Читать книгу Encyclopedia of Chart Patterns - Thomas N. Bulkowski - Страница 121

Table 9.2 shows general statistics for this chart pattern.

Number of formations. I found 1,223 patterns in stock data from July 1991 to October 2019 in 703 stocks. However, after slicing‐and‐dicing the bear market samples, there were too few to present here. So the following tables only show bull market statistics. Not all stocks covered the entire range, and some no longer trade.

Reversal (R), continuation (C) occurrence. Because the inbound price trend can come from any direction, we can't claim this pattern is a bottom or a top. But we can compare the inbound trend with the direction after the breakout. If the two are in the same direction, then the pattern acts as a continuation of the prevailing price trend.

If price enters the pattern from the top and exits out the top, that's a reversal. The same can be said if price enters from the bottom and breaks out downward (it's also a reversal).

Upward breakouts act as continuations most often, so we know the inbound price trend must have been upward, too. Downward breakouts act more often as reversals (suggesting price was trending upward into the pattern).

Reversal/continuation performance. Reversals for both breakout directions show better performance than continuations.

Average rise or decline. The average rise or decline isn't exceptional. As I mentioned, this pattern is a mid‐list performer, so don't expect a standing ovation.

Table 9.2 General Statistics

Description	Up Breakout	Down Breakout
Number found	551	455
Reversal (R), continuation (C) occurrence	25% R, 75% C	76% R, 24% C
Reversal/continuation performance	47% R, 41% C	–15% R, –14% C
Average rise or decline	43%	–14%
Standard & Poor's 500 change	12%	–2%
Days to ultimate high or low	243	51
How many change trend?	55%	25%

Standard & Poor's 500 change. The performance of the chart pattern beats the tar out of the index. I can't think of any pattern that failed to beat the general market results. That suggests the measure favors the chart pattern.

The chart pattern is performing at its best, from the breakout to the ultimate high or low. But the index, using the same dates, may fall well short of what it's capable of. However, the numbers also show how the general market assists individual stocks to perform. The market rises during upward breakouts and falls during downward ones.

Days to ultimate high or low. How long will your trade last? It lasts as long as you do not close out your position. However, I measured the average hold time from the breakout to the ultimate high or low.

Pop quiz: If it takes 243 days for price to rise 43% after an upward breakout, how long should it take price to drop 14% after a downward breakout, assuming the same velocity? Answer: 79 days. However, the table shows that it completes the trip in just 51 days. Thus, price drops much faster after a downward breakout than it rises in an uptrend. Often, price drops twice as fast.

How many change trend? In a gauge of how well price moves more than 20% from the breakout, this pattern does well. However, it's still a mid‐list performer. Have I mentioned that?

Table 9.3 shows failure rates for the broadening pattern. For example, I found that 15% of the patterns with upward breakouts failed to see price rise more than 5% after the breakout. Downward breakouts failed almost twice as often. Yuck.

Notice as you scan down the list how failure rates increase. Half of all upward breakouts will see price fail to rise 25%. Downward breakouts see half the patterns failing to rise more than 10%.

If you want to average 50% on your trades, 72% of them will fail to meet the threshold after an upward breakout. And that's if you trade it often and perfectly. You could make more or less, depending on your skill and the situation (such as just after a bear market ends when even the losers are winning). Sprinkle in some losing trades and your winners will have to make even more to reach your 50% target.

Table 9.4 shows breakout‐related statistics.

Breakout direction. The breakout direction is almost random with a slight advantage going to upward breakouts.

Yearly position, performance. I sorted the breakout price into the yearly high–low range and mapped performance on top of it. The table shows that breakouts occurring near the yearly low do better than those near the yearly high. It suggests this chart pattern does well with bottom‐fishing strategies (buy low, sell high). Avoid momentum trading this pattern (buy high, sell higher).

Throwbacks and pullbacks. Throwbacks and pullbacks occur about two‐thirds of the time. Price leaves the pattern for 6 days until it reaches the apex, either rising or falling 6% during the journey (depending on the breakout direction), and the stock returns to the breakout in another 6 days, for a 12‐day roundtrip.

Table 9.3 Cumulative Failure Rates

Maximum Price Rise or Decline (%)	Up Breakout	Down Breakout
5 (breakeven)	83 or 15%	129 or 28%
10	71 or 28%	106 or 52%
15	62 or 39%	62 or 65%
20	32 or 45%	44 or 75%
25	40 or 52%	28 or 81%
30	30 or 58%	22 or 86%
35	24 or 62%	19 or 90%
50	56 or 72%	35 or 98%
75	52 or 82%	10 or 100%
Over 75	101 or 100%	0 or 100%

Table 9.4 Breakout and Post‐Breakout Statistics

Description	Up Breakout	Down Breakout
Breakout direction	55% up	45% down
Performance of breakouts occurring near the 12‐month low (L), middle (M), or high (H)	L 58%, M 43%, H 40%	L –17%, M –14%, H –12%
Throwbacks/pullbacks occurrence	68%	63%
Average time to throwback/pullback peaks	6% in 6 days	–6% in 6 days
Average time to throwback/pullback ends	12 days	12 days
Average rise/decline for patterns with throwbacks/pullbacks	39%	–13%
Average rise/decline for patterns without throwbacks/pullbacks	50%	–16%
Percentage price resumes trend	72%	48%
Performance with breakout day gap	48%	–16%
Performance without breakout day gap	42%	–14%
Average gap size	$0.50	$0.40

Notice that performance improves if a throwback or pullback does not occur.

After a throwback or pullback completes, we see that price resumes trending upward after an upward breakout but struggles to drop after a downward breakout. Be careful shorting this pattern after a downward breakout. A pullback may see price drop as far as it's going to.

Gaps. Regardless of the breakout direction, a breakout day gap helps performance. That's good news. Why? Because I measured performance from the opening price the day after a gap to the ultimate high or low. Thus, you can buy into the situation after you see a gap and participate in the better‐performance party.

Table 9.5 shows size‐related statistics.

Height. For both breakout directions, broadening patterns taller than the median height performed better than did their shorter counterparts.

To use this finding, measure the height of the pattern from top to bottom and divide by the breakout price. If the result is higher than the median listed in the table for the associated breakout direction, then the pattern is tall.

Width. Wide patterns performed better than narrow ones. Take the difference between the end date and start date of the pattern and compare it to the median width in the table. Wide patterns will exceed the median.

Height and width combinations. Tall and wide patterns outperform all other combinations. As a general rule for this pattern, avoid tall and narrow patterns with upward breakouts and avoid short and narrow patterns with downward breakouts.

Table 9.6 shows volume‐related statistics.

Volume trend. I used linear regression to determine the volume trend. I found that it trends upward in almost two of every three patterns.

Rising/Falling volume. Upward breakouts don't show a big performance difference between volume trending upward or downward. Downward breakouts tend to prefer falling volume for better performance.

Table 9.5 Size Statistics

Description	Up Breakout	Down Breakout
Tall pattern performance	45%	–17%
Short pattern performance	40%	–12%
Median height as a percentage of breakout price	10.6%	11.4%
Narrow pattern performance	40%	–13%
Wide pattern performance	46%	–16%
Median width	50 days	48 days
Short and narrow performance	40%	–11%
Short and wide performance	40%	–14%
Tall and wide performance	48%	–17%
Tall and narrow performance	39%	–16%

Table 9.6 Volume Statistics

Description	Up Breakout	Down Breakout
Volume trend	63% upward	62% upward
Rising volume trend performance	42%	–13%
Falling volume trend performance	43%	–16%
Heavy breakout volume performance	44%	–15%
Light breakout volume performance	39%	–12%

Table 9.7 How Often Stops Hit

Description	Up Breakout	Down Breakout
Pattern top	74%	2%
Middle	22%	17%
Pattern bottom	2%	70%

Breakout day volume. Heavy breakout volume helped the pattern perform for both breakout directions. That finding matches conventional lore where lots of sources claim that a high volume breakout helps. Here's the proof for this chart pattern. However, a survey of the chart patterns says it's only true for upward breakouts, not downward ones.

Table 9.7 shows how often price reaches a stop location. I split the pattern into thirds and found how often price as it moved from the breakout to the ultimate high or low slid into one of the thirds.

For example, I found that price on the way to the ultimate high (upward breakouts) would touch the top of the chart pattern 74% of the time. So a stop placed there would trigger too often to be viable. If you moved the stop to the bottom of the pattern, price would reach it only 2% of the time, but if it were to trigger, you might see a big loss.

Thus, stop placement, where you have to balance the loss size with how often it'll be triggered, is something traders need to master.

Table 9.8 shows the performance over three decades.

Performance over time. The 2010s, for upward breakouts, showed the worst performance for this chart pattern compared to the other two decades. For downward breakouts, the 2000s were worst. I excluded the two bear markets during that decade, too.

Failures over time. The most recent decade, the 2010s, showed substantially higher failure rates than the 1990s. I'm not sure why that is.

Table 9.9 shows busted pattern performance.

Busted patterns count. A quarter to nearly half of broadening patterns will bust. That means price moved no more than 10% away from the breakout price before reversing and closing beyond the other side of the pattern.

Table 9.8 Performance and Failures Over Time for Bull Markets

Description	Up Breakout	Down Breakout
1990s	41%	–16%
2000s	50%	–13%
2010s	36%	–14%
Performance (above), Failures (below)
1990s	10%	20%
2000s	13%	34%
2010s	23%	33%

Table 9.9 Busted Patterns

Description	Up Breakout	Down Breakout
Busted patterns count	146 or 26%	220 or 48%
Single bust count	77 or 53%	140 or 64%
Double bust count	45 or 31%	10 or 5%
Triple+ bust count	24 or 16%	70 or 32%
Performance for all busted patterns	–14%	33%
Single busted performance	–23%	49%
Non‐busted performance	–14%	43%

Busted occurrence. I counted how often price single, double, or triple (or more) busted a pattern. See the Glossary for details. Single busts happened the most. You might think that triple busts would be rare, but they often come in second place (which they do here, but only for downward breakouts).

Busted and non‐busted performance. I compared the performance of all busted patterns, single busted patterns, and non‐busted patterns. I wanted to know if busted patterns performed better than non‐busted ones.

The table shows that single busted patterns performed better than the other two categories. The problem with a single bust is you don't know ahead of time that a pattern will single bust.

Non‐busted patterns performed better after downward breakouts compared to all busted patterns. All busted patterns means it's the average for the three busted types (single, double, and three or more).