Imagine someone named Xiaoyou. He used to chase rallies, but later came up with a rule for himself. If $UNI falls at least 4% from open to close one day, then closes with a bullish candle the next day (closing above the open), he buys at the next day’s close, holds for 4 days, and sells at the close of the fourth day. While he’s holding a position, including on the day he sells, even if the same pattern appears again, he doesn’t open a new position.

I applied this rule to completed daily candles for Binance spot UNI/USDT, from August 17 to November 14 (UTC, 90 candles in total). The November 15 candle was still incomplete as of the morning of November 15 Beijing time, so it wasn’t included.

The pattern appeared on 8 days during those 90 days, but Xiaoyou actually made only 6 trades according to the rule. When it appeared on October 13 and November 9, he was still holding his previous position or it was the day he was due to sell, so he didn’t open a new one.

The first trade was a big one. On August 28, the price fell from 4.685 to 4.437, a drop of 5.29%. The next day, August 29, it closed with a bullish candle, so he bought at 4.682. Four days later, on September 2, it closed at 5.854, earning him 25.03%. Xiaoyou told his friends, “A bullish candle after a big bearish one—it actually works pretty well.”

Things changed while he was out of the market. After selling on September 2, he didn’t meet the entry conditions again until September 29—a gap of 27 days. On September 23, $UNI touched its highest price of the 90-day period, 10.942 intraday, and closed at 9.242. It had been 21 days since he sold; the intraday high was 86.91% above his selling price, and even the closing price was 57.87% higher.

He missed that move. On September 29, he bought back in at 8.889 according to the rule. Four days later, on October 3, it closed at 9.052, for a gain of just 1.83%. The next four trades never came close to the first trade’s return:
Bought at 7.322 on October 9 and sold at 7.541 on October 13, gaining 2.99%;
Bought at 7.431 on October 15 and sold at 7.318 on October 19, losing 1.52%;
Bought at 7.161 on October 30 and sold at 7.22 on November 3, gaining 0.82%;
Bought at 6.889 on November 6 and sold at 6.671 on November 10, losing 3.16%—the biggest loss.

Across all 6 trades, 4 were profitable and 2 were losses, for a compounded return of 26.08%. The first 3 trades were all profitable, with a compounded return of 31.13%; the last 3 had a combined compounded return of -3.85%. If he had bought at the August 17 close of 3.286 and held until the November 14 close of 4.888, the return would have been 48.75%—considerably more than Xiaoyou’s rule made. He didn’t stay in for that big rally. Over the same period, $BTC went from 64532.1 to 64150, a change of -0.59%, essentially flat.

The lowest price during these 90 days came on August 18, at 3.227. The November 14 close was 55.33% below the September 23 high.

Xiaoyou made 25.03% on his first trade, but later missed the biggest part of the rally by selling too soon. There were only 6 trades in 90 days, and another period might not look the same.

These are historical data from completed candles. They don’t indicate what will happen next and are not trading advice. If you also wait for a bullish reversal after a big bearish candle, would you care about the gains you missed after selling?