What happens after Bitcoin hits an all-time high? 13 cases
After 13 new Bitcoin highs on Coinbase, price was up 30 days later in 53.85% of cases, with a middle move of +0.40%. Every case is listed with its date.
On Coinbase BTC-USD, Bitcoin was higher 30 days after a new high in 7 of 13 cases (53.85%), and the middle result was +0.40%. Across all days from July 2016, the same figures were 57.09% and +2.83%. So a new high did not, in these cases, put Bitcoin ahead of an ordinary day. In the 90 days after April 13, 2021, the lowest close was −50.31% below the high. In the 90 days after October 12, 2017, Bitcoin ended +173.44% higher.
These are past cases, not a forecast.
What counts as a new Bitcoin high here?
We use the highest daily close so far in our data, which is Coinbase spot Bitcoin priced in US dollars (BTC-USD). A close is the last price of a day, in UTC. We use closes, not the highest price touched during the day.
A rally makes many new highs in a row. The data holds 194 closes that beat every close before them. Counting each would give one rally dozens of "cases". So a case is the first new-high close after 30 or more days without one. We also need 365 days of earlier data, so a case could start from July 2016.
Two early highs passed those rules but are left out: December 9, 2016 (a close of $773.22) and February 23, 2017 ($1,186.90). Bitcoin had traded above $1,000 in late 2013, before our data starts. So the first was not a true all-time high, and the second was too close to the 2013 peak to call. That leaves 13 cases, the first in April 2017.
What did Bitcoin do after each new high?
This is every case. The day columns show the change from the high close to the close that many days later. The last column is the lowest close in the next 90 days compared with the high close. Zero means no close in that time was lower.
| Date of the high | Close | 7 days | 30 days | 90 days | Lowest close within 90 days |
|---|---|---|---|---|---|
| 2017-04-26 | $1,298.44 | +18.07% | +75.03% | +97.47% | 0.00% |
| 2017-08-05 | $3,243.49 | +19.44% | +38.69% | +121.06% | −0.66% |
| 2017-10-12 | $5,440.00 | +4.85% | +16.67% | +173.44% | 0.00% |
| 2020-11-30 | $19,713.94 | −2.72% | +46.58% | +129.44% | −8.48% |
| 2021-02-08 | $46,448.10 | +3.20% | +20.33% | +25.54% | −3.44% |
| 2021-04-13 | $63,588.22 | −11.15% | −21.86% | −47.96% | −50.31% |
| 2021-10-19 | $64,303.14 | −6.17% | −11.52% | −34.34% | −35.36% |
| 2024-03-04 | $68,360.14 | +5.49% | −3.47% | −0.91% | −14.77% |
| 2024-11-06 | $75,646.55 | +19.63% | +32.05% | +29.28% | 0.00% |
| 2025-01-21 | $106,159.26 | −4.59% | −7.36% | −17.56% | −28.17% |
| 2025-05-18 | $106,473.46 | +2.42% | −1.77% | +10.32% | −5.14% |
| 2025-07-10 | $116,031.30 | +2.79% | +0.40% | +6.30% | −6.71% |
| 2025-10-05 | $123,520.79 | −6.84% | −17.85% | −26.65% | −31.44% |
The cases are uneven. In the first five, every 90-day result was positive. Of the last six, three ended 90 days later below the starting close.
How does that compare with an ordinary day?
Here is each measure across every day from July 19, 2016, with the same forward windows. The middle value is the median: sort all the results and take the one in the center.
| Measure | After a new high (13 cases) | All days (about 3,700) |
|---|---|---|
| Up after 7 days | 61.54% | 54.01% |
| Middle move after 7 days | +2.79% | +0.60% |
| Up after 30 days | 53.85% | 57.09% |
| Middle move after 30 days | +0.40% | +2.83% |
| Up after 90 days | 61.54% | 58.99% |
| Middle move after 90 days | +10.32% | +9.28% |
| Middle lowest close within 90 days | −6.71% | −12.94% |
| Fell 20% or more within 90 days | 30.77% | 35.10% |
With 13 cases, none of these gaps is large enough to read as a pattern. The 7-day figures look better after highs and the 30-day figures look worse. We cannot tell either from chance. The shallower lowest close after highs (−6.71% against −12.94%) is the one measure where highs looked calmer, but the two worst cases still dropped −50.31% and −35.36%.
The 13 cases are also not 13 independent tests. Several sit in the same rally. If we keep only cases at least 90 days apart, 9 remain. For those 9, the 30-day middle move was −1.77%, and 4 of the 9 (44.44%) were up.
Did the early cases look like the recent ones?
No, and a single blended number hides that.
| Cases | Count | Up after 30 days | Middle move after 90 days | Middle lowest close within 90 days |
|---|---|---|---|---|
| 2017 to 2021 | 7 | 71.43% | +97.47% | −3.44% |
| 2024 to 2025 | 6 | 33.33% | +2.69% | −10.74% |
Five of the seven cases through 2021 were followed by large 90-day gains. Most of the six since 2024 ended with small gains or losses. We cannot say why, and each group is tiny.
What about Ethereum?
Coinbase ETH-USD has history from May 2016. The same rules give only 5 cases: stories, not rates.
| Date of the high | Close | 7 days | 30 days | 90 days | Lowest close within 90 days |
|---|---|---|---|---|---|
| 2017-11-23 | $406.61 | +6.74% | +80.22% | +106.67% | 0.00% |
| 2021-01-24 | $1,394.00 | −5.74% | +13.19% | +59.02% | −10.98% |
| 2021-04-01 | $1,968.38 | +5.74% | +49.81% | +15.67% | −8.09% |
| 2021-10-25 | $4,220.93 | +2.42% | +1.18% | −39.80% | −42.86% |
| 2025-08-22 | $4,831.24 | −9.75% | −7.95% | −41.41% | −41.41% |
Four of the 5 were up after 30 days. The last two cases (October 2021 and August 2025) were each about 40% lower after 90 days. Five cases cannot tell you which of those results is typical.
Where does the data stand now?
The latest new-high close for Bitcoin in our data was $124,720.09 on October 6, 2025, one day after the last case in the table. The data ends on October 19, 2026, 378 days later. The last close was 34.38% below that high.
For Ethereum, the latest high close was $4,831.24 on August 22, 2025, and the last close was 50.22% below it. This is a description of where two charts stand, not a prediction of what comes next. For more on how long recoveries took after past drops, see crypto drawdown recovery times.
What does this data not show?
It does not show why any case went the way it did. We did not look at news.
It does not show a tradable rule. These numbers ignore fees and slippage.
It covers Bitcoin on one exchange, plus a few Ethereum cases. Highs on Binance perps, which began in 2019, are only "highest close since the contract began," so we do not mix them in.
And the sample is small. Bitcoin had three streaks of highs: 2017, 2020 to 2021, and 2024 to 2025. Cases within one streak share most of their path. For how each streak lined up with the halvings, see Bitcoin halving price history. See also why past returns are not a forecast.
How we measured this
We used Coinbase daily spot closes for BTC-USD from July 20, 2015 and for ETH-USD from May 2016, through October 19, 2026. Each forward move starts at the case's close and ends at the close 7, 30 or 90 days later. The lowest close uses closes only, not the day's low. The "all days" figures use every day with 365 days of earlier data and a full forward window. Windows that ran past the last bar were dropped. Nothing here includes fees.
Try this in Comparand
Ask your connected AI agent: "Get completed daily bars for Coinbase BTC-USD over my chosen dates. List each close that beat every earlier close, and mark the first one after 30 quiet days."
- venue_bars returns completed open, high, low and close bars for one symbol. Set the interval to 1d. It is free on every key.
- Your agent finds the highs from those bars. The tool does not return a list of highs.
Then ask: "Find the 10 closest past charts to Bitcoin's last 48 daily bars on Coinbase."
- shape_analogs with timeframe 1d searches only Bitcoin's own past on Coinbase, with daily bars from January 2017.
- Each match comes back with its own path and its forward move at 12, 24 and 48 days, in basis points (one hundredth of one percent).
- It returns no average and no hit rate. Each search costs $0.01 from your credits.
You can try 3 coins free, check pricing, or read the tool docs.
The short version
- After 13 new Bitcoin highs on Coinbase, 53.85% were up 30 days later, with a middle move of +0.40%. All days: 57.09% and +2.83%.
- Within 90 days, 4 of the 13 (30.77%) closed 20% or more below the high. The worst was −50.31%.
- The 2017 to 2021 cases and the 2024 to 2025 cases looked very different.
- Ethereum has only 5 cases, too few to read.
- These are past cases from a small sample, not a forecast.
Questions people ask
What happens to Bitcoin after it hits a new all-time high?
In 13 past cases on Coinbase BTC-USD (April 2017 to October 2025), Bitcoin was higher 30 days later in 53.85% of them, with a middle move of +0.40%. Across all days from July 2016 the same figures were 57.09% and +2.83%. The cases ranged from −21.86% to +75.03% after 30 days.
Does Bitcoin usually crash after an all-time high?
Not usually, but it did fall hard in some cases. Within 90 days, 4 of the 13 cases (30.77%) closed at least 20% below the high, and the worst was −50.31% after April 13, 2021. The middle case fell only −6.71% at its lowest close in those 90 days.
How many times did Bitcoin make a new all-time high?
Bitcoin made many new-high closes in a row during each rally, so we count the first one after 30 or more days without a high. That gives 13 separate cases on Coinbase BTC-USD from April 2017 to October 2025, with the latest on October 5, 2025.
Did Ethereum do the same after new highs?
There are only 5 Ethereum cases on Coinbase ETH-USD (2017 to 2025), too few to call a pattern. Four of the 5 were up 30 days later, but the last two cases (October 2021 and August 2025) were each about 40% lower after 90 days.
Can these numbers tell me what happens after the next high?
No. They are 13 past cases, not a forecast. Cases from 2017 to 2021 and from 2024 to 2025 looked very different, so one blended number would hide that.