How to Use Fibonacci Retracement Levels

Fibonacci retracement is a way to mark where a pullback might stall inside an existing trend, using horizontal levels at 38.2%, 50%, and 61.8% of the prior move. This page covers how the levels are drawn, how traders use them for pullback entries, and why they work as a self-fulfilling crowd map rather than a predictive rule. Research and education only — not financial advice.

The short answer

Fibonacci retracement takes one completed price swing and draws horizontal levels at fixed percentages of it — most commonly 38.2%, 50%, and 61.8% — to flag where a pullback inside a trend might pause before the trend resumes. You anchor the tool from a swing low to a swing high (in an uptrend) or high to low (in a downtrend), and it plots the retracement band automatically. It is not a signal to buy or sell; it is a map of likely reaction areas. The levels matter mostly because a large number of traders draw the same ones and act around them — which is the whole honest story, and the reason the rest of this page spends as much time on the caveats as on the setup.

Where the numbers come from

The ratios trace back to the Fibonacci sequence (1, 1, 2, 3, 5, 8, 13...), where each number sums the two before it. Divide a number by the one after it and the ratio converges on 0.618 — the "golden ratio," the 61.8% level. Divide by the number two places up and you get roughly 0.382, the 38.2% level. The 23.6% and 78.6% levels come from the same series. One honest footnote most tools bury: 50% is not a Fibonacci number at all. It survives on the chart by convention — the old Dow-theory idea that trends often give back about half a move — because traders found it useful, not because the math produced it. Keep that in mind whenever someone frames these levels as a law of nature. They are a widely shared habit dressed in arithmetic.

How to draw and use it: a step list

The mechanics are simple; the discipline is in what you do after the levels appear.

Find a clean, completed swing. Pick an obvious leg — a low to a high the whole market can see. If you have to squint to justify the anchor points, the levels built on them are noise.
Anchor the tool. In an uptrend, drag from the swing low to the swing high; the retracements plot below the high as pullback support. In a downtrend, drag high to low; they plot above the low as bounce resistance.
Mark the 38.2%-61.8% band. Treat the space between them as a zone, not three exact lines. Shallow pullbacks to 38.2% imply a strong trend; deeper ones to 61.8% imply a weaker one that is closer to failing.
Wait for a trigger inside the zone. Price touching a level proves nothing. You want a reaction — a reclaim bar, a higher low, participation. Pair the level with a defined trigger so the entry has a condition, not a hope.
Place the stop past invalidation. If price closes decisively beyond the 61.8% (or the 78.6% backstop), the pullback has gone too far to still be "just a pullback." That break is your stop, not a round number you would merely tolerate losing.
Define targets and size by risk first. Set your objective (often the prior swing high, or a Fib extension beyond it), then size the position so the loss at the stop is survivable — never the other way around.

A worked example

Say a stock rallies from a swing low of $40.00 to a swing high of $50.00 — a $10 move — then starts to pull back. Anchoring the retracement low-to-high gives:

Level	Math	Price
23.6%	50 − (0.236 × 10)	$47.64
38.2%	50 − 3.82	$46.18
50%	50 − 5.00	$45.00
61.8%	50 − 6.18	$43.82
78.6%	50 − 7.86	$42.14

A trader looking to join the uptrend on a dip watches the $43.82-$46.18 band. Suppose price grinds down to $44.10, prints a higher low, and reclaims $44.50 on rising volume — that reaction inside the zone is the trigger, not the arrival. A stop might sit just below the 78.6% at, say, $42.00: a close under it says buyers who were supposed to defend the pullback did not, so the thesis is wrong. If the entry is $44.50 and the stop is $42.00, that is $2.50 of risk; a target back at the $50.00 swing high is $5.50 of reward, roughly a 2.2-to-1 ratio. Run those exact numbers through the risk-reward calculator before deciding whether the trade clears your minimum, and size the share count with the position-size calculator so the dollar loss at $42.00 is one you can absorb.

Confluence beats a lone line. A Fibonacci level is far more interesting when something else agrees with it — a prior support level, a moving average, a VWAP, or an oversold RSI reading landing in the same zone. Alone, 61.8% is one crowded number; stacked with two other reasons, it is a location several kinds of traders are watching at once.

Self-fulfilling, not magic

The uncomfortable truth is that Fibonacci levels have no known mechanism that forces price to obey them. What they have is reflexivity: because charting platforms plot the same 38.2/50/61.8 levels for everyone, a crowd stacks bids near them, expecting a bounce — and the stacked bids can produce the bounce. The level works to the extent that enough people believe it and act, the same way a round number or a prior high earns respect beyond any fundamental logic. That is a real, tradeable effect. It is also a ceiling. When a genuine catalyst arrives — an earnings miss, a guidance cut, a downgrade — order flow from people reacting to information overwhelms the people trading the golden ratio, and price slices through every level without pausing. Habit builds these lines; news breaks them.

The honest caveats

Anchoring is subjective. Two traders pick different swing highs and lows and get different levels, then each points to whichever "worked" after the fact. If you can only justify the anchor in hindsight, you are curve-fitting, not analyzing.
Touches are not triggers. Price tags 61.8% dozens of times a week across a watchlist and reverses at only some of them. Without a confirmation and a plan, a level is a place to lose money precisely.
Strong trends ignore shallow retracements. A powerful move may only give back to 23.6% before continuing, leaving the 50-61.8% dip-buyers waiting for a pullback the trend never owes them.
It is a location, never a strategy. "Long the 61.8%" tells you where, not how much to risk, when you are wrong, or when to take profit. Those four answers are the actual trade.

A line is not a plan. A Fibonacci level with no written trigger, no stop beyond invalidation, no defined target, and no time-stop is a screenshot waiting to be rationalized after the move. The retracement can sharpen where you engage; it cannot decide your risk for you.

How we use levels on our desk

On our desk a Fibonacci zone is context, not a signal by itself — one input into where a pullback might stall and whether the crowd is defending it, weighed alongside structure, volume, and catalyst. A card is only published after it carries a written trigger, TP1/TP2, a stop, and a time-stop, and after it clears a catalyst check, adversarial review, and a liquidity screen. That format, and why it exists, is laid out on our signals page. The discipline point is the honest one: no drawing tool manufactures an edge. When we traded our raw scanner blind with every rule mechanically honored, the hypothetical backtest produced 161 simulated trades at a 46.6% simulated win rate and a 0.82 simulated profit factor — roughly negative expectancy per trade. A retracement map can make an entry legible and tighten a stop; it cannot make a weak signal worth taking. The unflattering numbers, and the audit behind them, sit in the open at our public record.

Common questions

What are the main Fibonacci retracement levels?

The most-watched levels are 38.2%, 50%, and 61.8%, with 23.6% and 78.6% as secondary markers. The 38.2% and 61.8% ratios come from the Fibonacci sequence; 50% is not a Fibonacci number at all — it is included by convention from the old idea that trends often retrace about half a move.

How do traders use Fibonacci retracement for entries?

They anchor the tool across a completed swing (low-to-high in an uptrend), treat the 38.2%-61.8% band as a potential pullback zone, and wait for a reaction inside it — a reclaim, a higher low, rising volume — as a trigger rather than buying the touch. The stop sits just beyond the level that would invalidate the pullback, often past 61.8% or 78.6%.

Does Fibonacci retracement actually work?

It is partly self-fulfilling: because everyone plots the same levels, clustered orders near them can produce the bounce. That reflexivity is real but has a ceiling — a genuine catalyst overwhelms the crowd trading the line and price slices through. On its own it is a map of reaction areas, not an edge, which is why confluence and a written plan matter far more than the level itself.

Is the 50% level really a Fibonacci level?

No. 38.2% and 61.8% derive from the Fibonacci sequence, but 50% does not appear in the math. It stays on the chart because traders found the halfway retracement useful, an idea that predates the Fibonacci tool. Treat it as a widely shared convention rather than a mathematical law.

See the process with your own eyes. The desk posts trigger-based cards to a public, timestamped record — losses included — and published the backtest where its own raw scanner loses. The scoreboard is free to watch. Join the floor →

Free to join · paid floors optional · research and education only

Last updated 2026-07-11 · ClaudeQuantAlgo Research Desk · research and education only.

Disclosures. ClaudeQuantAlgo is a research and education community. Nothing on this page constitutes financial, investment, legal, or tax advice, or a recommendation to buy or sell any security or derivative. No profit promises are made, ever — trading stocks, options, and forex involves substantial risk of loss; options positions can lose 100% of their value. The public scoreboard reflects a model ("paper") desk — no real money. Past performance — real, paper, or simulated — never guarantees future results. Hypothetical and simulated results have inherent limitations and no representation is made that any account will or is likely to achieve similar profits or losses. ClaudeQuantAlgo is not a registered investment adviser or broker-dealer. You are solely responsible for your own trading decisions. Never risk money you cannot afford to lose.