Stock-to-Flow, Uptober, and the Myth of the Bitcoin Calendar

Bitcoin closed October 2026 at roughly $77,430, down about 7.3% on the month. That made it the worst October in twelve years and, more striking for anyone who trades the calendar, the second red October in a row. For most of the past decade October was the month Bitcoin went up so reliably that traders gave it a nickname, “Uptober.” Two consecutive losing Octobers had not happened since 2011 and 2012, when Bitcoin was a hobbyist curiosity trading in single digits. The streak is broken.

The day after the month closed, the pseudonymous analyst PlanB, creator of Bitcoin’s stock-to-flow model, posted a short note to his followers: “Bitcoin closed October at $77,430. No Uptober unfortunately. But RSI is above 50 for the 3rd month in a row, signaling the birth of a new bull market.” Read that twice. The man whose entire public thesis is that Bitcoin’s price is governed by its programmed scarcity called the start of a new bull market using a momentum oscillator, not scarcity. He still says stock-to-flow points to $500,000. He just no longer seems to steer by it.

As this piece was written, Bitcoin traded near $80,400, a market capitalization of about $1.62 trillion on roughly 20.1 million coins in circulation, around 36% below the all-time high of $126,080 set on 6 October 2025. Stock-to-flow’s headline target of $500,000 is more than six times the current price. That gap is not new, and it is not the point of this article. The point is simpler and, for the predictions business, more useful: the past two years ran a clean experiment on a whole family of models that assume Bitcoin’s price is a function of time, and the calendar lost.

What stock-to-flow actually claims

Stock-to-flow is a scarcity ratio borrowed from commodities. Take the existing above-ground supply of an asset (the stock) and divide it by annual new production (the flow). Gold scores high, around 62, because the hoard is enormous relative to what miners add each year. Consumables like corn or wheat score near zero, because almost everything produced is used up within the year. In March 2019 PlanB applied the ratio to Bitcoin in a paper titled Modeling Bitcoin’s Value with Scarcity, when the coin traded near $4,000, and reported a near-perfect fit: a regression of market value on the stock-to-flow ratio with an R-squared around 95%.

The engine is the halving. Roughly every four years Bitcoin’s block subsidy is cut in half, so the flow of new coins drops while the stock keeps growing, and the ratio jumps. By CoinGecko’s own explainer, Bitcoin’s stock-to-flow sat near 27 before 2020, climbed to roughly 56 after the 2020 halving, and jumped to about 120 after the April 2024 halving, pushing it past gold. The model reads that rising ratio as a rising price, and the prices it produces are large.

In April 2020 PlanB extended the idea into a cross-asset version, S2FX, that plotted Bitcoin alongside gold and silver and projected an average price of $288,000 for the 2020 to 2024 cycle, implying a market value near $5.5 trillion. Today his headline figure is higher still: an average of roughly $500,000 across the 2024 to 2028 cycle, inside a band of $250,000 to $1 million. Those are the claims. The rest of this article is about the shape of the claim, not the size of the number.

Four models, one clock

Stock-to-flow does not stand alone. It belongs to a family of Bitcoin price models that share a hidden assumption: that price is, at bottom, a function of time. Each member dresses the assumption differently, but strip the costume and the skeleton is the same.

Stock-to-flow says price is a function of the halving schedule. The four-year cycle says price is a function of how many years have passed since the last halving, up for three and down in the fourth. Seasonality, the folk model behind “Uptober,” says price is a function of the calendar month. The power law, the heir most often floated now that stock-to-flow has stumbled, says price is a function of the number of days since the genesis block. Different clocks, one idea: tell me the date and I will tell you the price.

Cory Klippsten, chief executive of Swan Bitcoin, made the point years ago in a widely cited takedown of stock-to-flow: the fit is impressive because the model is, underneath, just price plotted against time. “Does not need S2F at all,” he said. “Just price vs time.” That sentence is the key that unlocks the whole family. If the real variable is the clock, then these are four versions of one model, and all four can be tested the same way, by asking whether the clock actually moved the price. From 2024 through 2026 it did not.

Uptober, quantified

Start with the softest member of the family, seasonality, because 2026 just put a clean dent in it. October really was Bitcoin’s best month for most of its traded history. On CoinGlass data compiled across 2013 through 2025, October closed green in 10 of 13 years, with an average return around 18.5% and a median near 12.7%. The best was October 2013 at roughly 61%. That record, the highest win rate and highest median of any month, is where the nickname came from.

Then came two autumns that broke it. October 2025 closed red, snapping a long winning streak. October 2026 closed red again, down about 7.3% near $77,430, which news.bitcoin.com called the worst October finish since 2014. A tighter dollar, wider credit spreads, and a major exchange hack late in the month all added to the selling. Back-to-back red Octobers had last happened in 2011 and 2012.

Here is the tell. A pattern that is right 10 times out of 13 is a frequency, not a force. Nothing in the Gregorian calendar reaches into order books and lifts bids in October. The month was a coincidence that paid off often enough to feel like a law, until the years it did not. Seasonality is the family member that most plainly confuses a correlation with a cause, which is exactly why it is useful here: it shows the flaw in miniature before we scale it up to stock-to-flow.

The halving arrived exactly on schedule. The price did the opposite.

Stock-to-flow’s defenders have a standard move when the price misses: markets are irrational in the short run, the model describes the long run, give it time. That defense works for most models because their main input is fuzzy. Stock-to-flow’s input is not fuzzy at all. It is the single most predictable fact in finance.

The April 2024 halving happened precisely as coded. At block 840,000 the block subsidy fell from 6.25 to 3.125 BTC, daily issuance dropped from about 900 coins to about 450, and the stock-to-flow ratio roughly doubled on schedule, to the day, with no surprise and no slippage. This is the one input stock-to-flow never has to guess. And then the price did the opposite of the forecast. Instead of climbing a staircase toward $288,000 and then $500,000, Bitcoin put in a cycle top near $126,000 in October 2025, only about 1.8 times the prior peak, then fell by roughly a third.

That is the cleanest falsification a deterministic model can suffer. When a model’s one certain input fires exactly on time and the output goes the wrong way, you cannot blame noisy data. The mechanism is what failed. The halving is real; the claim that the halving sets the price is what 2024 through 2026 disproved. A model gets no credit for predicting the sunrise if it also predicted the sun would rise in the west.

The four-year cycle breaks in public

The halving miss did not stay contained to stock-to-flow. It took down the broader four-year cycle, the rule of thumb that Bitcoin rises for three years after a halving and crashes in the fourth. The clearest obituary came from an institution, not a skeptic. In December 2025 Matt Hougan, chief investment officer of Bitwise, published a memo titled The Four-Year Cycle Is Dead. Welcome to the Ten-Year Grind. His argument: the forces that powered past cycles, the halving supply shock, interest-rate swings, and leverage booms, have weakened as spot ETFs, regulation, stablecoins, and tokenization pull Bitcoin into a slower, steadier regime. “We’re not in a four-year cycle anymore,” he wrote. “We’re in a 10-year grind upward.”

Even PlanB concedes the timing broke, acknowledging that the clean four-year pattern did not play out after the 2024 halving, which is part of why his toolkit now leans on moving averages and momentum rather than a countdown clock. You can watch the diminishing returns in the peaks themselves. Each cycle top has been a smaller multiple of the one before it: roughly $1,150 in 2013 to about $19,800 in 2017 was near 17 times; 2017 to the roughly $69,000 top of 2021 was about 3.5 times; 2021 to the $126,080 top of 2025 was only about 1.8 times. That is the inverse of what stock-to-flow predicts, where each halving is supposed to make the move larger, not smaller. Our own halving-cycle math coverage has tracked the same contraction from the supply side.

Why price versus time fits so well and predicts so badly

If these models are so fragile, why do their charts look so convincing? The answer is statistical, and it is the same answer for all four. Bitcoin has mostly risen over its history, and time only moves one way. Any quantity that climbs with time will correlate with any other quantity that climbs with time. Plot price against days, or against the ratio, or against years since the halving, and you get a tidy line, because both axes are secretly measuring the same thing: the passage of time during a long bull market.

Stock-to-flow has a sharper version of this problem, close to a tautology. As the analyst writing under the name Level39 showed in the Bitcoin Magazine study Klippsten endorsed, market value equals stock times price, while the stock-to-flow ratio is stock divided by flow. Both sides of the regression carry the stock term, so the model is partly correlating stock with itself. Adjust for that autocorrelation and the celebrated R-squared collapses toward zero. A fit that strong is a warning sign, not a badge of honor.

There is also the sample-size problem that no amount of curve-fitting can fix. The entire theory rests on four halvings. Four data points cannot support a law of nature, and because each halving arrives only once every four years, a genuine out-of-sample test takes a decade to run. Statisticians who examined the series found the deeper issue: price and the ratio are non-stationary trending series, the classic recipe for a spurious regression, two things that rise together without one causing the other. Vitalik Buterin, Ethereum’s co-founder, put the cost of ignoring this plainly in June 2022, as Bitcoin first broke the model’s lower band. Financial models that give people “a false sense of certainty and predestination that number-will-go-up,” he argued, “are harmful and deserve all the mockery they get.” The danger is not that a chart is wrong. It is that a confident chart leads people to size positions as if the future were already scheduled.

The chameleon that keeps changing color

A good model makes a falsifiable prediction and then lives or dies by it. Stock-to-flow has a habit of changing its number whenever the old one is missed. Nico Cordeiro of the research firm Strix Leviathan named this in June 2020, calling stock-to-flow a “chameleon model” and judging it “about as successful at forecasting Bitcoin’s price as the astrological models of the past.” Watch the headline figure migrate. The original 2019 model implied roughly $55,000 for its cycle; S2FX in 2020 raised the target to $288,000; the current framing is $500,000 across 2024 to 2028. Each revision arrived after the prior one looked unreachable, and each was fitted to the data already in hand.

The most concrete miss was the 2021 floor model, a set of monthly price floors PlanB published as a near-term test. August, September, and October 2021 came in at or above the floors, and he took a victory lap. Then November called for a floor of $98,000 and December for $135,000; Bitcoin closed November around $57,000 and kept falling. PlanB acknowledged the miss, calling it the first in ten years while insisting the longer stock-to-flow thesis was intact. By June 2022 the price had broken clean through the model’s lower band, and that November the FTX collapse took it under $16,000. A model you can always re-fit after the fact is not making predictions; it is drawing a line through yesterday and calling it tomorrow.

The blind spot: demand, not the date

Every model in this family shares one structural hole: they describe supply, or time, and say nothing about demand. Stock-to-flow knows how many coins exist and how fast new ones arrive. It does not know whether anyone wants them. And 2024 through 2026 was, above all, a story about demand, the one variable the calendar cannot see. Consider what actually moved Bitcoin across the cycle versus what stock-to-flow is built to notice.

The launch of US spot ETFs in January 2024 rewired who the marginal Bitcoin buyer is. That buyer is now a financial advisor allocating a client’s portfolio through a brokerage, not a cypherpunk counting coins against a halving clock, and the flows into and out of those funds have driven price far more than any supply ratio. Our coverage of the ETF era has tracked how completely that changed the demand side; a model that cannot represent an ETF inflow is a model missing the main event. The same goes for Bitcoin’s real adoption story, which in 2026 runs through payment rails and dollar settlement as much as through a scarcity countdown, as our reporting on Taproot dollar rails describes. None of that demand appears anywhere in a stock-to-flow chart, which is why the chart kept pointing up while the price went down.

The power law: better geometry, same clock

When stock-to-flow failed, many of its followers did not abandon time-determinism; they upgraded to a smoother version of it. The power law, popularized by the physicist Giovanni Santostasi, fits Bitcoin’s price to the number of days since the genesis block raised to an exponent of roughly 5.8, drawn as a straight corridor on a log-log chart. It genuinely fits the history better than stock-to-flow, and its backers present it as something like a natural law, comparing it to the scaling rules that govern cities and organisms.

But it fails the same two tests. First, it is still price versus time, so Klippsten’s objection applies without modification: a better geometry is not a different epistemology. Second, it does not model demand. A CryptoSlate analysis notes that the power law projects six figures after 2028 and seven figures later in the decade, yet it “does not mechanistically incorporate ETF demand or liquidity cycles,” the very forces that set the 2024 to 2026 price. And because its projected ranges span years and hundreds of thousands of dollars, it is nearly impossible to falsify on any horizon a trader could use. Being unkillable is not the same as being right. More often it means a claim has been stated too loosely to ever be wrong.

Gold, the counter-example the scarcity camp cites against itself

The scarcity thesis points to gold as its proof of concept: the hardest money, the highest stock-to-flow, a store of value for millennia. But gold is the counter-example, not the confirmation. Gold has the highest stock-to-flow of any real commodity, around 62, and yet nobody prices gold off that ratio. Gold’s price moves with real interest rates, central-bank buying, and the dollar. Its scarcity is a near-constant backdrop; the price swings come from demand.

That is the whole argument in one object. If the single most-cited high-scarcity asset in history has a price set by demand rather than by its stock-to-flow ratio, then a high ratio is at most a necessary condition for a store of value, never a sufficient one for a price forecast. Scarcity can make an asset investable. It cannot tell you what investors will pay. Stock-to-flow quietly assumes the opposite, and gold, the model’s own favorite example, is the standing refutation sitting in plain sight.

Even the model’s author steers by other instruments

The most telling evidence against calendar determinism is how its own champion behaves. PlanB still reaffirms the $500,000 figure, but his day-to-day market reads in 2025 and 2026 lean almost entirely on price-versus-time technicals and, increasingly, on mining economics. He treats the 200-week and shorter moving averages plus realized price as the bear-market floor, and uses the monthly RSI as a regime filter, which is exactly the call he made on 1 November when he declared a new bull market because RSI had held above 50 for three months. Not one of those tools is stock-to-flow.

He has gone further. In a September 2026 note he argued that mining difficulty “better reflects market price than models such as S2F or power laws,” effectively ranking the production side of the network above both of his scarcity models. There is a serious idea underneath: the all-in cost to mine a coin behaves like a soft price floor, because unprofitable miners eventually switch off and the periodic difficulty adjustment resets the economics for those who remain. We have examined that dynamic in detail in our work on mining margins. The point for this article is narrower. When the creator of stock-to-flow reaches past stock-to-flow for his real-time signals, the author is telling you how much weight the model can bear, and the answer is not much.

What actually moved Bitcoin in 2026

If the calendar did not set the price this year, what did? Policy and liquidity, mostly. The Federal Reserve under Kevin Warsh raised rates for the first time since 2023 in September 2026, then held the target range at 3.75% to 4.00% on 28 October in a 10 to 2 vote, with two officials dissenting in favor of another hike. Markets now treat a December increase as the base case. A tighter dollar and wider credit spreads, not a supply ratio, framed the autumn selloff, and those are the conditions a scarcity model is structurally unable to represent.

The sharpest single reminder came a year earlier. On 10 October 2025, at the very hinge of the cycle, a cascade of forced selling erased around $19 billion in leveraged positions in a day, the largest such event on record, and dragged Bitcoin from about $122,574 to $104,782 in hours. Nothing about Bitcoin’s scarcity changed that day. The supply schedule was identical at the close as at the open; only leverage, liquidity, and fear moved. A model that cannot see the biggest single move of the cycle, because that move had nothing to do with the number of coins, is not a map of the territory. It is a map of a different country.

How to read any Bitcoin price model

None of this means price models are useless, or that Bitcoin cannot reach high numbers. It means something narrower: a model that maps time to price will fit the past and miss the future, and you should grade any forecast on whether it escapes that trap. Four questions do most of the work.

• Does it model demand, or only supply and time? If it cannot represent an ETF inflow or a liquidity shock, it cannot explain the last two years.
• Is it falsifiable on a horizon you can actually use? A target with a wide band and no deadline can never be wrong, which means it can never be useful.
• Does it escape price versus time? If the real independent variable is the clock, the fit is a mirage.
• Would it have beaten simply holding Bitcoin, out of sample? A model that never issues a sell is just buy-and-hold with extra steps.

That last point matters for anyone tempted to trade a forecast. The US Securities and Exchange Commission approved spot Bitcoin ETFs in January 2024 and, in the same breath, kept repeating the oldest line in investor protection: past performance does not guarantee future results. A stock-to-flow chart is past performance drawn forward with a ruler. Treat the $500,000 figure the way you would treat any extrapolation, as a story about scarcity that may or may not come true, not as a schedule the network is obligated to keep.

The halving is real and the supply is genuinely fixed. Those facts are the foundation of the entire asset. What 2026 settled, with a second red October and a bull-market call built on momentum rather than scarcity, is that they are the start of the analysis and not the end of it. The calendar tells you when the coins arrive. It has never told you what they are worth.

Frequently Asked Questions

Did Bitcoin’s stock-to-flow model fail?

By its own terms, yes. The model’s one certain input, the April 2024 halving, arrived exactly on schedule, and the price then did the opposite of the forecast, topping near $126,000 in October 2025 before falling about a third. Its headline $500,000 target is more than six times the roughly $80,000 price of early November 2026. Critics including Vitalik Buterin argue such models give a false sense of certainty, and even the model’s creator now reads the market with moving averages and momentum rather than scarcity.

What is “Uptober” and did it happen in 2026?

“Uptober” is the nickname for Bitcoin’s tendency to rise in October, which closed green in 10 of 13 years from 2013 through 2025 with an average gain near 18.5%. It did not happen in 2026. Bitcoin fell about 7.3% to close near $77,430, the second red October in a row and the worst October since 2014, breaking the seasonal pattern.

Is the Bitcoin four-year cycle dead?

A growing number of analysts think so. Bitwise chief investment officer Matt Hougan argued in a December 2025 memo that the four-year cycle is dead and that Bitcoin has entered a steadier ten-year grind, as ETFs and institutional demand smooth out the old halving-driven boom and bust. Each cycle top has also been a smaller multiple of the last, the opposite of what a strengthening cycle would produce.

What is a better way to value Bitcoin than stock-to-flow?

There is no single clean answer, but the more honest tools describe conditions rather than promise a price. Realized price, MVRV, and mining production cost act as risk and floor gauges, and demand-side drivers such as ETF flows, Federal Reserve policy, and dollar liquidity have explained price far better than any supply ratio since 2024. The key is to prefer models that can represent demand and that can be proven wrong on a useful horizon.

Does PlanB still predict $500,000 for Bitcoin?

Yes. As of late 2026 PlanB still cites an average of roughly $500,000 across the 2024 to 2028 cycle, inside a $250,000 to $1 million band, and says the October 2025 high was not the top. In practice, though, his real-time calls rely on the monthly RSI, moving averages, and mining difficulty, not on stock-to-flow, which tells you how much confidence to place in the headline figure.

Priya Reddy covers Bitcoin markets, mining, and on-chain research for HOGE Wire.