Zigzag Structure

What is Zigzag Structure?

A zigzag reduces price action to a sequence of alternating up and down legs. The rule is a reversal threshold: while an up leg is running, the algorithm tracks the highest high, and only when price retreats from that extreme by more than the threshold (a percentage, a fixed amount, or an ATR multiple) does it confirm the high as a pivot and begin a down leg. Every move smaller than the threshold is discarded as noise.

The connected pivots are swing highs and lows, so a zigzag is best understood as a swing detector plus a magnitude filter, a close relative of Gann's swing charts, which applied the same idea with bar-count rules. It is rarely a signal by itself. Its job is data reduction: once price is simplified into legs, downstream logic can label higher highs and lower lows through a swing structure grammar, count waves, measure retracement ratios, or match patterns on the reduced series. The threshold sets the scale: small values expose minor swings, large values keep only major ones.

It matters because nearly every structural tool needs an answer to "what counts as a swing," and the zigzag is the most common answer. Its defining limitation is that the newest leg is provisional by construction: a pivot only confirms after price has already reversed by the full threshold, so the final segment can extend, move, or vanish as new bars arrive. Confirmed pivots are stable; anything built on the live leg inherits that lag.

How to calculate a zigzag

The zigzag is fully mechanical, so the clearest way to understand it is to walk the algorithm.

1. 1Choose a reversal rule: a percentage of price, a fixed point amount, an ATR multiple, or a bar-count pivot (a high with N lower highs on each side). Percentage and ATR rules adapt across price levels and volatility regimes; fixed points do not.
2. 2Track the running extreme of the current leg. In an up leg, record the highest high since the last confirmed low; in a down leg, record the lowest low since the last confirmed high.
3. 3Confirm a pivot when price reverses from that extreme by more than the threshold, then start the opposite leg from the confirmed pivot. Enforce alternation: one high, then one low, never two pivots of the same kind in a row.
4. 4Read the output as legs: compare pivot levels to label structure, measure each leg's length and duration, and treat the final, unconfirmed leg as provisional until the next reversal locks it in.

How it's calculated

A polyline connecting alternating swing highs and lows, keeping only reversals larger than a chosen threshold.

Platform parameter sets differ (TradingView uses deviation percent plus pivot depth, MetaTrader uses Depth, Deviation, Backstep) but all implement this extreme-plus-threshold recipe.

Gann swing charts replace the size filter with a bar count: the swing turns up after a set count of consecutive higher highs and down after the same count of consecutive lower lows (commonly 2 or 3).

Confirmed pivots never move, but the newest leg repaints until its reversal confirms, so zigzag describes past structure rather than signaling in real time.

How traders use it

• As the swing engine underneath structure logic: break of structure and change-of-character labels are typically computed by comparing zigzag pivots rather than raw bars, so the zigzag's threshold quietly decides what "structure" means on the chart.
• As a wave proxy: Elliott-style counts and harmonic measurements need discrete legs to work with, so zigzag pivots are the usual anchors for Fib retracement and extension ratios between swings.
• As pre-processing for pattern recognition: detecting double tops, head and shoulders, or triangles is far more tractable on a polyline of a few pivots than on raw candles, which is how many algorithmic pattern scanners are built.
• As a multi-scale lens: running several thresholds at once separates minor structure from major structure, the same idea formalized by swing magnitude filters and nested swing analysis across timeframes.

Zigzag Structure vs related concepts