What a Fibonacci retracement is (and what it is not)

Take any completed price swing: a low to a high, or a high to a low. A retracement asks a simple question about the pullback that follows: how much of that swing did price give back?

The grid answers in the swing’s own coordinates: 0 sits at the swing’s origin and 1.0 at its end, with the ratio stations in between. Give back a little of the move and the pullback stops high on the grid, near the .786 or .618. Give back half and it stops at the .50. Give back most of it and price carries down to the .382 or .236. The ratios are addresses on the swing — a ruler laid against it, read from the origin up.

Why traders watch these levels

Markets spend most of their time in pullbacks and rallies inside larger structures. Traders who buy trends want the pullback to end somewhere sensible; the retracement grid gives everyone a shared set of somewheres. The .50 and .618 in particular are watched by enough participants that reactions there become partly self-reinforcing: people expect a response at the level, so they act at the level, so the level responds.

What a retracement is not

A retracement level is not a floor, and it is not a prediction. Price slices through perfectly good .618 levels every day. A level only means: this is a measured, widely watched place where a reaction is plausible. Whether a reaction actually happens is something you observe, not assume — which is why FibSetups pairs every zone with a confirmation ladder instead of treating the touch as the trade.

The anchoring problem

Two traders can draw different retracements on the same chart because they anchored different swings. That is the real weakness of hand-drawn fibs: the level is only as good as the swing selection behind it. FibSetups solves this mechanically — swings must be confirmed before they anchor a grid, every qualifying swing gets a grid, and zones only form where independent grids agree.