Fibonacci retracement and extension levels

What are Fibonacci retracement and extension levels?

A Fibonacci retracement divides a completed price leg into familiar proportions such as 38.2%, 50% and 61.8%. Traders use those internal levels to describe how much of the move has been given back.

A Fibonacci extension projects beyond the original leg. Ratios such as 127.2%, 161.8% and 200% are often watched as possible continuation, resistance or target areas.

How to draw Fibonacci levels with consistent anchors

Anchor selection is the real work. Two traders who choose different legs produce different grids, so a backtest must define the choice before it can claim a win rate.

• Choose one completed directional leg with an obvious beginning and end.
• For a rising leg, measure from the swing low to swing high; reverse the direction for a falling leg.
• Keep the timeframe and price scale unchanged during the comparison.
• Do not replace an anchor merely because a later level looks cleaner.

Do Fibonacci levels predict bounces? Our study found a limit

Across our monthly and weekly comparison study, Fibonacci overlap by itself did not reliably separate future bounces from control locations. A placebo grid often identified many of the same pullbacks.

That result does not make the levels useless. Fib can describe location and trade geometry, but the statistics below belong to the complete setup that combines context, qualification, entry, stop and exit.

Fibonacci retracement levels during a pullback

A shallow retracement gives back a small part of the prior move; a deep retracement returns closer to its origin. The ratio is descriptive. It does not tell you whether the pullback is healthy, whether price will reverse there or where risk belongs.

When a retracement sits near separately defined support, traders call that confluence. The overlap may focus attention, but it should not inherit an unsupported probability claim.

Fibonacci extension levels for targets and resistance

Extensions are useful for planning beyond a prior high or low, where fewer obvious historical prices exist. A projected level can act as a target, a resistance candidate or a checkpoint for managing risk.

The target rule must say whether an order exits at the level, whether a close is required or whether the position trails after arrival. Those choices materially change the result.

Trend line and Fibonacci confluence without double-counting

Two tools pointing to the same area are not automatically two independent pieces of evidence. They may both be derived from the same price swings. Treat confluence as a location filter, then test the complete setup against a fair comparison.

TradingPal shows the visible line and nearby Fib context without publishing the internal tolerances or selection machinery that creates the candidate set.

Common Fibonacci trading mistakes

• Changing anchors until a historical reaction lands on a preferred ratio.
• Assuming 61.8% is inherently stronger than every other level.
• Calling nearby values confluence without one consistent distance rule.
• Using a target level without defining the entry and invalidation first.

The sample, instrument scope and dates belong to this exact setup book. Extreme drawdowns remain visible rather than being hidden.

Backtest window: 2021-12-06 to 2026-07-13. Data snapshot: 7/23/2026. Historical results can differ from live fills and do not predict future performance.