If you’ve been around a Bitcoin cycle or two, you’ve seen the charts: a monthly candlestick view with neat boxes counting the bars between every top and bottom, and a confident claim that the next move is therefore already on the calendar. The numbers vary — 23 bars, 47 bars, 35 and 12 — but the pitch is identical: Bitcoin runs on a clock, and if you can count, you can time it.
We have skin in this one. The bot’s strategy leans on the four-year halving cycle — it’s one of the few edges that’s actually survived our testing. So when a cycle-timing theory makes the rounds, “is it real?” isn’t idle curiosity for us. We pulled the monthly chart back to 2015, marked each cycle’s real top (the highest price actually traded) and real bottom (the lowest), and counted the months — without sliding the start line after the fact.
Theory 1: the 23-bar bottom
The first claim is tidy: a bear market runs 23 monthly bars from top to bottom, so once 23 print, the bottom is in. It’s also wrong by about half.
| Cycle | Real top | Real bottom | Bars |
|---|---|---|---|
| 2017–18 | Dec 2017 | Dec 2018 | 12 |
| 2021–22 | Nov 2021 | Nov 2022 | 12 |
Top to bottom, both times, was twelve months — not twenty-three. So where does 23 come from? You only get there by sliding the start line back to the cycle’s first peak (Bitcoin tends to double-top — an early spike, a dip, then the real high), and by calling the “bottom” the end of the following year’s accumulation rather than the actual low. Move both goalposts and 12 stretches to 23. Leave them where the price actually was, and it’s 12. The number isn’t measured; it’s chosen.
Theory 2: 47 / 35 / 12
The second version is sharper, and to its credit it mostly holds. It splits the cycle into a 35-bar climb and a 12-bar fall — 35 + 12 = 47 — and 47 monthly bars is almost exactly four years.
| Leg | Claim | What we counted |
|---|---|---|
| Bottom → top (the bull) | 35 bars | 35 — both cycles |
| Top → bottom (the bear) | 12 bars | 12 — both cycles |
| Bottom → bottom (the cycle) | 47 bars | 47 |
Those numbers are real. The 35-bar climb even pinned the most recent top to the month. Unlike the 23-bar story, this one isn’t fudged — the bars are where it says they are. Which raises the obvious question: is the clock real?
Sort of — and here’s why that’s not the win it sounds like
1. It’s just the halving. Forty-seven months is four years, and Bitcoin’s supply halves roughly every four years (2016, 2020, 2024). The “47/35/12 cycle” is the halving cycle with a tape measure laid over it. The rhythm is real; the precise bar counts are decoration on top of it.
2. It only looks this clean if you delete a crash. When we let the data pick the turning points objectively, an extra one shows up that the theory ignores: the March 2020 COVID crash — a ~73% drop, as deep as any cycle bottom — sitting right in the middle of a supposedly-smooth 35-bar climb. The neat 47 only appears once you decide, after the fact, that 2020 “doesn’t count.” That’s a choice, and it’s doing a lot of the work.
3. It’s two cycles. There are exactly two complete cycles in the modern data. You cannot confirm a three-number timing rule (35, 12, 47) on two examples — there is nothing left over to test it against. It fits because it was built to fit.
What it predicts now
Played forward, the theory makes a clean, falsifiable call: a top in late 2025 (which happened), then a bottom roughly twelve bars later — around late 2026. Here’s the tension. The timing might rhyme. The depth isn’t rhyming yet: the last two bottoms fell 73–77% from their tops, and so far this decline is around a third of that. For the pattern to truly repeat, Bitcoin would have to fall a great deal further, on schedule. Maybe it does — but “the bars say bottom in October” quietly assumes a crash that hasn’t happened.The honest conclusion
The four-year cycle is one of the few things that has survived our testing, and the bot uses it — but as a phase, not a stopwatch. “We’re late in the cycle, lean defensive” is robust to a cycle that runs 44 months or 50. “The bottom is exactly 47 bars out” is not — it snaps the moment the schedule drifts, and with two or three cycles on record, betting it won’t drift is its own act of faith. So we keep the rhythm and bin the bar counts. The clock is real enough to respect, and far too loose to set your watch by.