How it unfolds, step by step
The setup
Instrument R pulls back three times in the chosen window. Each time it turns near the same area: at ₹99.60, then ₹100.20, then ₹99.90. Draw that as a band and it runs roughly ₹99.50 to ₹100.50, about a rupee wide. Draw it as a line and it becomes ₹100.00 — a price none of the three reactions actually printed.
What happens next
The fourth visit arrives and goes straight through. Price closes at ₹98.70, then ₹97.40. That is below the band, and below any line drawn inside it. Nothing about the three earlier reactions was wrong. They described what happened at that area on three occasions. They made no claim at all about a fourth.
What it means — and does not
The width of a zone is the honest part of it. It is the width the observations actually had. A band says “price reacted somewhere around here, three times, in this window”. A line says “price turns at ₹100.00”. That is a stronger statement than anything on the chart supports. The count of touches is a record, not a probability. Several tests are just as easily the run-up to the visit that goes through, which is what happens here. Once a zone gives way it is often watched from the other side, with old support read as resistance. That role reversal is a sequence that happens. It is not a rule. /learn/breakout-retest-support-hold works through the version that holds. /learn/bull-trap-false-breakout and /learn/bear-trap-mechanics work through the two ways the same break turns out to have been false. A zone that narrows session by session into a tight box is a different shape again, and it belongs to /learn/range-compression-breakout.
- Price (₹)
- Zone cap ₹100.50 · 100.5
- Zone floor ₹99.50 · 99.5
Three pullbacks end at ₹99.60, ₹100.20 and ₹99.90 — a reaction zone of roughly ₹99.50 to ₹100.50, not a floor at ₹100.00. The fourth visit closes through it.
Show the underlying data
| Day | Price (₹) |
|---|---|
| 1 | 104 |
| 2 | 102.2 |
| 3 | 100.8 |
| 4 | 99.6 |
| 5 | 101.4 |
| 6 | 103 |
| 7 | 102.1 |
| 8 | 100.6 |
| 9 | 100.2 |
| 10 | 102.4 |
| 11 | 103.5 |
| 12 | 101.9 |
| 13 | 100.3 |
| 14 | 99.9 |
| 15 | 101.2 |
| 16 | 100.1 |
| 17 | 98.7 |
| 18 | 97.4 |
What this does not establish
The zone reading is broken by a close beyond the band that is not won back. Here that is the ₹98.70 close below the ₹99.50 floor, confirmed by the next one. The width is what makes that testable. A brief dip to ₹99.40 in a session that closes back at ₹100.10 refutes nothing. Two closes below the floor refute the reading entirely. Only a stated boundary can tell those two apart. People who act on a zone conventionally anchor a stop-loss beyond its far edge, not at the drawn line. The line was never where the reactions were. A stop placed inside the band is taken out by the very movement the band was drawn to hold. A zone only a few paise wide is a separate problem. The gap between the buying and the selling price, plus the cost of buying and selling once, can be bigger than the zone itself. Then the reading cannot be tested at all, rather than being merely wrong.
Things to check
5What window were the reactions counted in, and does the zone still describe a later one?
Is the boundary a band with a stated width, or a single price the reactions never printed?
What counts as a crossing here — a trade through the edge, or a close beyond it? Decide before looking.
Is the zone wider than the gap between the buying and the selling price, plus the cost of trading through it?
Check what price does after crossing the zone. A line on a chart cannot guarantee a reversal.
Common mistakes
4Turning ₹99.60, ₹100.20 and ₹99.90 into a line at ₹100.00. None of the three observations is that precise.
Counting touches as if each one raised the odds of the next hold. A level tested many times is also a level approached many times.
Redrawing the zone after it fails, so the chart still looks right. That takes away the study's power to be wrong, and with it any value it had.
Reading a line on a chart as orders waiting in the order book. The drawing and the book are two different things. The book can be empty exactly where the drawing looks thickest.
Three pullbacks ended at ₹99.60, ₹100.20 and ₹99.90. What has been observed?
- A floor at ₹100.00
- An area around ₹100 where price reacted three times in this window
- That the area is now likelier to hold than it was before the third reaction, because each test that holds adds weight to it
Show the answer
B. An area around ₹100 where price reacted three times in this window
The three reactions describe an area. Its width is the part they actually support, and a line at ₹100.00 is a precision none of them printed. Nor does a third test make a fourth one hold. Repeated visits are just as easily the run-up to the one that goes through, which is what this page's own example does.
Frequently asked
3Should a level be one exact price or a band?
A band, in most cases. It carries the uncertainty the reactions actually held. A single price stands up only where the reactions really did print at it: a round number, an earlier session's high, a settlement value. Even then, it is the area around it that gets tested.
How many touches make a zone reliable?
No number does. Touches are observations of what already happened. They say nothing about the next visit. /learn/sample-size-what-twelve-occurrences-prove works through why a handful of occurrences cannot carry that weight, whatever they are occurrences of.
Is a zone that fails a bad zone?
No. A zone describes a past reaction, and descriptions do not fail. Readings fail. The failure here is useful because it shows what three touches were, and were not, evidence of. That is why this library keeps the examples that go wrong.