Fibonacci Retracements: Why 61.8% Works Best on Past Charts

Fibonacci retracements are the 23.6%, 38.2%, 50%, 61.8% and 78.6% levels of a move, drawn between a chosen low and high. There's no evidence the market “knows” these ratios. If it reacts, it's because many traders watch the same lines, or because with that many lines one of them will always “work”.

What it is

The numbers come from the Fibonacci sequence and the golden ratio (about 1.618). The 50% level isn't a Fibonacci number, but it's added by convention. Then there are extensions (127.2%, 161.8%) and “cluster” zones.

The key problem is practical: the result depends on which low and which high you pick. Two traders will draw different grids on the same chart.

How people use it

What the research and evidence say

There's no solid research confirming that Fibonacci levels work better than any other ratios. With five to seven lines on every move, price will almost always react at one of them, which feeds the illusion that it works. The effect may exist as a self-fulfilling prophecy, when many participants place orders at the same lines.

Weaknesses

What order flow shows

Instead of asking whether the market will stop at 61.8%, order flow asks where participants actually left a trace: high-volume areas where the market accepted price, the start of a move where aggression appeared, a level where liquidity is waiting in the book. If such a level coincides with a Fibonacci line, what matters is the level, not the number.

How to use it sensibly

Related

Frequently asked questions

How do you draw Fibonacci retracements?

You draw a grid from the low to the high of a move (or the other way round). The result depends on which points you pick, and that's a subjective decision.

Which Fibonacci level is the most important?

61.8% is the one most often cited, but there's no evidence it works better than the others. Its popularity comes from tradition and the golden ratio.

Does Fibonacci work in the markets?

There's no solid evidence. If the market reacts at these levels, it's more often because of many similar orders than because of any property of the numbers themselves.

Educational content, not investment advice. No signals, entries, stops or position sizes.