Fibonacci Retracement in Crypto: How the Levels Work and Fail

Fibonacci retracement is one of the most popular drawing tools in fibonacci retracement crypto charting: you stretch a grid across a recent move and watch whether price pauses at 38.2%, 50% or 61.8% of the way back. Used carefully, it organises your thinking about a pullback. Used carelessly, it lets you see support wherever you want to. This guide shows where the numbers come from, how to draw the tool, how to turn the levels into a plan with a stop and a target, and where it fails. It builds on the support and resistance guide, the pullback trading guide and the combining indicators guide.

One thing up front: CoinSeekly does not draw Fibonacci levels. You will draw them yourself on a charting tool. Later in the guide we cover how the screener can still help you find the swings worth measuring.

Where the Fibonacci ratios come from

The Fibonacci sequence starts with 1 and 1, and each number after that is the sum of the two before it:

1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233...

Divide a number by the next one in the sequence and the result settles around a fixed value as the numbers grow. 55 divided by 89 is 0.618, and 89 divided by 144 is also about 0.618. This is the golden ratio (its reciprocal, 1.618, is the number that appears in extensions).

The common retracement levels are built from ratios between numbers in the sequence:

Level	How it is derived	Check
61.8%	A number divided by the next one	55 ÷ 89 = 0.618
38.2%	A number divided by the one two places later	34 ÷ 89 = 0.382
23.6%	A number divided by the one three places later	21 ÷ 89 = 0.236
78.6%	The square root of 0.618	√0.618 ≈ 0.786
50%	Not a Fibonacci ratio at all	Simply the halfway point

Two honest notes. First, 50% is not a Fibonacci number. It is on almost every chart because traders have long watched the halfway point of a move. Second, none of this proves that markets respect these ratios. The maths of the sequence is real, but the link to price is a trading convention, not a law, which is why the later sections focus on limits.

What a retracement level is supposed to show

Prices rarely move in a straight line. After a strong rise, price usually gives back part of the gain before the next leg, and traders ask how deep that pullback might go. The retracement tool divides the prior move into percentages so you have a handful of candidate areas where the pullback might stall.

The levels are measured as a share of the prior swing. If a coin rose by $100 and then falls back by $38.20, it has retraced 38.2% of the rise. A shallow pullback (23.6% to 38.2%) is often read as a sign of strength, because buyers stepped in quickly. A deeper one (61.8% to 78.6%) is read as a more serious test of the move. Beyond 100%, the entire prior move has been erased and the tool no longer applies.

You are not predicting the turn, you are marking areas where you would pay attention. The pullback trading guide covers the broader approach, and Fibonacci is one way of choosing where on the pullback to look.

How to draw it

In an uptrend

1. Pick a clear swing low that started the move and a swing high that ended it.
2. Select the Fibonacci retracement tool in your charting platform.
3. Click the swing low first, then drag to the swing high.
4. The platform draws horizontal lines at 23.6%, 38.2%, 50%, 61.8% and 78.6% between the two points, measured back down from the high.

In a downtrend

Reverse it. Click the swing high first and drag to the swing low. The levels now sit above the low and mark where a bounce might run out of steam before price resumes falling.

Wicks or closes

Decide whether you anchor to the wicks (the absolute high and low) or to the closes. Either choice can be defended. What matters is that you use the same one every time. Switching between the two depending on which gives a nicer-looking level is a way of fooling yourself.

Also pick a timeframe and stay with it. A swing on the daily chart is a bigger structure than one on the 15-minute chart, and the examples here assume daily candles.

A worked example with round numbers

The numbers below are hypothetical, chosen to make the arithmetic easy. They are not a real chart and not a prediction.

A coin rises from a swing low of $100 to a swing high of $200. The size of the move is $100. Each retracement level is the high minus that percentage of the move:

Retracement	Calculation	Price level
23.6%	$200 − (0.236 × $100)	$176.40
38.2%	$200 − (0.382 × $100)	$161.80
50%	$200 − (0.500 × $100)	$150.00
61.8%	$200 − (0.618 × $100)	$138.20
78.6%	$200 − (0.786 × $100)	$121.40

You can check one by hand: 0.618 × 100 = 61.80, and 200 − 61.80 = 138.20. Notice that the round $100 starting point makes the percentage and the dollar amount the same number, which is why it is a good one to practise on.

Now suppose price falls from $200 and stalls around $150, the 50% level. You do not know whether it will bounce. A hold at $150 and a rise tells you the pullback was moderate. A break below $150 puts $138.20 and then $121.40 in play. A close under $100 would mean the whole move has been erased and the grid is obsolete.

Extensions: using Fibonacci for targets

Retracements look backwards at where a pullback might end. Extensions look forwards to where the next leg might run. They are measured the same way but projected beyond the old high, as a multiple of the original move.

The usual extension levels are 127.2% and 161.8%. The 161.8% is the golden ratio itself. The 127.2% is the square root of 1.618. In practice they are simply "1.272 times the move" and "1.618 times the move".

Using the same swing from $100 to $200 (a move of $100), projected from the swing low:

Extension	Calculation	Price level
127.2%	$100 + (1.272 × $100)	$227.20
161.8%	$100 + (1.618 × $100)	$261.80

If the coin pulls back and then breaks above $200, those two prices are where some traders might take profit, because they represent a continuation of the same size relationship. Price may stop short or run past them. Treat extensions as a way to fix a target before you enter, not as a forecast.

Confluence: a level on its own is weak

A single Fibonacci line is rarely worth trading. The grid produces five levels across a move, so there is almost always one close to wherever price happens to bounce. What gives a level weight is confluence, meaning other independent reasons to care about the same area.

Look for lines that coincide with:

• Horizontal support or resistance. A 61.8% level that sits right on a prior swing low or a level that was resistance before is more interesting than one in empty space. See the support and resistance guide for how to find those levels.
• A moving average. If the 50-day or 200-day average passes through the same area, buyers who use that average and buyers who use the Fibonacci level are looking at the same price. The moving averages guide explains what those averages show.
• A round number. Prices like $150 attract orders simply because they are round. A 50% level at exactly $150 is a good example of a coincidence of that kind.

When three or four of those line up, you have an area, not a line. When only the Fibonacci grid points to a spot, treat it as the weakest kind of evidence. The combining indicators guide goes into how to stack tools without just piling up noise.

Here is BNB with its 50-day and 200-day moving averages. To use it for confluence, imagine drawing a retracement grid over a recent rise and check whether any level lands near one of the two averages. You are looking for lines that cluster in one area, not lines that merely exist:

Building a trade: entry, stop, target and risk:reward

A level is not a trade until you have an invalidation point and a target. Here is the hypothetical coin from the example again. It has run from $100 to $200, pulled back to the 50% level of $150, and that area also lines up with a rising moving average and a round number.

• Entry: $150 (the level, after you see price react there, for example a daily close back up).
• Stop: $130. This sits below the 61.8% level at $138.20 with a buffer, so a brief wick through the level does not take you out, but a real break does. If price is accepted below $130, the idea has failed.
• Distance to stop: $150 − $130 = $20 per unit.
• Target 1: $200, the prior high. Distance: $50.
• Risk:reward to target 1: $50 ÷ $20 = 2.5:1.
• Target 2: $227.20, the 127.2% extension. Distance: $77.20, so $77.20 ÷ $20 = 3.86:1.

Now size the position so the stop costs a fixed share of your account. With a $10,000 account risking 1% ($100) and $20 of risk per unit, the position is $100 ÷ $20 = 5 units, worth $750 at the $150 entry. If the stop hits, you lose $100. If target 1 hits, you make 5 × $50 = $250. You can plug your own entry, stop and account size into the position size calculator instead of doing this by hand.

The risk:reward is only attractive because the stop is tight relative to the target. If the nearest sensible stop were $100 away, the same trade would not be worth taking.

Where Fibonacci retracement fails

This section matters more than the rest, so read it before you put money behind a grid.

It is subjective. The tool depends entirely on which swing you choose. Two traders can look at the same chart, anchor to different highs and lows, and get different levels. Both feel confident. The grid cannot tell you which swing is the right one, and a different choice can move a level by a large amount. A consistent rule for choosing swings helps, but it does not remove the judgment.

Self-fulfilling at best. If a level works, one plausible reason is that many traders are watching the same line and place orders there. That is a real effect, but it is fragile: it depends on enough people drawing the same grid, and it can fail the first time a bigger force, such as news or a liquidation cascade, hits the market.

There are so many levels that something always holds. Five levels across a move, plus extensions, plus the lines you also draw on the next swing, make a very dense map. Price will touch or come near one of them almost every time it moves. Looking back, it is easy to say "it bounced at the 38.2%" and ignore the times it sliced through it. This is a known trap called confirmation bias: you remember the hits and forget the misses. The fix is to keep a record of every time you apply the tool, including the failures.

This guide offers no evidence that it works. We do not claim that price respects these levels more often than chance would suggest, and we cite no study that says so. Use it as a planning device for where to look and where to put a stop, not as a predictor.

A level is not a reversal signal. Price touching 61.8% does not mean it will turn. Wait for the market to show you something, such as a close back in the direction of the trend, rather than placing an order on the line and hoping.

Crypto adds its own noise. Wicks are long and fast, and a level can be pierced and recovered within hours. Daily closes and a stop buffer reduce, but do not eliminate, the shake-outs.

How to find the swings on CoinSeekly

Because CoinSeekly does not draw Fibonacci levels, the workflow is split: the screener helps you choose where to look, and your charting tool does the drawing.

1. Open the screener and look at coins in an uptrend that have pulled back from a recent high. The drawdown from the 30-day high and the trend filter help you find them.
2. Check the distance to nearest support and resistance. The screener finds these using mechanical swing pivots: a day whose low (or high) is the lowest (or highest) of the three days either side, over the last year of daily candles. Those pivots are often the same swing lows and highs you would anchor a Fibonacci grid to, so they give you an objective starting point.
3. Open the coin on a charting tool, draw the grid from the swing low to the swing high, and see whether any level lands near the nearest support the screener found. That is a confluence check you can do in a minute.
4. Browse the near support list and the deepest pullbacks list for coins that have already retreated toward an area worth measuring.

The screener's pivots are a simple mechanical rule, not zones and not Fibonacci levels. They are a starting point for choosing the swing, not a verdict.

A short checklist before you rely on a level

• Is the trend clear, and are you anchored to a clear swing using wicks or closes consistently?
• Does the level line up with support, a moving average, or a round number?
• Has price shown a reaction there, or are you guessing?
• Is the stop beyond the structure, and is the risk:reward what your plan requires?
• Is the position sized so the stop costs a small, fixed share of your account?

If you cannot say yes to most of these, the grid is decoration.

The bottom line

Fibonacci retracement divides a prior move into five familiar percentages, and the 50% is the odd one out because it is not a Fibonacci ratio. The tool is easy to draw and easy to over-trust. Its honest value is in planning: it gives you candidate areas for a pullback to end, a way to place a stop beyond a defined level, and extension targets such as 127.2% and 161.8% so you know the risk:reward before you enter. Its weaknesses are just as real: the swing is chosen by you, there are enough levels that something always seems to hold, and nothing here shows that it predicts price.

Treat any single line as weak and look for confluence with support, a moving average or a round number. Then size the trade with the position size calculator, and keep a record of how often the levels actually held. For more, read the support and resistance guide, the pullback trading guide and the technical analysis guide, or open the screener to find a pullback worth measuring.