Fibonacci Extension: Levels, How to Draw It and How Often Price Reached Them

A Fibonacci extension is a drawing tool that projects price targets from three swing points. Point A starts a move, point B ends it, and point C ends the pullback that follows. The tool takes the size of the move from A to B and adds ratios of it to C: 61.8%, 100%, 161.8% and 261.8% are the usual levels. TradingView calls it the trend-based Fibonacci extension and says traders use it to set profit targets or to estimate how far a price may move.

How the levels are calculated

MetaTrader 5, which calls the tool Fibonacci Expansion, describes the same construction: the first wave is the unit, the end of the second wave is the reference point, and levels are drawn at 61.8%, 100% and 161.8% of the unit. It presents the tool as a way to find where the third wave of an Elliott wave count may end. TradingView's Auto Fib Extension lists 61.8%, 100%, 161.8%, 200% and 261.8% as the most common levels and names 0.618, 1.0 and 1.618 as the levels most used for take profit orders.

Fibonacci retracement vs extension

Levels such as 123.6% or 138.2%, obtained by adding 100% to retracement ratios, appear in neither the TradingView nor the MetaTrader 5 help page cited here. The Fibonacci retracement page covers the two-point tool.

How to draw a Fibonacci extension

1. Find a clear move: a swing low A and the swing high B that ends the move, for a rising market.
2. Wait for the pullback to end at a swing low C, between A and B.
3. Click A, then B, then C with the trend-based Fib extension tool. The levels appear above C.
4. For a falling market, reverse the points: A is a high, B a low and C the high of the bounce; the levels appear below C.
5. Decide in advance what you will do at each level and what would cancel the setup, such as a close beyond A.

The charts below pick the points by rule: in the window, among all swing points with a set number of bars on each side, the move from A to B with the largest size where C stays between A and B.

Example 1: weekly Bitcoin, 2024 to 2026

A is the low of the week of August 5, 2024 at 49,000, B the high of the week of January 20, 2025 at 109,588 and C the low of the week of April 7, 2025 at 74,508. What followed: price first reached the 61.8% level, 111,951.38, on May 22, 2025, when the daily high was 111,980. The rise went on to 126,199.63 in the week of October 6, 2025, which is 85.3% of A to B, short of the 100% level at 135,096. Price then fell to 57,800.19 in the week of June 29, 2026, below C, and the last weekly close of the window, in the week of October 19, 2026, was 80,154.09.

Example 2: daily Bitcoin, a falling extension in 2026

Here the move is down. A is the high of May 6, 2026 at 82,850, B the low of June 5 at 59,130.91 and C the high of the bounce on June 15 at 67,292.15. A to B is 23,719.09, so the 61.8% level sits at 52,633.75. What followed: price broke below B on June 24 and made its low on July 1 at 57,800.19, only 40.0% of A to B below C. None of the extension levels was reached. Bitcoin then turned up and traded above A for the first time on September 21, at 87,395.67.

How often did price reach the extension levels?

We measured this on the daily S&P 500 index from January 2, 1970 to October 28, 2026 and on daily Bitcoin from August 2017 to October 2026. Method: a swing high is a bar whose high is strictly above the highs of the 5 bars on each side (then 10), and a swing low is the mirror on lows; a bar that is both is dropped. When two swing points of the same kind follow each other, we keep the more extreme one, so highs and lows alternate. Each three consecutive points A, B, C with C strictly between A and B form a setup, and the next swing point D measures the move: the distance from C to D as a share of the distance from A to B. A level counts as reached when that share is at least the level's ratio. This is the move that an extension drawn at C tries to forecast.

With 10 bars on each side the shares were 61.4%, 29.8%, 11.6% and 1.9% on the S&P 500, and 53.3%, 33.7%, 15.2% and 8.7% on Bitcoin, for the 61.8%, 100%, 161.8% and 261.8% levels in that order. The four samples count 647 and 319 setups on the S&P 500 (5 and 10 bars), 168 and 92 on Bitcoin; the median next swing was 74.3%, 77.1%, 64.7% and 63.4% of A to B. Other choices that read the same way are just as defensible (a high equal to its neighbors counted as a swing, the first of two same-kind points kept instead of the more extreme one, or no merging at all), and on the S&P 500 with 5 bars they give between 453 and 671 setups, with the 61.8% level reached in 59.7% to 63.5% of them. The conclusion does not change; the exact counts depend on these choices.

So the 61.8% level was reached a little more often than not, the 100% level about one time in three and the 161.8% level about one time in nine on the S&P 500. Extensions are often presented as the place where a move ends. To check that, we counted how many swings ended close to each ratio, within 0.05 of it. On the S&P 500 with 5-bar swings, 2.3% ended near 1.618, against 2.5% near 1.5 and 2.2% near 1.75; 9.1% ended near 0.618, against 11.7% near 0.5 and 6.6% near 0.75. The chart below gives the whole profile: the share generally falls as the ratio grows, with a few bumps, and none of them at a Fibonacci ratio. These are counts on swing points found by a fixed rule, without costs, not a trading test.

Researchers reached a similar view on the Dow Jones Industrial Average. In “Magic numbers in the Dow” (Cass Business School, September 2006, archived copy of September 18, 2009), Roy Batchelor and Richard Ramyar tested whether the ratios of successive trends cluster around round fractions or Fibonacci ratios: “A few significant ratios appear, but no more than would be expected by chance given the large number of tests we conduct.”

How traders use Fibonacci extensions

• Profit targets. A trader long from C may take part of the position off at the 61.8% or 100% level. On the S&P 500 numbers above, the 161.8% level was reached by about one swing in nine, so a target there is far less likely to be hit.
• Elliott wave targets. MetaTrader 5 presents the tool for the end of a third wave, measured from waves 1 and 2.
• Confluence. Some traders look for a level that lines up with a past support or resistance, or with the measured move of a chart pattern such as the ABCD pattern.

Limits of Fibonacci extensions

• The points are a choice. Another A or C gives other levels.
• Many levels. With four or five levels on the chart, one of them is usually near any turn.
• No special stopping point. In our count, moves did not end near Fibonacci ratios more often than near other ratios.
• No measured edge. Our bench tests chart patterns detected by rule, not drawing tools.

Related tools: the Fibonacci retracement, which measures the pullback that sets point C, and the measured move, another way to project a target from a first leg. More guides are in technical analysis.