Imagine someone named Xiao Tang who works the night shift at a convenience store by his apartment complex. After restocking the shelves and dimming the lights in the refrigerators in the early hours of the morning, he looks at the daily charts. He sets himself a rule: first, wait for at least 2 consecutive daily candles to close lower than the day before. Then, if the next candle has a lower shadow at least twice the length of its body, and if you divide the day's high-to-low range into three equal parts, its close is in the top third—right on the dividing line counts too—he calls it a hammer. He doesn't buy on the day of the hammer; he waits until the next day. If the next day's close is above the hammer's high, he buys at that day's close, with just one position. A spike above the high intraday doesn't count if the close doesn't finish above it. He also writes down two exit rules for comparison. With one, he waits for the first day that closes lower than the day before, then exits at that day's close. With the other, he waits until the close falls below the hammer's low; an intraday dip below it doesn't count. He also compares that with holding for 5 trading days.

Among the 90 daily candles from August 12 to November 9, there were 3 instances of at least 2 consecutive down days followed by a hammer like this. Only twice did the next day's close actually finish above the hammer's high.

The first was in October. On October 12, $SYRUP closed at 0.2415; on October 13, it closed at 0.2318; and on October 14, at 0.2116—2 consecutive lower closes. On October 15, it opened at 0.2117, reached a high of 0.2138, dipped to a low of 0.1942, and closed at 0.2102. The body was only 0.0015, while the lower shadow was 0.0160, or 10.67 times the body's length. The close also landed in the upper third of the day's range. On October 16, it opened at 0.2102 and reached a high of 0.2159, which was 0.0021 above the hammer's high of 0.2138, or 0.98% higher. But it closed at 0.2055, 0.0083 below 0.2138, or 3.88% lower. It broke above the level intraday but didn't close above it, so he didn't buy.

A few days later, the second setup appeared. October 16 closed at 0.2055, October 17 at 0.2009, and October 18 at 0.1943—another run of consecutive lower closes. On October 19, it opened at 0.1942, reached a high of 0.1958, dipped to a low of 0.1860, and closed at 0.1953. The body was 0.0011, and the lower shadow was 0.0082, or 7.45 times the body's length. On October 20, it opened at 0.1953, reached a high of 0.2041, dipped to a low of 0.1890, and closed at 0.2020. The close was 0.0062 above the hammer's high of 0.1958, or 3.17% higher; from open to close, it rose 3.43%. The conditions were met, so he bought at 0.2020.

On October 21, it opened at 0.2019 and reached a high of 0.2085, which was 0.0065 above his entry price, or 3.22% higher. It closed at 0.2018, 0.0002 below his entry price, or 0.10% lower. This was the first day after his purchase to close below the previous day's close. Under the first rule, he exited that day. The day's low was 0.1984, still above the hammer's low of 0.1860. If he had instead waited to exit until a close below 0.1860, it closed at 0.1817 on October 23, 0.0203 below his entry price, or 10.05% lower. If he had held for the full 5 trading days, it closed at 0.1733 on October 25, 0.0287 below his entry price, or 14.21% lower. If he had held all the way to November 9, it closed at 0.1353, 0.0667 below his entry price, or 33.02% lower; in between, on November 8, it had fallen as low as 0.1287, 0.0733 below his entry price, or 36.29% lower.

The third setup was at the end of October. October 27 closed at 0.1750, October 28 at 0.1674, and October 29 at 0.1508. On October 30, it opened at 0.1509, reached a high of 0.1566, dipped to a low of 0.1394, and closed at 0.1560. The body was 0.0051, and the lower shadow was 0.0115, or 2.25 times the body's length. On October 31, it opened at 0.1559, reached a high of 0.1624, dipped to a low of 0.1528, and closed at 0.1599. The close was 0.0033 above the hammer's high of 0.1566, or 2.11% higher; from open to close, it rose 2.57%. He bought another position at 0.1599.

On November 2, it opened at 0.1606 and reached a high of 0.1644, which was 0.0045 above his entry price, or 2.81% higher. It closed at 0.1544, 0.0055 below his entry price, or 3.44% lower. This was the first day after his purchase to close below the previous day's close. Under the first rule, he exited that day. The day's low was 0.1511, still above the hammer's low of 0.1394. If he had instead waited to exit until a close below 0.1394, it closed at 0.1332 on November 8, 0.0267 below his entry price, or 16.70% lower. If he had held for the full 5 trading days, it closed at 0.1414 on November 5, 0.0185 below his entry price, or 11.57% lower; that day's low fell to 0.1387, already below 0.1394, but it hadn't closed below that level, so under the second rule he still wouldn't have exited that day. If he had held all the way to November 9, it closed at 0.1353, 0.0246 below his entry price, or 15.38% lower; in between, on November 8, it had fallen as low as 0.1287, 0.0312 below his entry price, or 19.51% lower.

With the same rule, after the first purchase, exiting the next day meant a loss of just 0.10%; holding until a close below the hammer's low turned that into a 10.05% loss. After the second purchase, exiting early meant a 3.44% loss; holding longer instead led to a 16.70% loss, and even holding for 5 days meant an 11.57% loss. When you encounter a day like this, where the price closes down soon after you buy, would you still switch to waiting for a close below the low before exiting?