Suppose there’s a guy named Xiao Ke who repairs surveillance cameras for shops at the electronics market in the south of town. On August 17, $IOTX opened at 0.00258. Old Zhou, who often came by for parts, said he’d bought some and was just holding it, and not to chase it even if it went up over the weekend. Xiao Ke set himself a different rule: whenever the UTC Saturday daily candle closed above Friday’s close, he’d buy 1,000 USDT worth at Saturday’s close, then count forward three more daily candles that had closed—which meant selling at Tuesday’s close. If he hadn’t sold his previous position yet, he’d skip any new signal.

I applied this rule to Binance spot IOTX/USDT daily candles that had closed, from August 17 to November 14 (UTC, 90 candles in all). The November 15 candle hadn’t closed yet, so I left it out. During this period there were 13 UTC Saturdays and 13 UTC Fridays, with a Friday available for comparison before each Saturday. Saturday’s close was higher than Friday’s on 5 days and lower on 8; there were no ties.

None of the 5 signals occurred while the previous position was still open, and every sell date’s candle had closed. In practice, that meant 5 trades: 3 up and 2 down. No two trades overlapped on the same day, and he never bought back in the day after selling. After putting 1,000 USDT through 5 trades, he ended up with 1,264.02, for a compounded return of +26.40% and an average return of +5.06% per trade. The first 3 trades were 2 up and 1 down, compounding to +16.89%, with the account reaching 1,168.92. The last 2 trades were 1 up and 1 down, compounding to +8.14%. The best trade was buying at the September 19 close of 0.003277 and selling at the September 22 close of 0.003862, for +17.85%. The worst was buying at the September 12 close of 0.002948 and selling at the September 15 close of 0.002909, for -1.32%. Excluding the best trade, the remaining 4 trades were 2 up and 2 down, and still compounded to +7.26%.

If Old Zhou had bought at the open on August 17 and held until the November 14 close of 0.002426, his 1,000 USDT would have become 940.31, a return of -5.97%. Over the same period, $BTC rose +1.99% from the August 17 open to the November 14 close.

The turning point came with Xiao Ke’s time out of the market after mid-October. He sold his fifth position on October 13, with 1,264.02 in his account, and stayed out after that. On Saturday, October 17, the close was 0.003857, down 1.00% from the October 16 close of 0.003896, so there was no signal under his rule. But October 22 was a Thursday, and the price reached an intraday high of 0.005—the highest price of those 90 days. Compared with his October 13 sell price of 0.003771, that was another 32.59% higher. If Old Zhou had held from the open to that point, his account would have reached 1,937.98 (+93.80%). Xiao Ke was out of the market that day, still at 1,264.02. The highest close of those 90 days was 0.004227 on October 21. Old Zhou’s account would have been worth 1,638.37 at that close (+63.84%), while Xiao Ke was still out of the market.

From the intraday high of 0.005 to the November 14 close, the price fell 51.48%; it would need to rise another 106.10% to get back to that high. The interval was 23 days. On each of the four Saturdays from October 24 to November 14, the close was below the previous Friday’s, so he never bought back in. November 14 was a Saturday. The intraday low was 0.002285 and the close was 0.002426, both the lowest of the 90 days. The close was down 9.71% from November 13’s close of 0.002687.

Old Zhou held all the way through and is now down 5.97%; Xiao Ke finished 5 trades up 26.40%. He missed the October surge, but also sat out the subsequent pullback. These are historical results based on closed candles; they don’t indicate what will happen next and aren’t investment advice. If you could either buy after a UTC Saturday closes higher and sell at Tuesday’s close, or hold all the way through, which would you choose?