A Fibonacci retracement can be drawn between almost any two points on a chart, and it always produces the same tidy grid. That is the problem. The tool looks objective while the only objective input is where you decided to click. A Fibonacci level is not a forecast. It is a place where a lot of traders happen to be looking at once, and it becomes tradeable only when it lands on structure the chart already had.

The evidence supports that skepticism. Roy Batchelor and Richard Ramyar tested whether the ratios of successive price trends in the Dow Jones Industrial Average from 1914 to 2002 cluster around Fibonacci ratios. A few significant ratios appeared, but no more than chance would produce across that many tests. The ratios are not a hidden law of markets. What makes them useful is more ordinary: everyone can see them.

What Is Fibonacci Retracement?

Fibonacci retracement measures a completed price move and marks the percentages of it where a pullback might stall. You anchor the tool to one swing and the platform draws the levels. The percentages come from the Fibonacci sequence, where each number is the sum of the two before it: 1, 1, 2, 3, 5, 8, 13, 21, 34, 55. Divide any number by the one after it and the result converges on 0.618, the golden ratio. Two places along gives 0.382, three places gives 0.236, and 78.6 percent is the square root of 0.618. Applying these ratios to markets traces back to Ralph Nelson Elliott's 1946 monograph Nature's Law: The Secret of the Universe. One number on the tool is neither his nor Fibonacci's: 50 percent comes from Dow theory, and it sits there because traders watch it.

The Fibonacci Retracement Levels and What Each One Means

The standard Fibonacci retracement levels each describe a different depth of pullback:

• 23.6 percent: a shallow pause, typical of a fast trend.
  
• 38.2 percent: a normal correction, often the first level to produce a reaction.
  
• 50 percent: the psychological midpoint, not Fibonacci but heavily watched.
  
• 61.8 percent: the golden ratio, the level most traders anchor to.
  
• 78.6 percent: a deep retracement that puts the trend in question.
  

The tool divides one completed swing into levels. Each level describes a depth of pullback, not a prediction that price will stop there.

How to Draw Fibonacci Retracement Correctly

Direction is the most common mistake. In an uptrend, drag the tool from the swing low to the swing high, so the levels sit below the high and mark potential pullback support. In a downtrend, reverse it. Drawn backwards, every level lands where price has already been.

• Use the extremes, including wicks. Anchoring to candle bodies shortens the range and shifts every level.
  
• Pick swings a stranger would also pick. If you have to hunt for the anchor points, the swing is too small.
  
• Redraw when structure changes. The levels belong to the swing, not the chart.
  

Left: the tool dragged in the wrong direction on an uptrend. Right: the correct anchoring, low to high, marking possible pullback support.

How to Use Fibonacci Retracement in a Trade

Knowing how to use Fibonacci retracement means treating a level as a place to start looking, not an entry signal. Price reaching 61.8 percent tells you nothing on its own. The level narrows the chart to an area where you can demand evidence.

Entry. Wait for a reaction at the level: a rejection wick, an engulfing candle, or a clear stall in momentum. Entering on the touch bets that the number is magic. Entering on the reaction responds to what buyers actually did.

Stop loss. Place the stop beyond the next level down. If price is reacting at 61.8 percent, the idea is not disproved until it closes through 78.6 percent. A stop in the noise between two levels gets taken out by ordinary volatility.

Take profit. The first target is the swing high that anchored the measurement. If your stop is as far away as your target, skip the trade.

The Fibonacci Golden Pocket

The Fibonacci golden pocket is the band between the 61.8 and 65 percent levels. It draws more attention than any other zone, for a circular reason. Carol Osler's study of a foreign exchange dealing bank's order book, published in the Journal of Finance in 2003, showed that take-profit orders cluster at specific prices and that this clustering produces the bounces. A level works when real orders rest on it. The golden pocket attracts orders because it is the most widely taught level, and those orders make it react.

Bitcoin's rally from the September 2024 low near 52,560 dollars to the January 2025 high around 109,350 dollars puts the 61.8 percent retracement at about 74,245 dollars, with the golden pocket running down to roughly 72,400. The March 2024 all-time high sat at 73,750 dollars, inside that band. Bitcoin then bottomed on 7 April 2025 near 74,637 dollars, half a percent above the 61.8 percent level and right on the old high. The ratio did not stop the market. It coincided with a structural level thousands of traders already watched, and together they concentrated enough buying to end the decline.


The April 2025 low formed where the golden pocket and the previous all-time high overlapped. The confluence did the work, not the ratio alone.

Fibonacci Retracement vs Extension

Fibonacci retracement vs extension comes down to whether you are measuring inside a move or beyond it.

Fibonacci retracement

• Measures: a pullback inside a completed move
  
• Levels sit: between the swing low and the swing high
  
• Key ratios: 23.6, 38.2, 50, 61.8, 78.6 percent
  
• Used for: finding an entry with the trend
  

Fibonacci extension

• Measures: continuation past the end of the move
  
• Levels sit: beyond the prior swing high or low
  
• Key ratios: 127.2, 161.8, 261.8 percent
  
• Used for: setting targets once the trend resumes
  

Use them in sequence: retracement for the entry, extension for the target.

When Fibonacci Works and When It Fails

The case for it. The most methodical test is a 2022 study in Expert Systems with Applications that used an algorithm to identify Fibonacci retracements across three equity markets. It found a positive relationship between the width of the Fibonacci zone and the probability of a bounce, while noting this does not by itself imply a profitable strategy.

The case against it. Batchelor and Ramyar's Dow study found no meaningful clustering at Fibonacci ratios once the number of tests was accounted for. The tool is also subjective where it matters most: two traders anchoring to different swings get different levels, and both can defend the choice.

Market regime is decisive. Retracement levels assume the trend is intact and the pullback temporary. On the next Bitcoin swing, April 2025 low to the October 2025 high around 126,198 dollars, the golden pocket sat between roughly 92,700 and 94,300 dollars. In November 2025 price cut straight through it to roughly 80,600 , past even the 78.6 percent level near 85,671. Nothing was wrong with the drawing. The market had stopped correcting and started reversing, and no ratio protects you from that.

Treat Fibonacci retracement as a filter, not a signal. Reducing a chart to five places worth watching is useful. Telling you price will turn at one of them is beyond the tool. The levels that matter sit on something the market already respected: a previous high, a broken support, a volume shelf. Find that overlap and wait for the reaction. The ratio is a reason to pay attention, not a reason to be in a trade.