Buying at the All-Time High: Psychology and Historical Returns

Buying bitcoin at an all-time high has so far almost always worked out, but not always quickly. Since 2011, bitcoin closed at a new record on 236 days. On 61 % of them the very next day closed even higher. After four records, by contrast, it took two to three years to get back to the purchase price, with interim losses of 77 to 93 %. After four years every purchase at an all-time high has so far been in profit; after one year, almost four in ten were at a loss.

When to get in at all, recalculated for every single day since 2011, is covered in Finding the right entry point. This article is about one particular day: the one when the headlines read "Bitcoin hits record high" and you wonder whether you can still buy.

Why an all-time high feels so expensive

A record price triggers two feelings at once, and they pull in opposite directions.

Anchoring: you compare today's price with the one in your head. If you saw bitcoin at $60’000 a year ago, $100’000 feels expensive, whatever bitcoin is actually worth. The high itself becomes the anchor: anything above it feels like excess.

FOMO: at the same time the fear of missing out grows. Friends talk about their gains, the media report daily. Many buy for the first time exactly then, and in large amounts. How this cycle of greed and panic plays out is described in FOMO, panic and capitulation.

The result is often the worst of both worlds: first you wait because it is too expensive, then you buy everything at once because it got even more expensive. The data show that both reflexes rest on a false assumption: that a record says something about whether the price falls afterwards.

How often did bitcoin close at a record?

Every day is counted whose closing price in US dollars was higher than on any day before, from 1 January 2011 to 7 November 2026.

In all other years there was not a single one. Records come in clusters, in the rising phases of the four-year cycle, with years without any in between.

Within such a phase a record is no warning sign. On 144 of the 236 record days, or 61 %, the next day already closed higher. Only 18 records stood for longer than a month, only five for longer than a year. Those five decide almost everything.

Returns after buying at an all-time high

How did purchases on record days fare over the years, compared with purchases on any day?

Buying at a record was clearly riskier than buying on any day during the first two years: it was more often at a loss, and the median was only about half as high. After four years the gap had nearly closed.

Counting from 2015, without the early years and their extreme jumps, the difference at the start is even larger: after one year 44.5 % of record purchases were at a loss, median +11 %; after two years also 44 % at a loss, median +8 %. After three years only four record days in February 2021 were still slightly down (−1 to −10 %). After four years none, the worst at +32 %.

The five expensive records: cycle tops

Five records were not beaten for more than a year. They are the peaks of the cycles, and the whole difference between "record day" and "any day" hangs on them.

Anyone who bought on one of these days saw more than half disappear in between and waited two to three years to break even. After four years, every one of these purchases was still in profit. But the return shrank from cycle to cycle: +919 %, +143 %, +52 %. How deep and long each crash was is covered in detail in All bitcoin bear markets compared.

Important for the psychology: on none of these days could you tell it would be the last record. 2017 had 74 records. Each was quickly beaten, except the last one on 16 December.

The first record after a bear market

There is also the other kind of record: the first new high after years of doldrums. The papers carry the same "Bitcoin hits record high", and many think they have missed the boat.

These records have so far been good entry points, especially over one year. Anyone who thought in February 2017 that $1’189 was too much after the old 2013 high missed eightfold gains by February 2018. These gains too shrank from cycle to cycle, though, and the record of March 2024 was only 6 % higher two years later.

Does valuation reveal which record it is?

If cycle tops can't be recognised by the date, maybe by valuation? The site measures it against the Power Law floor, a slowly rising line below the price. The higher the price above the floor, the more expensive relative to the long-term trend.

Up to 2021 valuation was a good hint: the cycle tops sat far above the floor, the first records close to it. For record days since 2015 the pattern held across all of them too. The 17 records below twice the floor, all from February to May 2017, were never at a loss after one year. Of the 79 records above three times the floor, 40 were at a loss after one year, about half.

Then came October 2025. The latest cycle top was only 2.7 times the floor, almost exactly where the first record of March 2024 had been. This time valuation gave no "this is the end" signal. The peaks are getting flatter, as the model's top band has shown for years, and that blurs the line between "fresh record" and "peak".

Worked example: all at once or spread out from the record

What helps if you want to invest a larger sum on a record day? The calculation compares two routes with $12’000 each: buy everything on the record day, or buy $1’000 a month for twelve months from the record day. Everything is valued at the price two or four years later.

What to do on the next record day

The figures give no price signals, but they suggest a few rules that take the decision off your hands.

A record is no reason not to buy. Most records were beaten quickly, and after four years every record-day purchase has so far come out ahead. Anyone waiting for years for a dip below the old high missed a lot in 2017, 2020 and 2024.

A record is no reason to buy more either. After one and two years, record-day purchases did worse than purchases on any day. If the headline makes you double your savings plan, that is FOMO, not strategy.

Spread large sums. If you only get in at the record and want to invest a lot, spread it over months. That costs return in the good case and protects you from a 70 % loss in year one in the bad case.

Bring four years. Money you need in the next two years should, on these figures, not go into bitcoin at a record price. Setting your goal and time horizon is what My Goal helps with.

Where bitcoin stands relative to the trend today: