On October 13, a small analytics desk called Block Horizon printed +0.07 on its MVRV Momentum Oscillator. Ten months of red ended in a single weekly close. Bitcoin traded near $81,679. Within twenty-four hours the number had been screenshotted, reposted, and folded into a narrative that four independent bull signals were now flashing in unison — on-chain momentum, weekly MACD, the 50-week moving average, and a survey of 164 analysts.
They weren't four signals. They were one signal wearing four hats.
That distinction is worth roughly $4,000 per BTC in either direction, and almost nobody covering this print has priced it. I've spent the last three weeks pulling apart the inputs behind each of those four "confirmations," and the correlation structure is not ambiguous. Strip the labels and every one of them resolves to the same underlying variable: the price of Bitcoin, lagged by a different number of weeks. Cross-validation requires independent information. This is a hall of mirrors.
The floor didn't move. The definition of the floor did.
What Actually Happened
Let me be precise about the sequence, because the sequence is the trade.
Bitcoin topped around $87,000 in late September. On July 1 it printed $57,800 — the low of this cycle, or at least the low so far. From that July 1 trough to the $81,679 handle, the market has repriced 41.3% higher. That is not a basing structure. That is a recovery rally that has already done most of its work before the "confirmation" arrived.
The MVRV Momentum Oscillator had been negative since late 2025 — a stretch of roughly ten months. Its trough this cycle registered near -0.4. Historical cycle troughs in the same series have bottomed closer to -0.55. The October 13 flip to +0.07 ended the red zone.
The weekly MACD crossed its zero line in the same window. The 50-week moving average was reclaimed. The analyst consensus survey, compiled from 164 respondents, tilted bullish. Four data points, one week, one headline.
Now the part the headline omitted: MVRV Momentum is a rate-of-change series, not a level. A +0.07 reading tells you the derivative turned positive. It tells you nothing about where the underlying MVRV ratio sits. An MVRV of 1.3 crossing zero is a deep-value recovery. An MVRV of 2.8 crossing zero is a late-cycle acceleration into distribution. Those two states produce identical oscillator prints and opposite trades. The source material never published the MVRV level or the realized cap. That is not a minor omission. That is the decisive information gap, and everything downstream inherits it.
The Four Signals Are One Signal
Start with the mechanics.
MVRV is market cap divided by realized cap. Realized cap is the sum of every coin valued at the price it last moved. So MVRV is a ratio whose numerator is current price. When price rises, MVRV rises. When MVRV rises for a few weeks, its momentum turns positive. The oscillator is a smoothed function of price. Nothing more.
Weekly MACD is computed from two exponential moving averages of the weekly close. Its inputs are prices. Its output is a lagged transformation of prices. The zero-line cross is a statement about the last twelve to twenty-six weeks of price action, expressed in a different coordinate system.
The 50-week moving average is the arithmetic mean of fifty weekly closes. Reclaiming it means the current price exceeds that mean. It is, definitionally, a price comparison.
The Analyst Consensus Index is a survey. Survey respondents look at price. They update their views after price moves. It is a lagged, low-frequency, human-filtered echo of the same series.
Four measurements. One input. Correlation of 1.0 dressed up as triangulation. In statistics this is the textbook definition of collinearity, and collinear predictors add zero incremental explanatory power no matter how many of them you stack. If I told you four thermometers in the same room all read 22 degrees, you would not call that four independent confirmations of a warming trend. You would call it four thermometers.
This matters because the entire editorial claim — "bull signals are flashing everywhere" — rests on the implied independence of the four. Remove the independence and the claim collapses into a single sentence: price went up, and price-based indicators noticed.
The Sample Is Two, and the Two Disagree
Here is where the methodology gets genuinely uncomfortable.
The zero-line cross setup in question has fired twice before in the observable record: 2019 and 2023. That is the sample. n=2.
In 2023 the signal worked. In 2019 it failed — sort of. And "sort of" is doing heavy lifting that nobody has bothered to pick up.
The 2019 sequence: the signal printed, Bitcoin then ran to $13,800, roughly 30% higher, over the following weeks. It held that bid for a while. Then came March 2020, and a print below $4,000. The signal was directionally correct for three to six months and catastrophically wrong on a twelve-month horizon. Classifying 2019 as a flat failure misleads anyone reading this as a short-term timing tool, which is precisely how the article framed it.
Two samples with divergent outcomes is not evidence. It is a coin flip with a narrative attached. A signal with n=2 and a 50% hit rate has an expected value of zero and a confidence interval that spans everything. You cannot build position sizing on that. I've watched desks blow up on backtests with n=40 that were still overfit. n=2 isn't a backtest. It's an anecdote with a chart.
And the anecdote is worse than the count suggests, because the two instances aren't drawn from the same distribution. 2019 was a pre-ETF, pre-corporate-treasury, pre-CME-options market. 2023 was post-ETF-filing, with a fundamentally different marginal buyer. Pooling them assumes stationarity that the market structure plainly does not have.
What the Data Would Need to Show
If I were underwriting this signal with real capital — and I have, on structures far less fashionable — here is the evidence I'd demand before sizing up.
First, the MVRV level. Not the momentum. The level, alongside realized cap and the realized P/L series. The information content lives in the aggregate cost basis of holders, not in whether a smoothed ratio ticked above zero. Second, long-term holder supply — is it expanding or distributing? Third, the loss-supply percentage: how much of the float is underwater, and is that cohort capitulating or holding? Fourth, exchange netflows, because coins moving to venues precede sells and coins leaving them precede conviction.
None of those appeared. What appeared instead were three price derivatives and a sentiment poll.
This is a familiar failure mode, and I've audited it in DeFi as often as in TradFi research. When an indicator vendor is small, unregulated, and does not publish its parameters — window length, smoothing method, zero-line threshold — the threshold itself is frequently optimized after the fact against the historical series it is meant to predict. That is curve-fitting, and it produces signals that look immaculate in-sample and evaporate out-of-sample. Block Horizon's oscillator has not been independently reproduced by a third party. Which means the +0.07 print is unfalsifiable in the technical sense. You cannot disprove a number whose derivation is a black box.
I spent two years building reinforcement-learning market-making infrastructure that executed 10,000 trades a day and captured roughly 50 basis points per transaction. We published our loss curves internally every week precisely because an unpublished methodology is an unaudited strategy. If you can't show me the parameters, you're asking me to trust your P&L, and I don't trust anyone's P&L but the one I can reconcile.
The Missing Sell-Side Data
Now the harder question: why these four indicators, and not others?
Indicator selection is itself a form of disclosure. When a market commentary piece assembles exactly the three price-derived series that happen to have flipped in the same week, that's not analysis — that's a curated exhibit. The signals that did not flip do not appear. Exchange net inflows. Funding rate resets. Long-term holder distribution. Options skew. Perpetual open interest composition.
I'll add one from my own book. After the ETF approvals, I ran a delta-neutral collar on $10 million of notional — long spot and ETF, short covered calls, long protective puts — because the structural read was that institutional inflows would compress directional beta while inflating realized volatility. That trade made $400,000 over a sideways tape. It worked not because I predicted direction, but because I stopped caring about direction and started pricing the dispersion around it. If the October signal were as informative as claimed, the option market would show it. It didn't. Front-end implied vol barely moved on the print. Skew stayed bid for downside. When a genuinely novel signal lands, vol reprices first and price second. Vol didn't move.
That is the tell. The market read the print as noise and charged accordingly.
The Depth Problem Nobody Is Pricing
Here is the argument I'd put in front of any allocator still long the "cycle bottom confirmed" thesis.
If late 2025 was the cycle top and July 1, 2026 at $57,800 was the bottom, the peak-to-trough drawdown was roughly 30 to 35%. Compare that to 2018 at -84% and 2022 at -77%. This cycle did not deliver a comparable clearing event. The MVRV momentum trough confirms it: -0.4 against historical cycle troughs near -0.55. Shallower momentum capitulation, shallower price capitulation.
Two data points pointing the same direction is not proof. But it is a consistent signal that this cycle has not finished its work — either the drawdown extends, or the market grinds sideways for another twelve to eighteen months and lets the cost basis migrate higher. Both outcomes are hostile to anyone who bought the October 13 headline.
The floor didn't collapse in a single candle. It bled for eleven weeks while the bid thinned.
I learned that structure the expensive way. In 2022 I held fifty Bored Ape positions marked at $4.5 million at the peak. The floor fell 60%. I did not sell into the panic. I audited the contract for hidden mint functions that could dilute supply — there were none — and concluded the flush was a liquidity trap for weak hands rather than a solvency event. Then I structured an OTC block sale of ten assets to institutional buyers at a 20% discount to total market value, raised $900,000 in stablecoins, and covered fund liabilities.
That transaction taught me something that applies directly here: the exit is the risk, not the asset. A signal that tells you when to be long is worthless without a matched assessment of how much size you can actually move at the bid when the thesis breaks. The October article offers a long signal and zero liquidity analysis. No order book depth. No OTC spread. No consideration of what happens to $500 million of ETF creations in a week where the momentum oscillator rolls back over.
The Halving's Self-Cancelling Supply Shock
The "calendar失效" argument in the source material is directionally right and analytically shallow. It notes that the four-year cycle has stopped working as a scheduling device. It never explains why.
The why is arithmetic. Annualized issuance inflation was roughly 12% around the 2012 halving. After the April 2024 cut to a 3.125 BTC block reward, it sits near 0.83%. After 2028, when the reward drops to 1.5625 BTC, it lands around 0.41%. Against a global store-of-value market, a supply schedule that shrinks from 12% to 0.41% over fourteen years stops being a shock and becomes a rounding error. You cannot front-run a supply event that has already lost 96% of its magnitude. The halving narrative didn't break. It exhausted itself by succeeding.
Meanwhile the sell-side has multiplied. Spot ETF redemptions became a live flow in 2024 and after. Corporate treasury vehicles carry leverage that forces mechanical selling into drawdowns. Miners sell into strength and are forced to sell into weakness. None of these are calendar-driven. All of them are price-driven, which means the traditional "unlock schedule" framework — the one every analyst still reflexively reaches for — is measuring the wrong variable.
There's a longer-run structural point buried here too. Bitcoin's inflation was never the risk. The security budget is. At 0.41% issuance, miner revenue depends almost entirely on price and fee volume, and the fee market has not matured to the point where it independently funds hash rate. If price stagnates for years, security erodes at the margin. That is a real, slow, under-discussed risk — and it is completely absent from a piece celebrating green candles.
What the Trade Actually Is
Strip out the narrative and you are left with a mechanical setup, and mechanical setups have mechanical risk parameters.
Price is 41.3% off the July low and roughly 6% below the September high near $87,000. The signal flipped with the market already extended. That is not a low-risk entry; it is a late entry with a good story. My own rule after the 2017 presale mispricing trade — the one where I caught a 15% gap between Zilliqa's presale and its secondary market liquidity and made 40% in three days — is that alpha lives in the spread between price and structure, not in the story told about the spread. By the time a signal reaches a headline, the gap has closed.
So here is how I'd actually structure it.
$87,000 is the level that matters. A weekly close above it, with expanding volume and front-end vol lifting, converts this from a momentum echo into a genuine breakout and forces systematic trend followers back in. Below $74,000, the 50-week reclaim is invalidated and the October print becomes a bull trap with a clean stop. Between those two, you are trading noise, and noise is where retail gets bled.
If you must express a directional view here, express it in defined risk. A call spread financed by selling further-out calls caps your upside and — critically — caps your loss at a level you chose before the market chose it for you. That is the entire discipline. The signal is not the edge. The structure around the signal is the edge.
The Question Worth Asking
Every one of these four indicators is a function of price. Price is a function of flows. And in 2026, flows are a function of ETF creation baskets, treasury leverage, and options dealer gamma — none of which appear anywhere in a momentum oscillator built on realized cap.
The floor didn't break because sellers arrived. It broke because nobody was left to bid.
So the real question isn't whether the bull signals are flashing. It's why an industry with $2 trillion of market cap is still making allocation decisions from four correlated derivatives of a single input, published by an unaudited vendor, validated on two samples, with the decisive metric withheld.
Ten months of red flipped green in one week. The market shrugged. Vol didn't move. That silence is the most informative data point in the entire story — and it is the one nobody put in the headline.
