”, your size has to be small, or the trade is not worth taking.
Let the market show the turn
Many traders wait for a sign that sellers are tiring before they act: a close back above the previous day’s high, a failed new low, or a clear change in the hourly structure. None of these is guaranteed either, but they are evidence from price, not from a count. Our chart of the day walks through these patterns on real recent bars, and the gold cheat sheet lists this week’s levels and events.
How this fits with our other studies
This post looks at days. If you want the bigger picture:
- Should you buy the dip in gold? tests buying 5% to 25% falls from the record high.
- What happens to gold after a big monthly drop? does the same for months that fell 6% or more.
- Gold after big up and down days looks at single days that moved 2% or more.
- Gold flash crashes covers falls that happen in minutes rather than days.
All three reach a similar conclusion to this one: past falls give context on what can happen, but none of them gives you a timing signal by itself.
Key takeaways
- In 26 years of COMEX gold closes, a streak of 4 to 6 down days did not make the next day, week or month more likely to be up than an ordinary day.
- After four down days, a fifth came about half the time; after six, a seventh came 58% of the time. No streak is “due” to end.
- A lower close came within two weeks in about three cases out of four, the same as normal, and the worst tenth of 5-day streaks saw another fall of nearly 5%.
- The 200-day trend was the clearest difference: 5-day streaks above it were followed by a higher close a week later in 21 of 30 cases, below it in 14 of 34. Small samples, so context only.
- Fast falls come with wide ranges, so cut your size, never raise it, when gold has been falling for days.
Next step: open the trading psychology hub and run the pre-trade checklist before your next trade on a falling market, or work through the School of Gold stage on risk if position sizing still feels unclear.
Sources
- XAUUSD Global calculation from COMEX gold futures daily closes (GC=F via Yahoo Finance, stored in the live
gold-daily.json), 30 August 2000 to 19 November 2026; 6,582 closes. Streaks, forward changes, further falls and the 200-day split computed for this article on 20 November 2026 - XAUUSD Global spot gold feed (
spot-close.jsonandlive.json), closes for 13-19 November 2026 and the 19 November range and 20-day average range; Friday price $3,520.1 at 05:37 GMT on 20 November 2026 - Drivers of the November 2026 fall as reported in our 20 November digest (Reuters via MarketScreener, Investing.com)
- The account sizes, risk levels and stop distances in the sizing table are hypothetical examples
Educational content only, not financial advice. Past results do not predict future returns. Trading gold with leverage carries a high risk of loss.
Frequently asked questions
Does gold bounce after 5 down days in a row?
Not reliably. In COMEX gold closes from 30 August 2000 to 19 November 2026 there were 71 times gold closed lower five days in a row. The next day was up 56.3% of the time, close to the 52.5% for any day, and the median next-day change was +0.11%. Five sessions later gold was up only 53.5% of the time, slightly below the normal 55.5%.
How often does a gold losing streak keep going?
More often than most traders expect. After four down closes in a row, gold closed lower a fifth time in 71 of 140 cases (50.7%). After six, it fell a seventh time in 18 of 31 cases (58.1%). The longest streak in our data was 11 sessions, in September 2000, and there was a 10-session run in July 2015.
Is it a good idea to buy gold after several red days?
A streak alone is not a reason. In our data the result depended on the bigger trend. After five down days while gold was above its 200-day average, the next five sessions were up 70.0% of the time (30 cases). Below the 200-day average, they were up only 41.2% of the time (34 cases). Both samples are small, so treat this as a warning about context, not as a rule to trade.
Why does gold keep falling day after day?
Long streaks usually have one steady driver behind them, such as rising interest-rate expectations, a strong dollar or forced selling during a market sell-off. When that driver does not change, sellers keep coming back each day. A streak ends when the driver changes or when selling is exhausted, and a count of red candles cannot tell you which day that is.
Is a streak on spot gold the same as on futures?
Not always. Spot XAUUSD and COMEX futures close at different times and can differ by a few dollars, so one can show a down day while the other shows a small up day. In the week to 19 November 2026, our spot feed showed four lower closes in a row while COMEX futures showed only two, because futures closed slightly higher on 17 November.
Comments
No comments yet.
More in Strategies & technical analysis
- How to Identify Support and Resistance on a Gold Chart
- Order Flow on Gold Futures: Footprint Charts, Delta and Heatmaps Explained for XAUUSD Traders
- ICT Concepts for Gold Trading: Liquidity, FVGs and Market Structure
- London Close Reversal on Gold: Does XAUUSD Really Turn at 4 pm London? We Tested 601 Days
- How to Build a Trading Plan for XAUUSD
Latest on the blog
- The Philly Fed Survey for Gold Traders: Diffusion Index, Prices Paid and the Thursday Reaction
- What Is the Fed Beige Book? A Gold Trader's Guide to District Reports, Price Clues and Market Reaction
- Trading Psychology Quotes That Actually Help: 12 Lines From Famous Traders and What They Mean for Gold
- Gold Trading Tax in the UK: Spread Betting vs CFDs vs Futures, Explained From HMRC's Own Guidance