Does buying the dip work? What Bitcoin did after every −7% and −10% day since 2013

The question behind "buy the dip"

Every sharp red day in Bitcoin brings the same posts: "buy the dip", "this is the discount", "we bounce from here". The idea is that a big one-day fall is an overreaction, and the price tends to come back. It is easy to remember the times that worked, like March 13, 2020, the day after Bitcoin's biggest crash in years. It is harder to remember the times when the first big drop was only the start.

So we counted them. We took every daily close of Bitcoin from January 2013 to November 5, 2026 (Bitstamp, about 5,000 days) and found every day it fell 5%, 7%, 10% or 15% or more. Then we looked at what happened next: the next day, the next week, the next month and the next three months, and how much further the price fell before any bounce. We repeated the test for Ethereum (from 2017) and Solana (from 2020) on Binance.

First drop vs any drop

There is a trap in this kind of study. Big drops come in clusters: in a crash, Bitcoin can lose 7% three days in a row. If you count every one of those days, the last one, right before the bounce, makes "buying the dip" look great. But you only know it was the last one afterwards.

So we report two versions. Every drop counts each day separately. First drop counts only a drop with no drop of the same size in the 7 days before it: the day a trader first sees a big red candle and has to decide. The first-drop version is the honest one for anyone thinking about acting on it.

What happened after a big red day

For comparison, on an ordinary day since 2013, Bitcoin was higher a week later 54% of the time, with a median gain of 0.65%.

Counting every drop, dips look good. After all 145 days with a fall of 7% or more, Bitcoin was higher a week later 58% of the time, with a median gain of 2.2%. After a 10% day, 59% and 4.0%. After a 15% day, 67% and a median of +21% a week later.

The first-drop numbers are much weaker. After the first −7% day in a cluster, Bitcoin was higher a week later only 48% of the time, with a median of −0.2%. That is slightly worse than an ordinary day. After the first −10% day it was 53%, about the same as an ordinary day. In other words, most of the "bounce" in the every-drop numbers comes from the later days of a crash, which you cannot pick out in advance.

The second drop usually comes

The most practical number is how much further the price fell first. In the 7 days after the first −7% day, Bitcoin traded at least another 5% lower in 70% of cases, and at least another 10% lower in 42%. On an ordinary day, the chance of a further 5% fall within a week is 45%.

And getting back to the price before the drop took time. After the first −7% day, Bitcoin regained the previous day's close within a week only 27% of the time, within 30 days 60% of the time and within 90 days 78%. After the first −10% day the numbers were almost the same (26%, 58% and 77%). So roughly one in four big drops had not recovered even three months later.

For anyone using leverage, this matters more than the bounce rate. A 3x long opened on the first −7% day would have faced a further 10% drop, a 30% loss on the position, in four cases out of ten.

Above or below the 200-day average

Context made a clear difference. We split first drops by whether Bitcoin was above or below its 200-day moving average, a simple way to tell an uptrend from a downtrend.

• In an uptrend (above the 200-day average), a first −7% day was followed by a higher price a week later 54% of the time and a month later 58%, with a median 30-day gain of 3.7%. That is close to normal uptrend behavior: the trend was still the main thing.
• In a downtrend (below the 200-day average), the same drop was followed by a higher price a week later only 39% of the time, with a median 7-day change of −2.1%, and the 90-day median was below zero.

The same split holds for −5% days: above the 200-day average, a 30-day median of +2.8%; below it, −3.4%. A big drop in a falling market was more often a step down than a discount.

Each cycle, dips bounced less

The edge, small as it was, has faded. For first −7% days in 2013–2016 the median 7-day change was +2.2%. In 2017–2020 it was −0.1%. In 2021–2026 it was −1.5%, with only 38% of cases higher a week later.

Big drops have also become rare. Bitcoin had 25 days of −7% or worse in 2013, 22 in 2017 and 14 in 2021, but only 1 in 2023, 3 in 2024, 2 in 2025 and 1 so far in 2026 (February 5, −14%). Its average daily move fell from 4.6% in 2013 to about 1.6% in 2025–2026. A −7% day in 2026 is a much bigger event, relative to normal, than it was in 2013. Since 2023 a −5% first drop has been followed by a higher price a week later in exactly half of the 24 cases.

Ethereum and Solana

For Ethereum, dips were worse than for Bitcoin. After a first −7% day, ETH was higher a week later 42% of the time and a month later 42%, with a 90-day median of −7.6%. In 2021–2026 the median next-day move after a first −7% day was −1.7%: the second day tended to be red too. Buying ETH's first big red day was below a coin flip at every horizon we tested except the next day.

Solana was in between: after a first −10% day it was higher a week later 61% of the time, but a month later only 45%, with a median of −4.0%. Solana's very large swings (a single month can move it 50%) mean any short-term pattern is noisy.

How to read this

1. A big red day is not a discount by itself. For Bitcoin, the first −7% day was followed by a week that was no better than average, and for Ethereum it was worse.
2. Expect a second leg. Seven times out of ten, Bitcoin fell at least another 5% in the following week. Plan any position size around that, not around the bounce.
3. Check the trend. Drops in an uptrend behaved like normal uptrend days. Drops below the 200-day average were more often followed by lower prices.
4. The huge crash days were different. After the rare −15% days, Bitcoin's 30-day median was +17%, but there were only 16 such first drops in 14 years, most of them before 2021.

You can follow Bitcoin's current trend, 200-day average and levels on our Bitcoin price analysis page, and compare with Ethereum and Solana. Our Bitcoin volatility study shows how much calmer the daily moves have become. Kumocast provides free market data and analysis, not trading signals or financial advice.