Suppose there’s a person named Xiao Yi. At the close on August 22, $YB closed at 0.0912. He set himself a rule: look only at daily candles that have closed. Yesterday must be a down candle, with the close below the open; if they’re equal, it doesn’t count. Today’s open must be above yesterday’s close; if they’re equal, it doesn’t count. And today must also be an up candle, with the close above the open; a tie doesn’t count. The rule does not require today’s open to be above yesterday’s high or yesterday’s open—only above yesterday’s close. If the conditions are met, he buys at the close and sells at the following day’s close. If the conditions still hold on the day he sells, he sells and then buys back in. He doesn’t add to his position while holding. His colleague Xiao Cang converted all 1000 USDT into coins at 0.0912 on the same day, planning to hold until the end. The replay below does not include fees.

Over these 90 days, there were 5 days that met the rule. Xiao Yi bought 5 times and sold 5 times, never buying back in, and ended with cash. On all 5 occasions, the open was neither above the previous day’s high nor above the previous day’s open; it was only 0.10% to 0.46% above the previous day’s close.

After sitting out 15 candles, he bought at the close on September 7: the open was 0.096, above the previous day’s close of 0.0959, and the close was 0.1025. At the close the next day, September 8, it was 0.1061, a gain of 3.51%, bringing his account to 1035.12. The September 8 close was 16.34% above the starting price, the highest close in these 90 days, and the last time a new closing high was set; there were 4 such instances in total. The day he bought, September 7, was already the third time a new closing high was set. The next day, September 9, the high reached 0.1167—the highest price in these 90 days—but he had already sold. Xiao Cang’s account stood at 1163.38 on September 8.

After sitting out 15 candles, he bought at the close on September 24: the open was 0.0879, yesterday’s close was 0.0875, and the close was 0.0899. The next day’s gain was 2.89%, bringing his account to 1065.06.

After sitting out another 24 candles, he bought at the close on October 20: the open was 0.0614, yesterday’s close was 0.0613, and the close was 0.0623. At the close the next day, October 21, it was 0.0659, a gain of 5.78%, bringing his account to 1126.60—the best of the 5 trades.

After sitting out another 12 candles, he bought at the close on November 3: the open was 0.0594, yesterday’s close was 0.0593, and the close was 0.0602. The next day’s gain was 2.49%, bringing his account to 1154.67, the strategy’s highest balance.

After sitting out another 4 candles, he made his final trade on November 9: the open was 0.045, above yesterday’s close of 0.0449, and the close was 0.0494. He bought, then lost 1.42% by the next day, bringing his account to 1138.31. It was the only losing one of the 5 closed trades, and the day of the strategy’s deepest drawdown. After sitting out another 9 candles, he stayed out until the end. On November 14, when the 90-day low close of 0.0425 occurred, as well as the day’s low of 0.0405, he had already sold. Xiao Cang’s account stood at 466.01 that same day.

Results over these 90 days (drawdown calculated using daily closes):
Xiao Cang, fully invested the whole time: 562.50 USDT, -43.75%, with a maximum drawdown of 59.94% (from the September 8 peak close of 0.1061 and account balance of 1163.38 to the November 14 close of 0.0425 and account balance of 466.01). On November 19, the close was 0.0513; it would take another 106.82% gain to return to that peak close. Over the same period, $BTC fell 19.58% from its August 22 close to its November 19 close. Both started with 1000 USDT.
Xiao Yi held overnight only on days when “yesterday was a down candle, today opened above yesterday’s close, and today closed as an up candle”: 1138.31 USDT, +13.83%, with a maximum drawdown of 1.42% (from his November 4 balance of 1154.67 to 1138.31 after selling on November 10). Of the 5 trades he sold, 4 were profitable and 1 was a loss; none broke even. The +5.78% trade contributed the most. Excluding it, the other 4 trades would still have ended at 1076.13 USDT from a starting balance of 1000 USDT, a gain of 7.61%.

This is a replay of one rule using historical Binance spot YBUSDT and BTCUSDT daily candles (UTC daily candles, closed data from 2026-08-22 to 11-19), and is not a prediction of what will happen in the future. Across these 90 days, the 5 overnight trades added up to +13.83%, with the best trade on the night of October 21; Xiao Cang reached 1163.38 on September 8, but ended with only 562.50. Would you use “open higher after a down candle, then close as an up candle” to decide whether to hold overnight?