Double tops and double bottoms: counted over ten years
Double top
| Instrument | Found | Counted | Hit | Miss | Undecided | Hit share | Average result, ATR |
|---|---|---|---|---|---|---|---|
| Reliance Industries | 0 | 0 | 0 | 0 | 0 | — | — |
| HDFC Bank | 0 | 0 | 0 | 0 | 0 | — | — |
| ICICI Bank | 1 | 1 | 0 | 1 | 0 | 0.0% | -1.13 |
| Infosys | 0 | 0 | 0 | 0 | 0 | — | — |
| Tata Consultancy Services | 1 | 1 | 0 | 1 | 0 | 0.0% | -1.05 |
| State Bank of India | 0 | 0 | 0 | 0 | 0 | — | — |
| ITC | 0 | 0 | 0 | 0 | 0 | — | — |
| Larsen & Toubro | 0 | 0 | 0 | 0 | 0 | — | — |
| Axis Bank | 0 | 0 | 0 | 0 | 0 | — | — |
| Bharti Airtel | 0 | 0 | 0 | 0 | 0 | — | — |
| Nifty IT | 1 | 1 | 0 | 1 | 0 | 0.0% | -1.97 |
| All instruments | 3 | 3 | 0 | 3 | 0 | 0.0% | -1.38 |
| Any session, same test, textbook direction down | 26729 | 7432 | 27.8% |
double top: the definition above applied to every occurrence the rule found. Shares are hits divided by counted.
Double bottom
| Instrument | Found | Counted | Hit | Miss | Undecided | Hit share | Average result, ATR |
|---|---|---|---|---|---|---|---|
| Reliance Industries | 0 | 0 | 0 | 0 | 0 | — | — |
| HDFC Bank | 2 | 2 | 1 | 1 | 0 | 50.0% | +0.49 |
| ICICI Bank | 1 | 1 | 0 | 1 | 0 | 0.0% | -1.13 |
| Infosys | 0 | 0 | 0 | 0 | 0 | — | — |
| Tata Consultancy Services | 2 | 1 | 0 | 1 | 0 | 0.0% | -1.62 |
| State Bank of India | 2 | 2 | 1 | 0 | 1 | 50.0% | +2.00 |
| ITC | 1 | 1 | 1 | 0 | 0 | 100.0% | +2.21 |
| Larsen & Toubro | 0 | 0 | 0 | 0 | 0 | — | — |
| Axis Bank | 0 | 0 | 0 | 0 | 0 | — | — |
| Bharti Airtel | 0 | 0 | 0 | 0 | 0 | — | — |
| Nifty IT | 0 | 0 | 0 | 0 | 0 | — | — |
| All instruments | 8 | 7 | 3 | 3 | 1 | 42.9% | +0.63 |
| Any session, same test, textbook direction up | 26729 | 10163 | 38.0% |
double bottom: the definition above applied to every occurrence the rule found. Shares are hits divided by counted. Left out: fewer than 20 sessions of history after the completing close: 1.
- Up session (close above open), hollow
- Down session (close below open), filled
- Line of the figure
- Target level
- Cancel level
- Span of the figure
- Breakout close
The rule that found it
Rule DB-1: pivots low-high-low; the two lows differ by at most 0.15 of the height of the middle high; that high lies at least 5 ATR above the higher low; the lows are 10 to 80 sessions apart; the first low came after a fall of at least one such height inside the 80 sessions before it; it counts as formed only when a session closes above the middle high within 30 sessions after the second low, with no close below either low in between.
Rule HIT-IN: a figure found by its written rule is counted at the close that completes it (for a candle model, the close of its last candle). From that close C, and the ATR of that session, the textbook direction follows from the figure: down for head and shoulders, double top, descending triangle and bearish engulfing; up for double bottom, ascending triangle and bullish engulfing. It is a hit if one of the next 20 sessions closes at least 2 ATR from C in that direction before any session has closed 1 ATR or more from C the other way. It is a miss if a close 1 ATR against comes first. If neither happens within 20 sessions it is undecided and counts as not a hit: it stays in the denominator. Only closes decide, never highs or lows. A figure with fewer than 20 sessions of history after it is left out and counted separately. The result of an occurrence is the distance from C to the close that decided it (or to the 20th close), in ATR, positive in the textbook direction.
- Up session (close above open), hollow
- Down session (close below open), filled
- Line of the figure
- Target level
- Cancel level
- Span of the figure
- Breakout close
The rule that found it
Rule DT-1: pivots high-low-high; the two highs differ by at most 0.15 of the trough's depth; the trough lies at least 5 ATR below the lower high; the highs are 10 to 80 sessions apart; the first high came after a rise of at least one trough-depth inside the 80 sessions before it; it counts as formed only when a session closes below the trough low within 30 sessions after the second high, with no close above either high in between.
Rule HIT-IN: a figure found by its written rule is counted at the close that completes it (for a candle model, the close of its last candle). From that close C, and the ATR of that session, the textbook direction follows from the figure: down for head and shoulders, double top, descending triangle and bearish engulfing; up for double bottom, ascending triangle and bullish engulfing. It is a hit if one of the next 20 sessions closes at least 2 ATR from C in that direction before any session has closed 1 ATR or more from C the other way. It is a miss if a close 1 ATR against comes first. If neither happens within 20 sessions it is undecided and counts as not a hit: it stays in the denominator. Only closes decide, never highs or lows. A figure with fewer than 20 sessions of history after it is left out and counted separately. The result of an occurrence is the distance from C to the close that decided it (or to the 20th close), in ATR, positive in the textbook direction.
The written rules found 3 double tops and 8 double bottoms. No double top was a hit (0 of 3); the double bottoms were 3 of 7 (42.9 percent), with one left out for want of 20 sessions of history. Both counts are small, and each is shown beside the share the same test gives from any session.
Two rules, two counts
Rule DT-1 and rule DB-1 are mirror images. For the double bottom: pivots low-high-low; the two lows differ by at most 0.15 of the height of the middle high; that high lies at least 5 ATR above the higher low; the lows are 10 to 80 sessions apart; and the first low came after a fall of at least one such height in the 80 sessions before it. It counts as formed only when a session closes above the middle high within 30 sessions after the second low, with no close below either low in between. The double top is the same, upside down.
Each is then judged by the site's one definition from its completing close: a double top is a hit if a close 2 ATR lower comes within 20 sessions before a close 1 ATR higher, and a double bottom the same way upward.
The double tops
Three double tops were found, one each on ICICI Bank, Tata Consultancy Services and Nifty IT, and all three were misses: 0 of 3, an average result of -1.38 ATR. The yardstick for the downward test is 27.8 percent, so three failures are not far from what chance alone could give: at that rate, three misses in a row happen about four times in ten.
The count is too small to say the double top fails. It says only that this rule, on these 11 instruments, found three and that none of the three fell 2 ATR before rising 1 ATR.
The double bottoms
Eight double bottoms were found, seven of them counted: three hits, three misses and one undecided, 42.9 percent against a yardstick of 38.0 percent for the upward test. The average result was +0.63 ATR. By instrument: HDFC Bank one hit and one miss, ICICI Bank one miss, TCS one miss (a second was left out), State Bank of India one hit and one undecided, ITC one hit.
The seventh share is above its yardstick by about five points, which on seven occurrences is a difference of about one result. A reader should not build on it. Over this decade the upward test was passed more often than the downward one from any session, so each figure is compared with the yardstick of its own direction and never with the other.
Why bottoms and tops are not compared
It is tempting to set the double bottom's 42.9 percent beside the double top's zero and conclude that bottoms work and tops do not. The two numbers sit on different yardsticks, 38.0 percent up and 27.8 percent down, and on different small samples of seven and three. Their gap is a few individual results.
The page keeps the two counts separate because the question for each is how it compares with its own baseline, and the honest answer for both is that the sample is too small to tell.
The two drawn examples
The hit is a double bottom on State Bank of India, completed on 24 May 2018 at a close of 268.50 with an ATR of 8.30; on 13 June a close came 2.31 ATR above C. The miss is a double top on Tata Consultancy Services, completed on 16 May 2022 at 3,376.80 with an ATR of 71.51; the next close, 17 May, stood 1.05 ATR above C and cancelled it.
Both are real stretches of history drawn from the stored series, and both are included in the counts above.
At a glance
Read next
Questions about the count
How many double tops and bottoms were found?
Three double tops and eight double bottoms in ten years of daily history across the 11 instruments.
Did any double top work?
By this definition, no: none of the three reached 2 ATR down before a close 1 ATR up. The sample is three.
What was the double bottom share?
Three hits in seven counted, 42.9 percent, against 38.0 percent for the same upward test from any session. One occurrence was left out because fewer than 20 sessions of history followed it.
Why are tops and bottoms on separate yardsticks?
Because the same test run upward and downward gives different shares over these histories, 38.0 and 27.8 percent, so each figure is compared with its own direction.
Can the double bottom share be trusted?
Not for conclusions. Seven occurrences make one result worth about 14 points.
Open a practice account with FxPro
A chart pattern is not an order. An order is placed on a platform, carries its own stop and take profit, and has costs. A practice account lets you see the order ticket and the platforms before any money is at stake.