Basic statistical flaws of bitcoin’s four-year price ‘cycle’

Bitcoin (BTC) price predictions from believers in its supposed four-year price cycle were so inaccurate that many have started joking about a five-year cycle.

At least a five-year cycle, as the joke goes, could offer some hope for a higher BTC price in 2026.

The idea that BTC follows a four-year cycle at all originates from the cadence of its coinbase reward halving every four years. Because the supply of BTC issuance programmatically decreases every four years, it is easy to invent a statistical model about that halving’s supposed effect on price.

However, this ignores the reality of financial markets where millions of investors discount future prices based on all presently known information.

Indeed, the halving is always known in advance and never comes as a surprise. Therefore, investors can model out the supply of BTC for hundreds of years.

Just as there’s no sustainable way to make money trading “cycles” of quarterly earnings, annual tax filings, or seasonal harvests — because these cycles are broadly known and continually discounted in advance on a daily basis — the halving is simply part of a set of knowledge from which investors make decisions every day, not every four years.

The statistical shortcomings of the four-year cycle

BTC only has a tiny bit of history on which to base any claims of repetition. Almost all cycle proponents implicitly treat its four, four-year periods since 2009 as robust evidence of repetition.

However, with such a tiny number of repetitions, there’s no meaningful way to distinguish random chance from a genuine pattern.

Also, cycle theory suffers from a statistical error called the multiple testing problem. In statistical fields like genomics where researchers might run 10,000 separate hypothesis tests on a large data set, dozens or hundreds of results might exceed their standard alpha level of 5% and appear to be statistically significant. 

However, treating these outliers as compelling evidence ignores the responsibility of every statistician: p-value adjustment.

Once a statistician adjusts p-values to account for how many hypothesis tests occurred, that evidence of statistical significance usually disappears.

In the same way, backtesting a numerous variety of time periods on BTC’s price will certainly yield statistically significant “cycles.” This is merely the law of large numbers.

That one time period correlates with BTC prices, however, isn’t evidence of its predictive power. This is the multiple testing problem.

Read more: What’s PlanC? Bitcoin investor PlanB’s stock-to-flow price model has failed

Survivorship bias, non-stationarity, and the base rate

Survivorship bias also runs rampant among BTC investors. When the four-year cycle was “working,” proponents like Plan B’s Stock-to-Flow and other technical analysts gained immense fame.

Eventually, of course, their price predictions failed and cleared the way for other dubious models.

Survivorship bias is the human tendency to focus on success while ignoring losses. The reality, as 2025 has proven, is that the four-year “cycle” isn’t doing well at predicting the price of BTC.

In addition, cycle theory suffers from non-stationarity. Non-stationarity in a time series is where statistical properties, such as mean and variance, change over time. 

Fans of cycle theory often treat BTC’s return-generating process as if it maintains the same structural rules in response to halvings.

However, new liquidity, regulations, macro adoption, mining practices, and market participation have changed dramatically since 2009. Any pattern from BTC’s tiny, early‑stage, low‑liquidity regime is unlikely to generalize to the highly financialized, modern regime. 

In statistical terms, shifts in a market environment can terminate the predictive power of any model based on old parameters.

Cycle theory also usually ignores base rate changes. Extremely high volatility and large speculative booms are common among small, thinly traded assets.

Just because BTC was highly volatile in the past with a few four-year periods that people cherry-picked as a frame for historical rallies, its base rate explains why these outsized returns aren’t indicative of future returns.

A proper statistical approach starts from the base volatility of the asset and asks whether BTC’s pattern is unusual relative to that baseline. Most cycle theorists don’t even attempt this.

Beautiful, non-falsifiable curves

Finally, cycle theory is curve fitting. Most visual arguments for the four-year cycle rely on stylized, visually appealing, log‑price charts with hand‑drawn cycle bands, smoothed curves, or fitted bands. This is curve fitting disguised as simplicity.

With enough free choices — log scale versus linear scale, arbitrary start dates, trend line slopings, etc. — almost any noisy, upward‑drifting series can be made to appear cyclical.

Instead of sticking with the predictions of four-year cycle theorists from prior years, almost all BTC investors continually re‑tune and modify their predictions to fit the asset’s latest price move, which is a hallmark behavior of curve fitting.

Curve-fitting also introduces another statistical failure of cycle theory: Non-falsifiability. Robust hypotheses should have clear falsification criteria. In practice, four-year cycle narratives are extraordinarily squishy.

Technical analysts routinely revise price targets, or modify time windows. Statistically, if the four-year hypothesis cannot be falsified by any pre-determined path of future prices, it’s functionally meaningless as a predictive model.

Got a tip? Send us an email securely via Protos Leaks. For more informed news, follow us on X, Bluesky, and Google News, or subscribe to our YouTube channel.