There is a specific arithmetic error that markets commit after a crash. Prices fell, positions were liquidated, therefore leverage is gone, therefore the system is safer. It is intuitive. It is also wrong.
This week, as the industry marked the first anniversary of the October 10, 2025 cascade โ the session that vaporized roughly $19 billion in open interest in a matter of hours โ a comfortable story settled over the tape. The weak hands are gone. The market has deleveraged. The worst is behind us.
The data disagrees. And the disagreement is not marginal. It is structural.
Bitcoin did not deleverage. It re-leveraged into a thinner book. That single distinction separates the trader positioning for a durable bottom from the trader walking into the second leg of a liquidation cycle.
I have watched this movie before. In 2020, I built a short thesis on Yearn Finance's early vaults by looking past the headline APY and into the capital efficiency underneath it. The mechanism then, as now, was identical: the market mistook the disappearance of one cohort of leverage for the disappearance of leverage itself. It never works that way. Leverage doesn't retire. It migrates, and it re-prices.
Let me show you where it migrated to.
The Anniversary Nobody Wanted to Celebrate
On October 10, 2025, a US tariff threat from the Trump administration collided with an already-fragile derivatives market, and the result was a cascade. Approximately $19 billion in positions were liquidated across centralized venues in a single session โ a scale of forced closure that resets a market's psychology for months, sometimes years. Spot Bitcoin, which had printed an all-time high near $126,000 that same month, began a grinding retreat that has now carried it roughly 35% below that peak, to around $82,699.
The mechanism of that event matters, because it is the same mechanism running today, only quieter. When a macro shock hits a market whose leverage is concentrated in a narrow band of prices, the first liquidations trigger a reflexive loop: forced selling pushes price into the next band, which triggers the next wave of forced selling, and so on until the collateral is exhausted or the venues intervene. The October cascade was not a single event. It was a sequence of them, chained together by the geometry of where the leverage sat.
A year later, that anniversary has become a Rorschach test. Optimists see healing. The absence of a repeat cascade, the return of positive funding, the resilience of the $80,000 handle โ all of it reads as convalescence. One prominent voice, Eric Conner, a co-author of EIP-1559 and a well-known figure in the Ethereum ecosystem, went further: positioning is thin, he argued. The market is under-owned. The pain is behind us.
That is a claim about market structure. And market structure is measurable. So I measured it.
What I found is that the three core derivatives indicators โ estimated leverage ratio, open interest, and funding rate โ are not telling a single story. They are telling three different stories, and the most optimistic reading requires you to ignore two of them. Let me take them one at a time, because the way they interact is the entire point.
Indicator One: The Leverage Ratio Is Rising, Not Falling
CryptoQuant's Estimated Leverage Ratio โ the ratio of open interest to exchange reserves, a proxy for how much leverage is stacked on top of available collateral โ sat at roughly 0.234 at the start of the week. It now reads approximately 0.256, with the broader figure hovering near 0.250.
That is an increase. On a metric explicitly designed to measure how crowded the leverage in the system is, the market got more crowded, not less, over the exact window in which the "we have deleveraged" narrative was being loudly repeated.
Read that again, because it is the crux of the entire anniversary debate. The single most direct measure of systemic leverage rose during a period that was universally described as post-deleveraging consolidation. If the market had genuinely flushed its leverage, this line would slope down. It slopes up.
There is a methodological caveat, and I want to be honest about it, because I spent the early years of my career auditing smart contracts precisely because I learned that undisclosed methodology hides the real story. CryptoQuant does not publish the precise composition of ELR in the discussion at hand โ specifically, whether it captures only perpetual futures, or whether it also folds in spot margin, collateralized lending, and on-chain DeFi leverage. That matters enormously. If ELR includes DeFi borrowing, then a meaningful part of the rise could be coming from on-chain credit markets that never participated in the October cascade at all. I flag this as a medium-confidence inference, not a fact.
But here is the thing: whichever way the composition cuts, the direction of the move is not in dispute. Leverage, by the broadest available measure, expanded.
Indicator Two: Open Interest Is Falling
Now the contradiction. Aggregate open interest across venues declined from roughly $154 billion to about $142 billion over the same period. On its face, this is the deleveraging signal the optimists are pointing to. Fewer contracts outstanding means fewer positions at risk. Simple.
Except it is not simple, and the reason it is not simple is arithmetic. If open interest falls while the leverage ratio rises, the only way both can be true is if exchange reserves fell faster than open interest did. The denominator shrank more than the numerator. That is not the signature of a healthy market shedding risk. That is the signature of collateral leaving the system while positions stay put โ or, more precisely, of positions being closed involuntarily rather than voluntarily.
This is where I want to introduce the distinction that I think the entire "did we deleverage?" debate is missing, and it is a distinction I first learned the hard way during the 2017 ICO audit cycle. When I audited the fund-distribution logic of three ICO projects that year, I found reentrancy vulnerabilities that let value leak out of the contracts in ways the token holders never saw in the price chart. The lesson was permanent: price action and position data tell you what happened; only the mechanism tells you why.
Apply that lens here. Open interest fell. But the fall came predominantly through liquidation โ forced closures โ rather than through traders voluntarily reducing size. You can see the fingerprint of forced closure in the single-day data: on the move down to $80,393, roughly $1 billion in positions were liquidated in 24 hours, and $930 million of that โ 93% โ was long. That is not a market trimming exposure. That is a market being carried out on stretchers.
When leverage exits through liquidation, the collateral behind it is destroyed, not redeployed. The remaining book is smaller but more concentrated. And a smaller, more concentrated book sitting on top of thinner liquidity is precisely the condition that makes the next cascade worse, not better.
Indicator Three: Funding Rates Collapsed
Here, finally, is the one hard data point that genuinely supports the deleveraging thesis. Funding rates on perpetual futures fell from an annualized 26.9% on Deribit to 7.1%, and to roughly 3.5% on OKX on a seven-day average. That is a decline of roughly 74% at the top end. In plain terms: longs are paying far less to hold their positions. The speculative premium has been squeezed out of the perp market.
I want to give this its due, because it is real. A funding rate falling from the high twenties to single digits is genuine cooling. It tells you the most frothy, most crowded expression of long leverage โ the perpetual futures basis trade โ has deflated.
But notice the scope. Funding rates measure perpetual futures and nothing else. They are silent on spot margin, silent on collateralized lending desks, and silent on DeFi credit markets. So when someone cites falling funding as proof of system-wide deleveraging, they are generalizing from one chamber of the house to the entire structure.
The three indicators do not agree, and the disagreement is the signal. Funding says the perp market cooled. Open interest says positions closed. But ELR says the system as a whole carries more leverage than it did a week ago. The only coherent reading that satisfies all three is this: the perp market cooled through forced long liquidation, the collateral base shrank, and whatever leverage remains is now denser relative to the smaller book it sits on.
Leverage doesn't disappear during a crisis. It concentrates.
The Thinner Book Problem
There is a second-order consequence of this that rarely makes it into the commentary, and it is the part that worries me most.
When open interest contracts and exchange reserves contract alongside it, market depth thins. The order book gets shallower. And here is the mechanical reality: the same dollar of leverage that was comfortably absorbed by a deep book six months ago now moves price further, triggers more liquidations, and cascades faster. The leverage ratio does not merely describe crowding. It describes fragility.
This is the mechanism I modeled during the 2020 DeFi Summer, when I assembled a team of four analysts to stress-test the capital efficiency of Yearn's early vaults. The headline yield was magnificent. The structure underneath was brittle, because that yield depended on a liquidity depth that would vanish the moment it was actually needed. We published the deleveraging prediction, and it played out โ not because we were clever, but because we read the mechanism instead of the marketing.
The same brittleness is present here, and it is quantified by Glassnode's liquidation cluster mapping. Below the market, at approximately $75,000, sits a dense band of levered positions waiting to be liquidated. That is not a support level. It is a magnet. When price approaches a cluster of this density, the reflexive dynamic is self-reinforcing: falling price triggers liquidations, liquidations push price down, which triggers more liquidations. The cluster is the fuel.
BTC is currently testing the $80,000 handle, which is both a psychological round number and a technical inflection. If it holds, the market gets another week to pretend the anniversary taught it something. If it fails, the path of least resistance runs straight into the $75,000 cluster, and the $1 billion of liquidations we just saw will look like a rehearsal.
The Exchange Risk Layer Nobody Discusses
There is an institutional dimension to this that deserves more attention than it gets, because it is where the cascade actually becomes systemic.
When a cluster of positions goes underwater faster than the order book can absorb, exchanges do not simply let the liquidations clear at market. They invoke auto-deleveraging โ ADL โ a mechanism that forcibly closes the opposing side of the book to socialize the loss. In extreme conditions, this means profitable traders get their winning positions closed against their will, because the venue's insurance fund cannot cover the gap. The October 10 event almost certainly triggered ADL across multiple venues. So did the more recent $1 billion flush.
The recurrence of these anniversary-scale liquidation events is not a coincidence, and it is not a sign that the market is healing. It is a sign that the risk controls at centralized venues remain a systemic weak point. Every cascade tests them, and every cascade reveals the same thing: under enough leverage concentration, ADL stops being a safety valve and becomes a transmission channel, spreading forced closures from the most fragile accounts to the most solvent ones.
This is why the leverage ratio matters more than the open interest headline. OI tells you how many positions exist. ELR tells you how much they are stacked on top of each other, and therefore how violently they will interact when the price moves against them. A falling OI with a rising ELR is a market that has replaced diversity of positioning with concentration of risk. That is the opposite of resilience.
Sentiment: The Divergence Nobody Is Pricing
Now let me leave the mechanics and talk about psychology, because the sentiment data is, in my judgment, the most under-weighted risk signal on the board.
The Fear & Greed Index โ Alternative.me's composite of volatility, momentum, volume, and social metrics โ reads 64. That is squarely in "Greed" territory. During the recent selloff, it dipped only to 59. It never crossed into fear. It never even approached it.
Hold those two facts next to each other. Bitcoin is 35% below its all-time high, and the market is still greedy. A market that is 35% off its highs and still registering Greed has not capitulated. It has not even begun to. Capitulation is a psychological event, and the psychology here is one of stubborn optimism โ the behavioral signature of participants who have not yet been forced to accept the loss.
I have written before about how "community" narratives in crypto function as financial mechanisms dressed in sociological clothing, and this is the same phenomenon in reverse. The "we're fine, we deleveraged" story is not an analysis. It is a comfort object. And comfort objects are exactly what you would expect to find at a market top that has not finished marking itself down.
The Consensus Safety Illusion
Here is where I part company with the optimists, and I want to be precise about why.
The argument that "positioning is thin, therefore the market is safe" commits a classic error: it treats the absence of visible leverage as the absence of risk. But risk does not vanish when leverage exits. It relocates. The October cascade did not eliminate the structural fragility of a highly levered, thinly collateralized derivatives market. It simply cleared one cohort of holders and handed their positions, at better prices, to a new cohort โ who then re-levered into the same fragile structure.
The market believes it is safer than it is. That belief is itself the risk.
This is a recurring pattern, and it is why I have learned to distrust moments of consensus calm. In 2021, when the NFT market was in full euphoria, I hedged against it โ put options on NFT index tokens, shorts on the underlying ETH pairs โ not because I had a crystal ball, but because the valuation metrics and the cultural narrative had decoupled completely. The community was loud. The fundamentals were quiet. The gap between the two is where risk accumulates. It generated $150,000 on that trade, and more importantly, it reinforced the discipline: when everyone agrees the danger has passed, check the mechanism again.
Conner may be right that his own positioning is light. He may even be right that institutional positioning is light. But "thin positioning" and "low leverage" are not the same claim, and the data supports only one of them. Open interest fell. Leverage rose. Those two sentences are not contradictory; they are complementary. They describe a market that shed holders without shedding risk.
Leverage doesn't forgive. It waits.
What Actually Changed
Let me be fair to the bull case, because I am not bearish for the sake of being bearish โ I have no interest in the emotionalism of either camp.
Something real did change. Funding rates normalized, which means the perp basis trade is no longer paying a punitive premium to stay long. That is genuine de-risking within the perpetual futures complex, and it is the one indicator pointing toward healing. If funding stays in single digits and price stabilizes above $80,000, you have the ingredients of a base โ not a bottom, but a base.
The question is what has to happen for that base to become durable. Three conditions, in my framework:
First, ELR needs to fall back below its start-of-week level of 0.234, ideally decisively. Until the broadest leverage measure turns down, the deleveraging thesis is unproven, no matter how many times it is repeated.
Second, the Fear & Greed Index needs to break below 50 and stay there for a meaningful period. A market that has not touched fear has not completed its psychological reset. Greed during a 35% drawdown is a warning, not a green light.
Third, open interest needs to stabilize rather than continue bleeding. Continued OI decline alongside rising ELR is the signature of an unstable equilibrium โ a smaller book carrying denser leverage, one shock away from a reflexive cascade.
Until those three conditions align, the anniversary narrative is a story the market is telling itself to avoid the discomfort of the data. And I have never found it profitable to trade the story over the data.
Positioning for the Second Leg
So where does that leave a macro watcher?
It leaves me treating $80,000 as the fulcrum of the entire setup. Above it, the market gets to believe its own narrative, and the anniversary passes as a scare. Below it โ specifically, a decisive close beneath it โ the path opens toward the $75,000 liquidation cluster, and the second cascade becomes the base case rather than the tail risk.
The asymmetry here is the whole trade. The downside is mechanical and mapped: $75,000 is not a guess, it is a cluster of positions that will be forced out if reached. The upside is narrative and unproven: the "we deleveraged" story requires you to ignore the leverage ratio that says otherwise.
I am not calling for a crash. I am calling for honesty about the structure. A market that mistakes liquidation for deleveraging is a market that will be surprised by its own fragility. The October 10 cascade was not the end of the leverage cycle. It was a clearing event โ and clearing events redistribute risk, they do not abolish it.
Leverage doesn't clear itself. It finds a new floor and rebuilds.
The only open question is whether that new floor is $80,000, or $75,000. The data, read honestly, is leaning one way. The sentiment index, still sitting at Greed with the market down 35%, is leaning the same way. The anniversary was supposed to be a moment of closure. Instead, it is a moment of measurement โ and the measurement says the market is not as safe as it feels.
That is the trade. Not a direction, but a divergence: between how the market feels about its leverage, and what the leverage actually is. When those two finally converge, one of them will have to move โ and the data has already told us which one.