Author: Jake Pahor
Translated by: TechFlow
Bitcoin moved higher while many traders were waiting for a September low. That happened even as several negative headlines hit the market at once: the Federal Reserve raised rates early Thursday Beijing time for the first time since 2023, the Bank of Japan lifted rates to their highest level in 31 years the next day, the U.S. Senate failed by one vote to advance the CLARITY Act, and spot Bitcoin ETFs posted their largest single-day outflow since June.
Even so, Bitcoin rose more than 5% on Friday Beijing time, wiped out about $190 million in short positions within an hour, and by Sunday had climbed back above its 50-week moving average for the first time since last November.
Pahor went back through daily scoring history and weekly Bitcoin closes from 2012 onward. He pulled every case in which Bitcoin had spent at least a month below the 50-week moving average and then posted its first weekly close back above that line. He found six such cases in total. Four marked the start of bull markets. Two turned out to be traps.
Current CSH dashboard reading
The CSH score closed the week at 37.4, up from 36.8 the previous Sunday. According to the article, that is lower than 65% of all days in the historical sample.
Bitcoin was at $81,062 on Sunday, or about A$113,800, and was up 4.9% for the week. On TradingView’s Bitstamp weekly chart, the 50-week moving average sat near $78,800, leaving Bitcoin roughly 3% above it. The weekly candle closes at 10:00 a.m. Beijing time on Monday.
Pahor places support in the bull market support band around $70,200 to $72,600. He also says a CSH score of 30 currently maps to about $69,700, putting those two reference zones close together again.
His plan is still to wait, but with a wider net. He has expanded his buy range from a CSH score of 10 to 30 to a new range of 10 to 35. A score of 35 now corresponds to roughly $76,800. He says he has not placed an order since August.
That change means the plan is now 4 points, or about $4,000, away from another buy. If Bitcoin revisits the support area many traders are watching, the score would be around 30 and the plan would buy on Monday. If that pullback never comes, the cash he had reserved for a September low will enter on schedule once the score allows it. He does not call the bear market over on one close. Consecutive weekly closes above the 50-week moving average are the real signal, he says, and that is the real test. Still, he adds that this week he stopped betting on lower lows.
Six reclaims of the 50-week average, four real bottoms
Pahor says he still remembers the feeling from March 2013. Bitcoin had just posted a weekly close back above the 50-week moving average at about $27,000, and he was frustrated with himself. His self-managed pension had missed the November 2022 low at $15,700 because of paperwork delays, and price was already 70% higher. He thought he was late. Bitcoin then went on to rise nearly fivefold.
That, he argues, is the trap built into this signal. It always comes after a rebound, so it almost always feels late. That is why he wanted to know what it had actually meant in past cycles.
His setup is simple. He looked at weekly closes from 2012 onward. A reclaim means Bitcoin spent at least one month below the 50-week moving average and then posted its first weekly close back above it. He then measured what happened next.
- October 2015: CSH score 19.5. Bitcoin stayed above the line for 134 weeks and was up 2.3x one year later.
- May 2019: score 47.5. It stayed above for 31 weeks and was up 1.5x after one year.
- January 2020: score 39.9. Bitcoin fell back below the line within 8 weeks, and the COVID crash later pushed price 35% below the reclaim level. A trap.
- May 2020: score 39.6. It held above for 62 weeks and was up 6.3x one year later.
- April 2022: score 65.4. Bitcoin fell back below the line in the second week and then dropped another 58%. A trap, and the one most people remember best.
- March 2023: score 41.9. It stayed above for 137 weeks, until last November, and was up 2.5x one year later.
In his reading, four were real and two were traps. The real signals stayed above the moving average for at least 31 weeks and, over the next six months, never closed more than 7% below the reclaim close. The traps failed within two months.
The candle itself tells less than many think
Pahor says the reclaim candle on its own is not the key separator.
The March 2023 reclaim was powerful, with a 32% weekly gain and a close 17% above the 50-week average. But the May 2020 signal that later led into a 6.3x move closed only 2% above the line. The 2022 trap closed just 0.5% above it. A bigger breakout helps, he says, but it is not the main difference between a real signal and a failed one.
He points to two factors instead.
First, how long Bitcoin had spent below the 50-week moving average. Before the real signals, Bitcoin had usually been below the line for about a year: 62 weeks, 49 weeks, and 49 weeks. The COVID case lasted only 7 weeks, but he treats that as a crash and V-shaped reversal rather than a standard bear market. The traps, by contrast, came after only 4 weeks and 13 weeks below the line. In his view, a market that has spent a year below the 50-week average has gone through enough time-based washout. He notes that the last three bear markets took 13 months, 12 months, and 12 months from top to bottom, and the reclaim usually came four to nine months after the low. If the market has been below the line for only a month, that process has not really played out.
Second, the CSH score on the day of the reclaim. Every real signal came with a score between 19.5 and 47.5, which he interprets as Bitcoin still being historically cheap when it turned higher. The 2022 trap came at 65.4, a historically expensive reading. He describes that move as a bear-market rally, and says the score had already flagged it.
Bear-market depth is less useful in his framework. The three major bear markets fell 77% to 86% from top to bottom. Two shallower declines, in 2019 and during COVID, were around 50%, and one of those reclaims was real while the other was a trap.
How the current setup compares
Bitcoin has now spent 45 weeks below the 50-week moving average, and the CSH score stands at 37.4. On the two factors he considers most important, Pahor says this setup looks like a real one.
The low was 53% below the October top, which fits the profile of a shallower bear market. This week’s candle was also fairly modest: Bitcoin rose 5% on the week and closed 3% above the moving average. He notes that traps can look like this too. So did May 2020.
One number is what made him stop waiting. In every real signal, Bitcoin had already risen 65% to 80% off the low by the time the reclaim happened. On the day itself, each one felt late. One year later, those cases were up 1.5x to 6.3x. This week, by comparison, Bitcoin is only 38% above the July low.
His plan is straightforward. If Bitcoin posts a weekly close back below the 50-week moving average within the next eight weeks, he will treat this as a trap. Both historical traps failed inside that window. If that happens, the plan will keep buying below a CSH score of 35, and he says he will stop talking about a bull market. If Bitcoin can keep closing above the line, he will say publicly that the bear market is over, even if that means his plan misses part of the move. He says that trade-off has always been part of his approach.
What he is no longer willing to do is sit on a pile of cash waiting for a price that history says usually does not come back. Even at $81,000, the CSH score of 37.4 still says Bitcoin is historically cheap, lower than 65% of days since 2013. As he puts it, you do not need to buy the exact bottom to make good money.
Jake’s workbench: rule changes, personal trades, and dates to watch
Pahor lists three items for the week.
First, he changed the plan. The old rule bought Bitcoin when the CSH score fell below 30. Since Aug. 22, the score has stayed between 30 and 39, so the plan did nothing while Bitcoin rose from $78,000 to $81,000. He then went back through the score history and tested which buy threshold ended up holding the most Bitcoin while keeping the sell rule unchanged, using three different starting points: 2017, 2019, and the November 2021 top. In each case, buying below 35 came out ahead. Buying below 30 made the plan wait too long.
Since the start of 2023, there have been 134 days with a score below 30 and 254 days below 35. With the same amount of money, that nearly doubles the number of buying days. On the sell side, starting to sell at 80 ranked first or tied for first in every test window, so that rule stays in place. The range is now 10 to 35. He says the rule change took 20 seconds, but it took a month of inactivity for him to realize it needed changing.
Second, he looked at his own trades. The rebuilt Today page plots every buy against the score on that day. The buys he feels best about mostly sit in the 20 to 30 range. The points from 2024 and 2025 that landed in the 70 to 80 range show where he should have taken profit instead of adding more. He built the tool himself, he says, and it still stung. That is exactly why it is useful.
Third, he highlights two dates. On Sept. 6, he wrote down two conditions for changing the rule on idle cash: 12 full months from the Oct. 7, 2025 top, and a weekly close above the 50-week moving average that holds. The first condition arrives 17 days later. The second is being tested now. If both are met, idle cash will stop waiting for lower lows and start entering according to the schedule.
In the product note, he says the app rebuilt the Today page this week. It now shows a one-line summary of the plan and the next order, plots every imported trade on the score, and adds a trade-file upload feature that supports export files from major Australian and global exchanges.
Macro events and the near-term calendar
The Federal Reserve raised rates by 25 basis points on Thursday morning local time, taking the policy rate to 3.75% to 4.00% in a 12-0 vote. It was the first hike since 2023. One more hike in December is already priced in, according to the article. Bitcoin made its weekly low before the decision and rose during the announcement window.
The Bank of Japan raised rates to 1.25% on Friday, the highest level in 31 years. Pahor says the market reaction told the same story.
On regulation, the CLARITY Act, a U.S. crypto market structure bill, failed to clear a procedural Senate vote on Tuesday by 49 to 50. Bitcoin fell below $75,000 that day and was back above $81,000 by Friday.
For ETF flows, spot Bitcoin ETFs saw $450 million in outflows on the day of the Senate vote, the largest single-day outflow since June. Two days later, they recorded $159 million in inflows.
He then lists the dates ahead:
- Monday, Sept. 21, 10:00 a.m. AEST: weekly close. A close above $78,800 would confirm the reclaim of the 50-week moving average, though the following week’s close matters more.
- Thursday, Sept. 24, 11:30 a.m.: Australian employment data from the ABS.
- Friday, Sept. 25, 6:00 p.m. AEST: Deribit quarterly options expiry, with about $14.7 billion in Bitcoin options and peak open interest near $72,000. He says that is another reason support could be tested.
- Tuesday, Sept. 29: Reserve Bank of Australia rate decision.
- Wednesday, Sept. 30, 11:30 a.m.: Australian monthly inflation data from the ABS.
- Thursday, Oct. 1, 10:30 p.m. AEST: U.S. inflation data, PCE.
- Friday, Oct. 2, 10:30 p.m. AEST: U.S. jobs report.
- Wednesday, Oct. 7: 12 full months from the top.
Whatever those numbers do, he says the plan stays the same: buy below 35, sell above 80, and wait in between.
His closing point is practical. If the market proves you wrong this month, do not fight it. The early stage of a bull market and the late stage of a bear market are often when the biggest risk needs to be taken, and it rarely feels comfortable in real time. For people with jobs, families, and no time to watch the screen all day, the workable answer is to set the rules now and let the plan make the decisions.