Written and copyright © 2011-2013 by Thomas N. Bulkowski. All rights reserved. Revised October 14, 2011.
Busted Triple Top Summary
Busted triple tops are triple tops in which price breaks out downward and drops less than 10%, turns around, and then closes above the top of the triple top.
The following move in the busted direction (upward) averages 43% and the median gain is 34% (both numbers for single busted patterns only), or a 20% median gain for all busted triple tops.
Those numbers represent perfect trades so expect different results from actual trades.
To avoid failures of a busted triple top, trade only those in an upward price trend (which is nearly always the case) when they appear within 100 days (about 3 months) of the start
of the up trend, measured from the trend start to the top of the first peak in the busted triple top.
If price closes below the bottom of the busted triple after trade entry, then sell.
Single Busted Triple Tops
Busted triple tops come in three varieties: single busts, double busts, and three or more busts.
The chart of Bed Bath and Beyond (BBBY), pictured on the daily scale, shows an example of a single bust.
Price forms a triple top at A, B, and C, with the three peaks at nearly the same price. The chart pattern confirms as a valid triple top when price closes below the lowest valley
in the chart pattern. I show that by the horizontal red line at F. Price continues down to D before stopping at a support area.
Notice that the decline from F to D is less than 10% of F. In other words, price drops less than 10% before beginning a recovery. That recovery takes price up, closing above
the top of the triple top at E.
The close above the highest peak in the triple top means it busted the downward breakout.
Although not shown, price continues higher by more than 10%, confirming that this move is a single bust. That means price busted the downward breakout and continued moving
higher until the stock climbed at least 10% above the highest peak in the chart pattern. To qualify as a single bust, price must rise at least 10% above the top of the
For a single bust, look for:
- Price must confirm the triple top by closing below the bottom of the triple top.
- Price must drop less than 10%.
- Price rises and closes above the top of the triple top.
- Price continues rising at least 10% before either closing below the bottom of the triple top or dropping by more than 20% from a peak.
The last point, 4, means the ultimate high must be at least 10% above the top of the triple top. If price fails to climb more than 10%, then it could be forming
a double bust.
Double Busted Triple Tops
The chart of MGE Energy, pictured on the daily scale, shows an example of a double busted triple top.
Price forms the triple top at peaks A, B, and C. The chart pattern confirms when price closes below the lowest valley in the chart pattern. That price is
at D and price closes below it on the way down to E.
The drop from D to E is less than 10%.
Price reverses direction at E and moves higher, to F. At F, price closes above the highest peak in the triple top, confirming a busted triple top chart pattern. Just after it
busts the triple top, the up-move falters and the stock drops down to G.
Notice that G again closes below the horizontal red line. That line represents the lowest valley in
the chart pattern. A second close below this line busts the triple top a second time.
If the stock climbed higher than 10% above the highest peak in the triple top (meaning point F was more than 10% above the peak at A in this example),
the chart pattern would only be a single bust even though price may have dropped to G.
Since I captured this chart pattern recently, additional price action may change a double bust into a triple bust (or more).
For a double bust, look for these elements.
- Price must confirm the triple top by closing below the bottom of the triple top (below the red line at D).
- Price must drop less than 10% (the move from D to E).
- Price rises and closes above the top of the triple top (this happens at F).
- Price fails to rise at least 10% above the highest peak in the triple top before reversing (the rise from the blue line to the peak at F).
- Price closes below the lowest valley in the triple top a second time (point G).
Triple Top, Triple Busts
Triple busts are rare, but the word triple is misleading. I counted every pattern that had three or more busts.
The chart of Met Life is an example of a multiple bust triple top.
The triple top occurs at A, B, and C. This triple top is not wide, but it is tall. Price drops after peak C,
but does not close below the bottom of the triple top until D. I show a red line connecting the lowest valley in the chart pattern as the benchmark.
Price rises up to E and closes above the highest peak in the chart pattern, busting the triple top for the first time. Then price reverses and closes below the bottom of the pattern, at F.
That is the second bust. The chart pattern busts again at G when it closes above the ABC peaks.
For a triple (or more) busted triple top, look for the following:
- Price must confirm the triple top by closing below the bottom of the triple top (point D).
- Price must drop less than 10%.
- Price rises and closes above the top of the triple top (point E).
- Price fails to rise at least 10% above the highest peak in the triple top before reversing.
- Price closes below the lowest valley in the triple top a second time (point F).
- Price must drop less than 10%.
- Price rises and closes above the top of the triple top again (point G).
- Additional down and up cycles may continue, busting the triple top.
Methodology for Testing Busted Triple Tops
I found 1,033 triple tops in 613 stocks dating back as far as July 1991 to September 2011. Few stocks covered the entire period. All of the triple tops I found manually either
using a historical search or real time (looking at my stocks each day). The real time editions prevented any look ahead bias since I am not privy to future price movements.
I then used software to measure performance and flag potential busted chart patterns.
Gauging performance uses the same method as I used to catalog non-busted chart patterns. That is, the search for the new ultimate high or low proceeded as described in the
glossary. Thus, the numbers reported in Results should be considered perfect trades.
Busted Triple Top Test Results
The following numbers are the results from perfect trades in bull markets, unless otherwise noted. Do not expect actual trading results to match those discussed below. Use the numbers
only for comparison purposes with other chart patterns.
How often do triple tops bust?
- Single busts: 24% of the time.
- Double busts: 6% of the time.
- Triple or more busts: 7% of the time.
- All busted triple tops: 37% of the time.
Of busted triple tops, what is the frequency distribution?
- Single busts: 65% of busted triple tops bust only once.
- Double busts: 15% of them bust twice.
- Triple or more busts: 19% of them bust at least three times.
If you trade a busted triple top, there is a 65% probability that it will bust just once. Thus, 35% fail to show price rising by more than 10% above the top of the
What is the average rise for single busted triple tops? Answer: As measured from the top of the chart pattern to the ultimate high the rise averaged 43%.
- The average rise: 43%
- Median (mid range) rise: 34%
- The average rise (for comparison) of triple bottoms (busted and non-busted) is: 32% (from 610 triple bottoms in a bull market, updated to September 2011)
What is the rise from perfect trades after all busted triple tops?
- The average rise in a bull market: 29%
- Median (mid range) rise in a bull market: 20%
- The average rise in a bear market: 15%
- Median rise in a bear market: 13%
What is the failure rate of all busted triple tops? The answer appears in the below table.
Failure Rate for Busted Triple Tops
|Failure rate:|| 5% || 10% || 15% || 20% || 25% || 30% || 35% || 50% || 75% || >75% |
|Number of triple tops:||43||36||17||16||16||12||9||27||24||20|
For example, there were 43 triple tops that failed to show price rising at least 5% above the top of the triple top.
Those 43 represent 20% of all busted triple tops. On a cumulative basis (a running total), the 43 also represent 20% of all busted triple tops.
The median rise of all busted triple tops is 20%. You can see that in the 20% column where the cumulative line reads 51% (which is above 50%, but the 20% median doesn't change).
Thus, half of all busted triple tops will see price rise 20%, providing they are traded perfectly.
Trading Busted Triple Tops
I show a chart of El Paso Corp (EP) on the daily scale. An awkward triple top appears at ABC. The three tops are not exactly on the same price level, but I chose this
stock to show flaws, not perfection.
Point G is the lowest low between the three peaks. At D, price closes below G, confirming the chart pattern as a valid triple top. Price drops and makes a low 6% below
the price at G before turning.
Price climbs and busts the triple top at E. In this example, however, price tumbles to F. Notice that F is below G, but it does not close below G. Thus, this triple
top still qualifies as a single bust.
Price meanders upward, off the chart. In fact, a year later, in May 2011, the stock reached a high of 21.54 for a gain of 71% before it tumbled at least 20%.
I chose this example to illustrate and emphasize how bumpy the road to riches can be. A stop placed a penny below G would have cashed you out for a loss instead of holding on
to a big winner. Assuming an entry price a penny above the peak at A (12.59) and a stop a penny below the low at G (10.76), the loss would have been a massive 14.5%.
I tested the idea of buying at E and selling at various times in the future. Here is what I found.
Using a 50-day or 200-day simple moving average (SMA) as a confirmation tool did not help much. That is because price must climb to E, which means price is almost always above the
moving average. Thus, a comparison with price at the buy time below or above the SMA showed too few samples to be fair. Nevertheless, here are the results.
I checked smaller moving averages (9 and 20) and also show their results. Samples below 30 should be considered suspect.
- Buy if price above 50-day SMA, sell at ultimate high. Average gain: 29% from 213 patterns.
- Buy if price below 50-day SMA, sell at ultimate high. Average gain: 31% from 7 patterns.
- Buy if price above 200-day SMA, sell at ultimate high. Average gain: 30% from 194 patterns.
- Buy if price below 200-day SMA, sell at ultimate high. Average gain: 21% from 22 patterns.
- Buy if price above 9-day SMA, sell at ultimate high. Average gain: 29% from 215 patterns.
- Buy if price below 9-day SMA, sell at ultimate high. Average gain: 16% from 11 patterns.
- Buy if price above 20-day SMA, sell at ultimate high. Average gain: 28% from 219 patterns.
- Buy if price below 20-day SMA, sell at ultimate high. Average gain: 24% from 7 patterns.
Then I used a price-SMA crossover technique. Buy when the pattern busts providing price is above the SMA, and sell the day after price closes below the SMA (or the reverse,
buy when price is below the SMA and sell at the open the day after it closes above the SMA).
Gains were 3% (212 patterns) versus 2% (7 patterns) for above and below the 50-day SMA. For the 200-day SMA, the gains were 4% (184 patterns) versus 2% (22).
Finally, I tested the gains over time. That means buying when price busted the triple top and selling sometime in the future.
- Sell after 1 week: 2% from 220 trades.
- Sell after 2 weeks: 2% from 220 trades.
- Sell after 3 weeks: 2% from 220 trades.
- Sell after 1 month: 3% from 219 trades.
- Sell after 2 month: 4% from 216 trades.
- Sell after 3 month: 5% from 215 trades.
Again, the results are unimpressive.
What this tells me is that a mechanical exit does not work well.
Trend Start: An Idea That Works
I was paging through examples of busted triple tops and noticed the position in the price trend. Those that busted near the start of the trend tended to outperform. Those that
busted near the end of the trend were likely to lead to small gains or double/triple bust. So, I tested this idea and found it to be true. If the time from the trend start to the
first peak in the triple top is less than 100 days (call it 3 months), then the busted triple top outperforms those longer than 100 days by 37% (92 samples) to 23% (115 samples).
The glossary defines the trend start as: "Where the price trend begins, as measured from a major price turning point. To determine
the trend start, look backward in time for the lowest low followed by a rise of at least 20%, or the highest high followed by a decline of at least 20%. Those turning points, the lowest low
or highest high, represent where the trend starts. Whether to search for the lowest low or highest high, I use whichever drops below or rises above the bottom/top of the chart pattern first."
The chart shows how to make this work.
Begin with a busted triple top ABC. From the top of peak A, look to the left for a trend change -- price should either drop from a peak or rise from a valley by at least 20%.
In this example, looking backward in time, the trend start began with a low price and the stock climbed at least 20% before that low. The 20% rise is not the distance from
the trend start to peak A. Rather it is the rise from the trend start to D.
Price need not look like this V-shaped turn, from trending down to up leading
to the start of the triple top. If you imagine everything to the left of peak A as inverted, the stock climbs by 20%, peaks, and then slides downward into the triple top. The peak
is where the trend starts. In this drawing, the valley is where the trend starts.
Once you find the trend start, the time from that point to the first peak should be less than 100 days or about 3 months.
Entry Setup for Busted Triple Tops
As an entry setup, here are the rules for trading busted triple tops.
- Find a confirmed triple top.
- Locate the trend start leading to the first peak. If it is longer than 100 days, skip the trade.
- Price must not close lower than 10% below the bottom (lowest valley) of the triple top.
- Price rises and closes above the top (highest peak) of the triple top. This confirms a busted triple top.
- Buy at the open the next day.
- Place a conditional order to sell if price closes below the bottom of the triple top.
- Hold until the trend changes and then sell.
This entry setup can mean a large drop from the entry to the stop price, if things go wrong. Even if things go right, how do you detect a trend change to sell? If price climbs in a steady trend,
then selling when price closes below a trendline often works well. Perhaps other bearish chart patterns will appear and signal the move from bull to bear. Or perhaps you wish to use indicators.
Use whatever sell method works for you.
-- Thomas Bulkowski
Written and copyright © 2011-2013 by Thomas N. Bulkowski. All rights reserved. God is an atheist!