Bitcoin Cycle Models: Checking Them Against Reality

Overview

Bitcoin cycle models produce confident, specific price targets, and their published record is poor enough that the targets should be read as illustrations of a method rather than as forecasts.

The best-known example is the stock-to-flow model. It projected $98,000 for November 2021 and $135,000 for December 2021. Bitcoin closed that December around $47,000, then fell to $15,500 during 2022 — more than 80% below what the model projected for the period.

The same model's projection for 2025-2026 rose to roughly $500,000. Bitcoin peaked at about $126,000 in October 2025 and corrected to the $65,000 to $75,000 range in early 2026.

That last figure matters most for anyone who acted on these models. Price levels widely described as mathematical floors — including levels near $96,000 — were broken decisively. A floor that gets traded through is not a floor, and treating one as a guaranteed buying opportunity is how model-following produces real losses.

A Model Predicted $135,000; Bitcoin Closed at $47,000

The stock-to-flow model attracted an enormous following by appearing to explain Bitcoin's price through its supply schedule, and it produced dated, specific projections that can be checked.

For November 2021 it projected $98,000. For December 2021 it projected $135,000. Bitcoin closed December around $47,000 — roughly a third of the projected level.

The specificity is what makes the failure instructive. This was not a vague directional claim that could be reinterpreted afterwards. It was a numbered target for a named month, and the outcome was off by nearly $90,000.

Then It Fell to $15,500

The following year the gap widened dramatically. Bitcoin fell to $15,500 in 2022, more than 80% below what the model had projected for that period.

An 80% miss is not a model requiring calibration. It is a model whose central relationship did not hold, and the direction of the error matters as much as its size — the projection was high while the market fell, meaning anyone using it to decide when to buy was buying into a decline the model said could not occur.

This is the period that determines whether a forecasting method is useful. Models are easy to believe during rises, when almost anything bullish appears vindicated. What separates a model from a narrative is whether it holds when the market turns.

The Current Projection Is $500,000

Despite that record, the same model's average projection for 2025-2026 rose to around $500,000.

Bitcoin peaked at approximately $126,000 in October 2025 — about a quarter of the projected level — before correcting substantially.

The pattern is consistent across cycles: the model projects a figure, the market reaches a fraction of it, and the projection is subsequently revised upward for the next period rather than the method being abandoned. A forecast that is reissued at a higher number after each miss is not being tested by outcomes.

The Floor That Wasn't

The most consequential claim these models make is not the upside target but the floor — a level described as mathematically supported, where buying is presented as low-risk.

Levels around $96,000 were characterised in this way for 2026, with dips toward them framed as generational buying opportunities. Bitcoin traded in the $65,000 to $75,000 range in early 2026.

Someone who accepted that floor and bought near it would be down roughly 25 to 30%, having taken the position specifically because a model told them the downside was mathematically bounded. Upside targets that fail cost you an opportunity; floors that fail cost you money, which is why the floor claims are the genuinely dangerous half of these frameworks.

Why These Models Are Auto-Correlative

The technical criticism is more damaging than any individual miss, because it explains why the fit looked so convincing in the first place.

The models are auto-correlative: the prediction line is shaped by the same historical prices it claims to be predicting. When both the input and the output derive from the same price series, a close historical fit is close to guaranteed and demonstrates nothing about predictive power.

Critics have also noted that the approach ignores demand entirely — treating price as a function of supply issuance alone, which violates basic economics — and that regressing non-stationary variables against each other risks producing spurious relationships that look statistically strong and mean nothing.

Why They Survive Being Wrong

A reasonable question is how models with this record retain followings, and the answer is structural rather than psychological.

These models generally have wide bands of acceptable values rather than single lines. A projection expressed as a range spanning a factor of two or three can accommodate almost any outcome, and when the midpoint is missed the price usually remains somewhere inside the band — which is then presented as the model holding.

That flexibility makes them effectively unfalsifiable in practice. A framework that cannot be shown to be wrong also cannot be shown to be right, and the specific numbers extracted from it for headlines carry none of the band's ambiguity — which is how a $500,000 projection and a $126,000 outcome coexist without the model being retired.

What On-Chain Metrics Actually Measure

On-chain indicators such as MVRV Z-Score occupy a different category and deserve a fairer assessment than the price models.

These measure real properties of the network: how much of the supply is held at a profit, how long coins have been dormant, how much sits on exchanges. That is genuine data about the present state of the market, and it is not fabricated or curve-fitted.

The error is in the leap from description to prediction. Knowing that holders are heavily in profit tells you something true about current conditions and the potential for selling pressure. It does not tell you when anyone will sell, and the thresholds cited as cycle-ending signals are drawn from a handful of past instances rather than from any established relationship.

Four Halvings Is Not a Dataset

The deepest problem with cycle analysis is sample size, and it is rarely acknowledged in the material that presents it.

Bitcoin has experienced a small number of halvings — a single-digit count — and cycle theories are built by identifying patterns across them. Any claim about what typically happens a given number of days after a halving is drawn from a handful of observations.

No conclusion drawn from that many instances would be considered reliable in any other analytical field. And the surrounding conditions differed enormously between them: the arrival of institutional custody, spot exchange-traded funds, corporate treasury holdings and changed interest rate environments mean each cycle occurred in a materially different market from the last.

Why Specific Price Zones Are the Dangerous Part

The distinction that matters is between analysis and instruction. Explaining that a metric is historically elevated is analysis. Telling someone to sell above one number and buy below another is instruction, and it transfers no risk to whoever issued it.

Content presenting colour-coded price zones — take profit here, accumulate there — converts an unfalsifiable model into a specific action with real financial consequences for the reader and none for the author.

The 2026 outcome demonstrates the cost precisely. A published floor near $96,000 was broken by 25 to 30%, and anyone who bought there because the model described it as mathematically supported discovered that the support was a line on a chart.

Bitcoin may rise substantially from any level, and none of this argues about its long-term prospects. It argues that models with this record should not be the basis for deciding when to buy or sell, and that anyone publishing price zones is asking readers to take a risk they are not taking themselves.

Conclusion

The most widely followed Bitcoin price model projected $98,000 for November 2021 and $135,000 for December 2021. Bitcoin closed near $47,000 and subsequently fell to $15,500 — more than 80% below projection. Its 2025-2026 projection of roughly $500,000 met an actual peak near $126,000.

The technical objection explains the pattern. These models are auto-correlative, with the prediction line shaped by the very prices it claims to predict, and they ignore demand entirely. A convincing historical fit was guaranteed by construction and demonstrated nothing.

The floors are the part that costs money. Levels near $96,000 described as mathematically supported were broken decisively when Bitcoin traded at $65,000 to $75,000 in early 2026 — leaving anyone who treated that floor as a guaranteed entry down roughly a quarter.

This article examines the published record of forecasting models and is not investment advice. It contains no price predictions and no buy or sell recommendations. Cryptocurrency is highly volatile and carries risk of total loss; anyone considering it should seek independent financial advice.

Data Source and Attribution

Bitcoin MagazineBookmap analysisLBank (model comparison)

Model projections and their outcomes are drawn from published analyses of the stock-to-flow model, including its stated November and December 2021 targets, the subsequent 2022 low, and its 2025-2026 average projection. Actual market levels — the October 2025 peak and the early 2026 trading range — are as publicly reported. Methodological criticisms regarding auto-correlation, treatment of demand and non-stationary regression reflect published critiques of the model. No price forecast is made in this article.

FactsFigs reviews, cleans, and cross-checks every source dataset before shaping it into a data story. Each visualization is created and designed in FactsFigs Design Studio — an internal tool developed and owned by FactsFigs — and is the original work of a FactsFigs author, not an AI-generated copy of any existing graphic. Individual assets within a visual may or may not be produced with AI tools, but the design of the visual itself is solely FactsFigs' own.

This content is for information only and is not investment or financial advice. It contains no price targets or trading recommendations. Cryptocurrency carries substantial risk including total loss of capital.

2026-07-20