Black Friday: a whole free December, until 27 November. See the offer →
SonarSignal
Analysis6 min read

Stop-Loss on Bitcoin: How Far Price Pierces Below Obvious Lows

We measured 4,330 Bitcoin days: a stop 1 percent below the open triggers on 63 of 100 days. And below obvious lows, price often pierces deeper still.

Share:XTelegramWhatsAppRedditFacebookLinkedIn

A stop 1 percent below the open would have been triggered on 63 out of 100 Bitcoin days. Normal daily noise alone is that large. We measured how tight a stop can realistically sit and how far price pierces below an obvious low before it turns. The result is uncomfortable for both sides: a tight stop is often triggered without anything having changed in the market. A stop just under the next low sits at a spot that price often visits.

What a stop is for

A stop-loss is an order that closes a position once price reaches a set level. It does not predict where price goes. It only limits how much a trade can cost at most, plus fees and slippage: a stop is usually executed at the market price and can fill worse than the level in fast moves. The real decision is therefore the distance. The farther away the stop, the less often it triggers, and the smaller the position has to be if the possible loss is to stay the same.

First measurement: daily noise

For every Bitcoin day since the start of 2015 we checked whether price, measured from the open, would have dropped by at least x percent. The basis is 4,330 daily candles of BTC/USD from Bitstamp up to 8 November 2026. The range between the daily high and low has a median of 3.8 percent of the open, and 2.9 percent over the last 24 months. A stop would have triggered on this many days out of 100:

– 1 percent distance: 63 days (last 24 months: 56)

– 2 percent: 40 days (32)

– 3 percent: 26 days (18)

– 5 percent: 12 days (5)

– 8 percent: 5 days (less than 1)

– 10 percent: 3 days (less than 1)

On about every third day on which a 1-percent stop would have triggered, price still closed above the open. A stop inside the normal daily range therefore measures noise, not a trend change. This applies to daily candles. On hourly candles the swings are smaller, but the logic is the same: the stop has to fit the volatility of the timeframe you trade on.

Bar chart: share of Bitcoin days on which a stop at 1 to 10 percent distance would have been triggered

Second measurement: the pierce below the obvious low

Traders talk about a liquidity grab or a stop hunt. Stop orders pile up below a clearly visible low. Price dips just below, triggers them and turns back. Whether anyone does this on purpose cannot be read from price data. What can be measured is the pierce itself.

We count the lowest point of eleven days, five days before and five after, as an obvious low. Since 2015 there were 270 of them. For each we checked whether price undercuts it within the following 60 days and what happens next:

– 164 of 270 lows (61 percent) were undercut.

– In 120 of those 164 cases (73 percent) price closed back above the low within three days, in 82 cases on the same day. That is the classic wick.

– These wicks pierced a median of 1.7 percent below the low. In a quarter of cases it was more than 4.1 percent, in one in ten more than 7.1 percent.

– The other 44 cases (27 percent) did not come back. There, the lowest point of the following ten days lay a median of 13 percent below the low, and in one in ten cases more than 39 percent.

A stop behind the low would often not have survived the wick: at 1 percent distance, 64 of 100 wicks would have triggered it, at 3 percent 28, at 5 percent 22 and at 8 percent 7.

For highs, meaning the stops of short positions, the picture is similar but less clear: 190 of 255 highs (75 percent) were exceeded, 54 percent of those closed back below within three days, and the wick had a median height of 1.6 percent. The difference from the lows is probably explained by Bitcoin's uptrend since 2015. We cannot prove that.

Two panels: share of wicks after the pierce and share of wicks that a stop 1 to 8 percent behind the low would have triggered

And what comes after the triggered stop?

Anyone stopped out below a low often considers it a mistake afterwards. The data are less clear-cut: across all 164 pierces, price stood back above the low 14 days later in 54 percent of cases, and not in 46 percent. That is roughly a coin toss.

Two cases show both sides. On 5 August 2024 Bitcoin pierced 7.4 percent below the low of 5 July (53,550 US dollars) and closed back above it the same day, at 54,014. In June 2022 price ran below the low of 26 May (28,003 US dollars) and did not come back: in the ten days after the break it fell as far as 37 percent below that low. Both cases sit in the same data, and at the moment of the pierce you cannot see which one it will be.

Two Bitcoin daily chart sections: wick on 5 August 2024 and further fall in June 2022 below an obvious low

What the measurement shows

– A stop tighter than the daily range (for Bitcoin roughly 3 to 4 percent) is triggered by noise.

– A stop right behind an obvious low sits where price often pierces: 61 percent of these lows were undercut within 60 days, and a good one in four wicks went beyond 3 percent.

– A wide stop does not protect against every loss. If price falls through, the further drop has a median of 13 percent.

– The price of more room is position size. If the possible loss stays the same, the position gets smaller as the stop gets wider.

Limits of our measurement

The figures come from daily candles from Bitstamp that start at 0:00 UTC. Anyone trading on hourly candles sees smaller swings in absolute percent. That a low is the lowest point of eleven days is our own definition. With 21 days, 72 percent of pierces closed back within three days, with 41 days 67 percent, on only 27 cases. The period 2015 to 2026 contains a long uptrend. Fees, spreads and slippage are not included, and this is not a test of a trading strategy: the measurement says nothing about whether a stop makes money. Past prices are not a forecast.

The stop in the Trading Desk

In the Trading Desk you set the stop as a distance in the launch pad of a box, and take-profit is set automatically by Fibonacci by default. The risk modules calculate with it differently. In Safety Focus (Mod 1) the stop sits strictly outside the signal zone. In Mathematical Precision (Mod 2) the quantity is calculated from the distance between entry and stop, whether the stop is 1 or 10 percent away. This is exactly where our measurement applies: the distance decides how often the stop triggers and how large the position is. How the modules work in detail is covered in our articles on risk strategy and on the box.

Conclusion

A stop is not a forecast but a limit with a deductible. Our measurement shows that tight stops pay for daily noise and that the spot right behind an obvious low is visited often. Whether a stop holds depends on the distance and the position size, not on hoping that price turns exactly there.

Not investment advice. This post is for information and education. It is not investment advice and not a recommendation to buy or sell. Trading cryptocurrencies and leveraged products can lead to the total loss of the capital employed.
Share:XTelegramWhatsAppRedditFacebookLinkedIn

More posts