Imagine a person named Xiao An, who helps people scan old files into digital copies at a print shop in a county town. People tell him that once he likes an investment, he should just keep holding it. But he worries that if he does, he’ll get caught in the next pullback. So he set a rule for $AR , using 10,000 USDT as his starting capital: based on UTC daily candles, if the previous calendar week (Monday close to Sunday close) was down, he buys at the close of the following Monday and sells at Friday’s close. If the previous week was up, he stays out for the entire week. If he’s still holding at the end of the window, it’s valued at the final close.

We recalculated this using Binance spot daily candles (UTC, August 24 to November 21, 2026, 90 closed daily candles; determining whether the previous week was up or down requires closed candles from before those 90 days, the earliest of which is August 17):

August 24 was the first Monday in the window. The previous week, August 17 to 23, closed from 1.819 to 2.206, up 21.28%, so he stayed out. That week, from Monday to Friday, closed from 2.313 to 2.092, down 9.55%.

His first buy was at the August 31 close of 2.070 (the previous week fell from 2.313 to 2.045, down 11.59%). He sold at the September 4 close of 2.631, up 27.10% by Friday, with about 12,710.14 in cash. He stayed out the following week, September 7 to 11, when the price fell from Monday’s 3.015 close to Friday’s 2.616 close, down 13.23%. On September 14 he bought again at 2.823, then sold at 3.654 on September 18, up 29.44%, leaving him with about 16,451.60 in cash. Two days after he sold, on Sunday, September 20, the price closed at 4.804—the highest close in these 90 days, 31.47% above his sale price. He was already out of the market.

For the next three weeks, the previous week had closed higher each time, so he stayed out: September 21 to 25 rose just 0.11%, from 4.642 to 4.647; September 28 to October 2 fell 1.53%, from 4.316 to 4.250; and October 5 to 9 fell 9.82%, from 4.541 to 4.095. On October 12 he bought at 3.750 and sold at 3.940 on October 16, up 5.07%, with about 17,285.15 in cash. He bought again on October 26. By the October 28 close of 3.548, his position had briefly reached about 17,921.60 on paper. But his rule said to wait until Friday, and he sold at 3.229 on October 30. This trade instead lost 5.64%, leaving him with about 16,310.27 in cash.

He made three more buys in November. He bought at 3.046 on November 2 and sold at 3.199 on the 6th, gaining 5.02% and bringing his cash to about 17,129.53. He bought at 3.207 on November 9 and sold at 2.988 on the 13th, losing 6.83%. His final buy was at 2.483 on November 16, and he sold at 2.197 on November 20, losing 11.52%. Seven buys and seven sells. At the end of the window, he was holding cash: about 14,121.49, a gain of 41.21%. On November 21, the price closed at 2.266, 3.14% above his sale price the day before, but he was already out.

· If he had simply put all his money in at the August 24 close of 2.313 and held until November 21, when it closed at 2.266, his 10,000 would have dwindled to about 9,796.80, a loss of 2.03%—about 4,324.69 less than under his strategy. The deepest drawdown for someone fully invested the whole time occurred on November 20, at 54.27%; for his strategy it was 21.20%, also on November 20;
· The November 21 close was 52.83% below the highest close of 4.804, and it would need to rise 112.00% to get back there;
· Over the same period, $BTC fell from 78,992.75 to 60,215.77, down 23.77%. The same 10,000 would have become about 7,622.95. Applying the same Monday-buy, Friday-sell rule would have turned it into about 8,307.53, a loss of 16.92%.

Xiao An finished checking the figures and stared at the line for September 18. “The highest close came on the Sunday after I sold. On October 28, my position hit its highest value on paper, but by Friday I sold at a lower price. My last trade in November lost the most.”

I didn’t tell him he shouldn’t have sold on September 18, or that he shouldn’t have bought again in November. These are all historical daily candles that have already closed; the latest one is November 21, and nobody knows what happens next. In this window, he ended up about 4,324.69 ahead of someone who bought and held, but from the peak closing value of about 17,921.60 down to 14,121.49, he had already given some of it back. Change the window and both 41.21% and -2.03% will change. Without that +29.44% trade from September 14 to 18, the earlier numbers would have been different too.

Have you ever been in a situation like this: your rules make you exit on Friday, the highest close happens over the weekend, and a few days later you buy back in because the previous week closed down?