The (Average True Range) measures how far a price moves on a normal day. For each day, the largest of three ranges counts: high to low, high to the previous close, low to the previous close. The rule forms a smoothed average of this over 14 days.
What the rule needs the ATR for
First, it confirms turning points: the wave 1 high only counts after a pullback of 3 ATR, a after 1 ATR. Second, the stop lies half an ATR below the start of wave 1. Third, the signals page shows the distance to the buy zone in ATR.
Why not just percent?
A calm symbol may move 1% a day, a restless one 5%. A fixed distance in percent would be too tight for one and too wide for the other. Measured in ATR, it fits every symbol. That is why the signals page also shows the distance to the buy zone in ATR.
Do the math
High 104, low 100, previous close 99: the ranges are 4, 5 and 1 – the largest is 5. It is smoothed like this: new ATR = (old ATR × 13 + today’s range) ÷ 14.
Model example: ATR in the setup
In the model the ATR stays fixed at 4.00; on Day 34 the price closes at 124.60 – about 3.2% per day. The high 150.00 (Day 18) counts as confirmed on Day 21: the daily low 136.70 is more than 3 × 4.00 = 12.00 below it. Buffer below the start: 0.5 × 4.00 = 2.00, so stop 100.00 − 2.00 = 98.00.
Model numbers, not real prices. Zone, targets and stop calculated with the constants of the current rule. An example for understanding, not a forecast and not investment advice. Real buy zones in the screener