Zig zag indicator

A filter drawn over price rather than a signal generator. It marks the extremes of a move, joins each to the next with a straight line, and discards everything between them, so a chart that showed hundreds of bars shows a handful of legs. The setting that decides what survives is a minimum reversal, stated either as a percentage of the current extreme or as a number of points, and it is the whole of the indicator: raise it and only the largest turns are drawn, lower it and the line begins to follow the noise it was meant to remove. Most implementations add a second input governing how many bars must separate two extremes, so a swing high cannot be recorded immediately beside another one.

Construction is mechanical. While price extends a rising leg, the leg's endpoint moves with each new high. It is fixed only once price has fallen from that high by at least the stated reversal, at which point the high becomes a confirmed turn and a falling leg begins from it. The same logic runs in reverse for a falling leg. Nothing about the rule is predictive: every point on the line is a price that has already traded, and the line simply states which of those prices the setting considered large enough to keep.

The trip is that the last leg is provisional, a behaviour usually called repainting. The most recent extreme has not yet been confirmed by a reversal of the required size, so the line redraws as price moves, and a leg visible this hour can be gone by the next. This matters far beyond appearance: a strategy tested against the indicator's finished lines is tested against knowledge that was not available at the time, which flatters results in a way a backtesting report will not flag. Practitioners are broadly agreed on that limitation and disagree about what remains: the common uses are measurement and labelling, anchoring a retracement grid, counting waves, or describing how far and how long the market's recent legs have run, rather than treating any leg as evidence about the next one.