Fibonacci retracement is the one tool in this series that comes with a medieval math backstory and a whiff of mysticism. Strip that away and it’s simply a set of horizontal lines drawn at fixed percentages of a price swing, marking spots where a pullback might stall. Whether that works because of cosmic ratios or because millions of traders watch the same lines is a fair debate — but the mechanics are dead simple. Let’s do the math.
The one-sentence job
Fibonacci retracement draws horizontal support/resistance levels at set percentages between a swing high and a swing low, to anticipate where a counter-move might reverse. After a big move, price often retraces part of it before continuing — and these levels flag the popular “part.”
The levels
You pick a significant swing (a high and a low), and the tool draws lines at these retracements of that move:
- 23.6% — shallow pullback.
- 38.2% — the first “real” Fibonacci level.
- 50% — not actually a Fibonacci number (more below), but widely watched.
- 61.8% — the golden ratio, the most important level.
- 78.6% — deep pullback (the square root of 61.8%).
The zone between 38.2% and 61.8% is often called the golden zone — the region where healthy pullbacks most commonly reverse and the trend resumes. On the chart above, price sells off from the swing high, then bounces and stalls right in that golden zone before rolling over.
Where the numbers come from
The ratios come from the Fibonacci sequence (1, 1, 2, 3, 5, 8, 13, 21, 34, 55,
89…), where each number is the sum of the two before it. Divide any number by the
next and you approach 0.618 (e.g. 55/89 ≈ 0.618) — the golden ratio. That’s
where 61.8% comes from. 38.2% is 1 − 0.618; 23.6% comes from dividing a number by
the one two places ahead; 78.6% is √0.618.
The 50% level is an imposter — it’s not derived from the Fibonacci sequence at all. It survives on the chart because of Dow Theory’s old observation that markets often retrace about half of a move. Traders lumped it in with the real Fibonacci levels, and it stuck.
How to read it
- Pullback targets. In an uptrend, a dip that holds the 61.8% (or the golden zone) and turns back up is a classic continuation entry.
- Confluence. Fibonacci levels are far more convincing when they line up with other evidence — a prior swing, a moving average, a round number. A level alone is weak; a cluster of reasons at the same price is strong.
- Extensions. Beyond 100%, extension levels like 161.8% project where a move might reach — used for profit targets.
Where it lies to you (and it lies a lot)
Be honest about this one — it has more caveats than most:
- It’s subjective. Which swing high and low do you use? Different choices give different levels, and it’s dangerously easy to pick the swing that makes the lines fit what you already believe. That hindsight bias is the tool’s biggest trap.
- Post-hoc pattern-matching. With five-plus levels drawn, price will hit one of them, and it’s tempting to declare victory after the fact.
- “It works because everyone watches it.” Much of Fibonacci’s predictive power is plausibly self-fulfilling — enough traders place orders at 61.8% that it becomes a real level. That’s a legitimate reason to respect it, but it’s not magic.
- The mysticism is optional. You don’t need to believe the golden ratio governs the universe to use these lines; you just need to know a lot of other people are watching them.
Used with discipline — a consistent rule for choosing swings, and confluence with other evidence — Fibonacci retracement is a reasonable framework for anticipating pullback zones. Used as a mystical oracle drawn to fit your bias, it’s astrology with numbers.
Now go test it, don’t trust it
Fibonacci is hard to backtest precisely because it’s subjective — which is exactly why you should try to make it mechanical (e.g. auto-detected swings) and test it, rather than eyeballing. Build a rules-based version in AlgoGen and see if the levels actually hold. Then compute and draw the levels yourself in Python, MQL5, Pine Script, EasyLanguage, or NinjaScript.
This post is educational, not financial advice. Indicators describe the past; they don’t predict the future. Backtest anything before you risk real money on it.
