利空叠在一起的那一周， $BTC 比特币反而拉了起来。美联储 2023 年以来第一次加息，一次 25 个基点；日本央行把利率加到 1.25%，是 31 年高位；(CLARITY 法案)在参议院没过；现货 ETF 出现 6 月以来最大单日流出。价格当天涨超 5%，一小时内大约 1.9 亿美元空单被打掉。周日收在 81062 美元，当周涨 4.9%，重新站上大约 78800 美元的 50 周均线，高出均线大约 3%，也是 2024 年 11 月以来第一次。当时 CSH 风险评分 37.4，好于历史上大约 65% 的交易日。

Since 2012, after falling below the 50-week moving average and adjusting below it for at least one month, the first time the weekly candle closes up and reclaims it—this has happened a total of 6 times. After 4 of them, a bull market followed: in October 2015 the score was 19.5, staying above the moving averages for 134 weeks, followed by a 2.3x move within a year; in May 2019 the score was 47.5, holding above the moving average for 31 weeks, followed by a 1.5x move within a year; in May 2020 the score was 39.6, holding for 62 weeks, followed by a 6.3x move within a year; in March 2023 the score was 41.9, holding for 137 weeks, followed by a 2.5x move within a year. The other 2 were bull-trap setups: in January 2020 the score was 39.9; it dropped back below the moving average within 8 weeks, and later fell 35% from the signal level. In April 2022 the score was 65.4; the very next week it lost the moving average, and the cumulative drop was 58%.

How much it rose in a single week and how far above the moving average it broke are not enough to tell what’s real. In the 2023 launch, the single-week increase was 32% with a 17% premium. In the 2020 May launch when the market produced a 6x run, the premium was only 2%. The 2022 trap had only a 0.5% premium. What can be separated are two things. Before a true signal appears, the price spends about 49 to 62 weeks adjusting below the 50-week moving average, close to a full round’s bear-market washout; traps last only 4 weeks and 13 weeks. The CSH on the day of a true signal falls within 19.5 to 47.5; the trap in 2022 was 65.4. The drawdown itself also can’t be separated cleanly: a deep bear can reach 77%–86%, a mild bear is about 50%, and within mild bears both true signals and traps have appeared.

This time matches the first two statements. The adjustment below the moving average lasted 45 weeks, close to the washing-out length of that set of true signals. The score is 37.4, which falls within the 19.5–47.5 range. Historically, when true signals appear, the price is often already 65%–80% away from the lows. After that, within one year it still tends to move 1.5x to 6.3x; in this cycle, from the July low it has only risen 38%. Two tiers of support below overlap: the core bull-market range is $70,200–$72,600, and a score of 30 corresponds to roughly $69,700.

A weekly close above it isn’t enough by itself. The next ~8 weeks are the verification window: if it falls back below the moving average, treat it as a trap and keep only buying when the score is below 35. Only after multiple consecutive weeks close above the moving average do we count the bear market as over. The buy zone mentioned in the article is also widened—from score 10–30 to 10–35. Take profit at 80+; otherwise, stay on the sidelines in between. A score of 37.4 is about 4 points away from 35, corresponding to a price of roughly $76,800. Previously, deploying spare cash was also constrained by another condition: being at least 12 months away from the top in October 2025—the article writes that as of October 7.

The most important thing afterward is to look at the next two weeks: September 24—Australia employment; September 25—about $14.7 billion in Bitcoin quarterly options expires, and $72,000 is marked as support; September 29—Australia RBA; September 30—Australian monthly inflation; October 1—US PCE; October 2—nonfarm payrolls. The rule is simple: buy in batches when the score is below 35, take profit when it’s above 80, and don’t chase based on gut feeling in between.