March 2020. I was brand new to the industry I’d just joined — crypto. I had recently moved from Russia to Singapore, I worked in client support, and I was waiting for my first paycheck. The only number that really mattered to me was the one that would land in my account at the end of the month.

So when the market broke, I watched it from the outside. I didn’t have the depth of knowledge — or the understanding — to see that a price this low was actually a rare opportunity, that these are the moments you’re supposed to buy. I simply didn’t see it.

Bitcoin had spent mid-February above $10,000. By the morning of March 12 it was sliding, and within hours it had fallen more than thirty percent, briefly trading under $4,000 after starting that week above $9,000. People later called it Black Thursday. By the daily close the low printed near $4,900; intraday it touched roughly $3,800. Not the three-thousand I sometimes misremember — but for anyone who was long with leverage that night, the difference didn’t matter.

I remember the clients. I worked in support, and I remember the ones who were screaming — who wrote about thoughts of suicide, who had lost hundreds of thousands of dollars in a single night. It is genuinely frightening. Every spike of this kind works the same way: it lets a few win and almost destroys the rest. And you sit on the other side of the screen reading all of it in real time.

A colleague asked me back then: “Sani, will you buy Bitcoin?” I smiled. That was all I did. One thought in my head: I just need to get to payday — why would I want Bitcoin? It was an abstraction to me. The rent was not.

Six months later I understood what I had walked past. Bitcoin closed that year near $30,000. The window I had watched from the outside turned out to be the cheapest entry of a generation, and I had spent it waiting for a salary.

I am not telling this for the regret. Regret is cheap and teaches nothing. I am telling it because six years later I am looking through the same kind of window — and this time I want to break down what is actually the same and what is completely different.

June 2026

This week Bitcoin fell below $67,000 for the first time in two months. On June 3 it traded near $66,700 after sliding about 6.75% in a day, down from an intraweek high near $75,850 — roughly a twenty-two percent move inside a single week. The 24-hour liquidation tally crossed $1.48 billion, the overwhelming majority of it long positions, and more than two hundred thousand traders were caught in it. Measured from the October peak near $126,000, Bitcoin is now down around forty-five percent.

The air around the asset is thick again — “scam” is everywhere, the obituaries are being drafted. It rhymes with 2020. And the first instinct is to overlay the two charts and shout: “buy the dip, like back then!” But before giving in to that instinct, let’s compare the two periods honestly. Not by feel — by facts.

1. Causes of the crash: then and now. In March 2020 there was one cause, external and sudden — the COVID pandemic. Lockdowns, panic, a global dash for cash, a liquidity vacuum that drained every risk asset at once. It was a shock from outside. In June 2026 the cause is internal and slow: record outflows from spot Bitcoin ETFs (roughly $2.4 billion left in May — the worst month of the year), institutional selling including Strategy unwinding, sticky inflation, and a strong dollar. Then, an instant blow. Now, a slow bleed.

2. Wars and geopolitics: then and now. In March 2020 there were no large-scale active conflicts — the whole world was staring at one point, the virus. Now the backdrop is fundamentally heavier: Russia’s special military operation in Ukraine is in its fifth year, and in parallel the United States, alongside Israel, is in a military conflict with Iran. So in 2020 there was a single global stress factor. In 2026 there are several at once, and any one of them can hand the market a fresh push downward at any moment.

3. Who’s in power and what the Fed rate is: then and now. Here is the most curious coincidence. In both cases the U.S. president is Donald Trump: in March 2020 it was his first term, now his second (since January 2025). But the Fed rate is the exact opposite. In March 2020 the Fed slashed rates to 0–0.25% in an emergency and launched $700 billion of QE: the medicine was being prepared in the very moment of the crash. In June 2026 the rate sits at 3.50–3.75% and hasn’t moved since March — the Fed paused after its late-2025 cuts and is holding high. The next Fed meeting is June 16–17, 2026, with an updated rate forecast (the dot plot) — the nearest point where the picture could shift. The same man at the wheel, and a mirror-opposite monetary policy. That, to me, is the single biggest difference between the two periods.

4. Oil and gold: then and now. Here is another important difference, and it’s about how Bitcoin itself behaves. In March 2020 everything fell at once: WTI crude crashed to about $25 a barrel (its lowest since 2003, as the OPEC+ deal collapsed), and gold sank to roughly $1,472 an ounce — investors were selling even safe-haven assets to raise cash. Bitcoin fell right along with them. In June 2026 the picture is reversed: oil is above $95 a barrel, gold is near $4,500 an ounce — both rising on a geopolitical risk premium and safe-haven demand. And Bitcoin is falling. In other words, it is behaving now not like “digital gold” but like a risk asset people are dumping. In 2020 it sank with the market in a shared panic; in 2026 it sinks alone while the classic havens climb. To me that’s a signal: the 2020 analogy is elegant, but Bitcoin is playing a different role inside it.

5. What comes next — two scenarios. On the optimistic side: the market needs a catalyst, and this time it won’t appear out of thin air the way it did in 2020 when Fed money lifted everything. The help would have to come from the Fed itself — resuming rate cuts (a new Fed Chair takes the seat in 2026, and a lot rides on the signals they send), plus a return of ETF inflows and at least a partial easing on the geopolitical front. In that scenario, today’s levels really would turn out to be the opportunity I once slept through.

On the cautious side: with the rate at 3.5–3.75% and several parallel conflicts running, there’s no money flood to wait for like in 2020. In that case the floor could land below current levels — the technical picture allows for zones around $63,000 and lower — and “buying the dip” by analogy to 2020 would become exactly the mistake I have spent years documenting in my own account: entering on hype, with no reason and no line that proves me wrong.

What I Actually Think

Honestly, what reminds me of March 2020 right now isn’t the numbers — we’ve already established the two periods differ on the data. It’s the feeling: the way Bitcoin is falling. The character of the move, the density of panic around it. Although, come to think of it, October 2025 felt just as similar — everything was sliding then too, and I wanted to draw the same analogy.

And here I have to be honest: in the strict sense, there is no analogy. For 2026 to truly repeat 2020, the miracle I described above would have to happen — the Fed cuts rates, fresh money flows into the system, geopolitics eases at least a little. None of that is happening yet, and leaning on “it bounced last time” isn’t analysis, it’s hope.

But now I’ll speak not as an analyst, just from personal experience. If I have spare funds — not money I’d be afraid to lose, but genuinely spare — I’d rather put them into Bitcoin. Even now, with more FUD around the word “Bitcoin” than I can remember in a long while: “scam,” “it’s over,” “it’s dead.” I heard it in 2020, I hear it now. This is not a trade recommendation and not a signal — it’s simply what I do, with my money and my horizon.

The difference between the me of 2020 and the me now is that back then I wasn’t holding back out of discipline — I had no reason to act and no reason to wait, I just didn’t see. Now I do see. And if I decide to enter, I know why, over what horizon, and what share of capital I’m willing to lose if I’m wrong. That’s the thing I lacked six years ago.

I’m not telling you to buy the dip. I’m not telling you to run. The whole point of The White Box is that I don’t sell certainty — I keep notes on my own uncertainty, paid for in real money. What you do with yours is up to you.

Six years ago I smiled and waited for a paycheck. This time I’m not smiling. I am watching, I am writing it down — and yes, for myself, I’ve already made my choice.

— Sani P.

Before you continue reading

This is the most important case of the three I’ve published. Not because the loss is the largest — though it is the largest. Because this is the only case that ended with me having to start a course of antidepressants.

I want to say this at the beginning, not at the end, because this is the substance of the case.

I am a professional. I work in the crypto industry. I know what a stop-loss is. I know you don’t average down on a losing position. I know you don’t trade with a deposit that has a specific life-purpose. I knew all of this and did the exact opposite.

I did it because I believed non-professionals with audiences of tens of thousands on YouTube and Telegram who, at exactly the same time, were calling on their followers to average down. They earn through affiliate arrangements with exchanges — the more their subscribers trade, the more they earn. And they were saying, every day throughout November, that BTC was “about to reverse.” I listened. I added.

Five lessons confirmed personally and burned in:

• You cannot keep depositing into a losing position. I deposited $555,341 across 24 averaging-down additions. Each one only enlarged the risk.

• You cannot trust a single influencer with affiliate income. If their pay depends on you trading, their analysis is a sales job.

• A stop-loss is not optional. It is a structural element of the position. I had none.

• You do not trade with money that has “a name” in your head. Mine had one. It was “Singapore home deposit.”

• A $104k loss, when the money had a name, can cause clinical depression. In my case it did. Antidepressants helped. Time helped. People close to me helped.

If I could give back $104k in exchange for reading the above two paragraphs ahead of time — I would do it without a second’s thought.

The rest of this essay is the math, the chronology, and the personal architecture of how all five of those rules got violated simultaneously over five months.

—$104,047.

That’s what I lost over 143 days on a single BTC long position that I added to twenty-four times on the way down, partially capitulated on, then added to again, and finally watched get force-liquidated on a day when four independent macro events fired simultaneously.

This is the third case study in The White Box series. The first one — Seven Months in Blur — was a 213-day slow grind on a single LONG. The second one — Five Days at the ATH — was a fast break from a record-OI cascade. This one is the one in between: a medium-speed catastrophe with the most mistakes layered on top of each other.

This is also the largest loss of the three. And the one with the most repeats of the same decision.

I’m going to break this trade down the same way: phase by phase, addition by addition, mistake by mistake.

🔍 New here?

The White Box is post-trade analytics for crypto traders. I take real trades — mine, and submissions from other traders — and break them down with the full math, the exact timestamps, the macro events that hit during the hold, and the cognitive patterns that produced the decision. Subscribe to get every case study in your inbox, plus the framework I’m building from each one.

👉 Subscribe (free)

🔗 Want the structured version with all the data tables?

This essay is the long-form analysis. The case page on the site has every phase, every funding number, the geopolitical timeline, the comparison table with Blur and October, and the rules. If you want the spreadsheet, go there:

👉 thewhitebox.net/#january31

Why I’m writing this one now

The October case ended at the cascade. The Blur case ended at capitulation. This one ended at the moment when four independent things converged on a Saturday afternoon, and the position closed itself.

But the case isn’t really about that moment. The case is about the twenty-four decisions before it.

October is a story about size and one tweet. Blur is a story about a chemical state and 213 days of hope. This one is a story about a particular kind of mistake that compounds — the addition.

If you’ve ever added to a losing position, this case is for you.

October 24, 2025. The first addition that didn’t need to be made.

Let me back up.

Before this position, I had three consecutive winning swing trades on BTCUSD inverse — September 10, October 12, October 20. Each one a 1-3 day hold. Each one a small profit. Each one confirming a calibration: I understand BTCUSD swings.

What I did on October 24 was open a new LONG at $111,290. Standard size. Held it for less than a day. Closed it October 25 at $111,434 — basically flat, fees ate the difference.

This is a moment I want to look at carefully, because in isolation it doesn’t look like anything. It looks like a small failed trade. It is the moment when the calibration started to drift.

The previous three trades had been clear setups: technical signal, defined exit, executed. This October 24 trade had no setup. It was “I think BTC should bounce here.” No level, no signal, no exit plan. The fact that it was tiny and flat hid what had actually happened: I had moved from setup-based trading to hope-based positioning while still using the same instrument and the same size patterns.

I didn’t notice.

October 26. The reopening that started the bag.

October 26, I reopened at $112,321. 137,368 contracts long.

This is the moment to which I will return many times in this essay, because it’s the moment the position stopped being a trade and became a thesis I had to defend.

The difference is invisible from inside. From inside, it just felt like “the price is good here, I’ll take size again.” But from outside — looking at the chart, looking at the funding rate, looking at the order history — there was no setup change between Oct 24 and Oct 26. Same instrument. Same direction. Same size. Just a re-press of the same button, two days after it didn’t work.

This is not how setup-driven trading works. This is how identity-driven trading works: I am someone who is long BTC right now. The trade is no longer about the chart. It’s about who I am in the market.

I want to be careful here, because this concept gets thrown around loosely. Identity-driven doesn’t mean ego-driven. It means: the position has become an answer to “what am I doing?” rather than “what is the chart doing?” Once that switch happens, you cannot use the same techniques to manage the position that you used to enter it.

I didn’t know this on October 26. I learned it over the next ninety-five days.

October 29 — November 18. The twenty-four additions.

Then BTC fell.

Over the next twenty-three days, I added to the position twenty-four times. Here is the receipt:

Date Add at Cumulative position Oct 29 $110,700, $110,205 174k contracts Oct 30 $106,930 187k Nov 3 $105,790, $105,600, $105,430, $105,950 207k Nov 4 $103,500, $101,100, $101,050 279k Nov 13 $98,200 289k Nov 14 $97,300, $97,000, $95,000, $95,200, $94,000 489k Nov 16 $94,250 634k Nov 18 $89,830 643k

A 5x increase in position size while BTC fell 20%. My average entry was getting better. My total risk was getting worse.

This is the part of the case that, looking back, contains the most mathematics and the least judgment. Each individual addition had a small justification: the price is lower than before, my average improves. No individual addition was crazy. It was the pattern of addition itself that was crazy.

I want to name this precisely, because it is the central mechanism of bag-holding through additions:

Each addition lowers the average price. Each addition raises the total exposure. The first thing makes you feel like you are managing the position. The second thing is the position managing you.

If your conviction in the thesis is high enough to add a fourth time, your conviction was wrong on the first add. Because the first add was a thesis-confirming trade. The fourth add is the chart-disconfirming trade. The fact that the second, third, and fourth adds all look the same from inside is the trap.

By November 18 I had 692,709 contracts long at an average of $100,193. The position had grown five-fold while my thesis had been disconfirmed every single time.

November 19-21. The cut that should have been the end.

On November 19, BTC was at $91,000. I started exiting.

Here is what I closed over three days:

Date Closed at Realized P&L Nov 19 $91,200 −$15,900 Nov 20 $87,600 −$13,800 Nov 21 $85,552 −$8,300 Nov 21 $81,850 −$7,250

About $45,000 of locked-in loss over three days. Position reduced from 692,709 → 212,726 contracts.

I want to pause here, because this is the part of the case where, if you’re reading this looking for the moment to learn from, this is the one.

On November 21, I had just done the right thing. I had locked in a partial loss. I had reduced my exposure dramatically. I had effectively admitted, with my own money and on my own statement, that the thesis was not working. The work of recovery from this position was already underway. All I had to do was let November 21 stand as the decision it was.

I didn’t.

November 24-27. The re-addition that owns this case.

On November 24, I added back. 13,539 contracts at $87,900. On November 27, I added more. 17,931 contracts at $90,195.

Position now: 244,196 contracts at an average price of $98,288.

This is the most expensive decision in this entire case study. Not because of the dollar amount of these two adds — they were small. The decision was expensive because it nullified the previous decision.

Three days earlier, I had told myself with my actions: the thesis is wrong, I am reducing exposure. Three days later, I told myself with my actions: no it isn’t, I was just temporarily scared, the thesis is still alive.

I’m going to be precise about what this is, because it’s worse than just a bad trade. It’s the dissolution of the previous decision. And the dissolution is more expensive than the decision itself, because once you can dissolve your own decisions in three days, you have removed the only tool you have left for managing a losing position.

Closing was the tool. I used the tool. Then I undid the use of the tool. Now I had no tool.

November 27 — January 31. Sixty-five days of hope.

For the next sixty-five days, the position did not change. 244,196 contracts. Average price $98,288. BTC slid from $90,000 → $85,000 → $80,000.

My thesis during this period, to the extent I can articulate it honestly: BTC will come back to $100,000 by January or February. I just need to wait.

That is not a thesis. That is a hope structured to look like a thesis.

The difference is in the form. A thesis says: if X happens, I exit. If Y happens, I add. If Z time passes without X or Y, I revise. A hope says: I will be right eventually.

The way you tell them apart is by looking at what would convince you you were wrong. A thesis has an answer. A hope has an internal protest.

For sixty-five days I had no answer. The position was a hope.

A note on funding, and why “right side” wasn’t enough.

In the October case, I paid funding sixteen times in a row — the market told me sixteen times that I was on the overloaded side, and I ignored it.

In this case, the opposite happened.

Over those 65 days, I had 315 funding events. 254 of them came in. I received money. The market was telling me, 254 times, that I was on the underloaded side — that shorts were overcrowded, that any squeeze would benefit me.

And I was still liquidated.

This is worth naming separately, because the lesson from October could be misread as “follow the funding signal.” It isn’t that. Funding is not a directional signal. Funding is a crowdedness signal. Being on the right side of crowdedness doesn’t protect you from being on the wrong side of price.

What happened in January was that price moved against me for reasons that had nothing to do with leverage and everything to do with macro: Iran, shutdown, silver, Microsoft. None of these are forecastable from the funding rate. The funding rate measured how the leveraged traders were positioned. The price was set by the institutional flows underneath.

I had been calibrating on funding as a safety signal. It is not a safety signal. It is one piece of information about one side of the market.

January 31, 2026. The four-event day.

By January 31, BTC had been declining for two months — slowly, on accumulating macro pressure, without a single sharp trigger.

Here is what happened on that Saturday:

Beginning of the day — the first US government shutdown of 2026 began. Political uncertainty into the open.

~14:30 UTC — explosion at Iran’s Bandar Abbas port. Geopolitical risk-off across all assets.

During the day — silver crashed −30%. Worst single-day move for silver since 1980. The traditional safe-haven trade collapsed.

Same day — Microsoft missed earnings. NASDAQ sold off into the close on tech contagion fears. Crypto correlates with NASDAQ in this regime.

Background — US Bitcoin ETFs had recorded $1.61B of outflows in January alone, including $818M on the 29th. The structural bid that had supported BTC for months was gone.

BTC opened January 31 at $84,106. By 08:00 UTC, $83,394. By 16:00 UTC, $81,154. By 17:11 UTC, $78,491.50.

At that price, my maintenance margin requirement was no longer met. The position closed itself.

244,196 contracts. Market sell. Realized loss on the liquidation event alone: −$49,836.

What the liquidation actually closed.

I want to be precise about the size of this loss, because the number on the liquidation event ($49,836) is not the full size of the case.

On the liquidation event: −$49,836 Earlier partial closes in November: −$45,250 Earlier swings and adjustments (the noise): −$8,000 or so Trading fees across 143 days: −$1,147

Total realized loss on the BTCUSD instrument, Sept 10 — Jan 31: −$104,047.

The liquidation is the headline. The November closes are the structural reality. A position that takes $104k to die was already broken in November.

A note on what the money was for.

I’m going to step away from the math for a few sections, because the math is not the part of this case that’s been most useful for me to understand in the months since.

The money in this position was not “free risk capital.” It was my savings for buying a new home in Singapore. A sum I had been collecting for years out of salary — not out of crypto profits — and was holding on Bybit because I thought of the position as a short-term maneuver. I’ll just run it through and withdraw.

This is a fact I usually don’t say in public. I’m saying it now because it explains the single most important thing about this case: why I could not close the position at a loss.

Closing at a loss would not have meant locking in a trading loss. Closing at a loss would have meant admitting I had lost the deposit for a home. Those two operations are mathematically identical and psychologically completely different.

When people say “never trade with money you can’t afford to lose,” they usually mean it as a risk-management rule about position sizing. The deeper rule is about what name the money has in your own head. This money had a name. The name was “new home.”

Every averaging-down of the entry price was an attempt — at the level of the nervous system — to preserve that name. Eleven times, over twenty-three days.

The market does not know what your money is for. And one of the most expensive lessons of my trading life is that any amplification of a position’s emotional significance through its tie to a life goal is a mechanism of losing, not winning.

A note on whose advice I was reading.

During those eight weeks in November — through the additions, through the partial capitulation, through the re-additions — I was reading several large crypto influencers very carefully. People with large audiences who regularly publish market analysis.

They were saying: BTC is about to correct and rise. They were saying: they are averaging into their position at these levels. They were saying: this is a buy.

They said this the entire time BTC fell from $112,000 to $90,000.

I know — because this is my industry — that many of these people operate on affiliate arrangements with exchanges. They bring users onto a platform, and for every dollar of trading volume those users generate, they receive a percentage. The more their subscribers trade, the more they earn.

This does not make them malicious. Most of them, I believe, sincerely think what they say. But it creates a structural conflict of interest: it is in their financial interest for their audience to trade — regardless of whether the audience wins or loses. And content that convinces audiences to hold and average down during a decline is particularly optimal for that business, because holding-and-adding generates the maximum trading volume per audience member.

I knew this. And I read them every day anyway. And every time they said “I’m adding,” I added too.

In the end, BTC did not rise. And my funds — let me note — did average the entry price down. From my own home deposit, which I had been collecting for years, I was financing their affiliate income through the trading volume they generated on me.

This is one of the most expensive lessons of the whole case. When a person whose income depends on your trades advises you to make trades — that is not objective advice. That is a job. I now read crypto influencers the way I read advertising. Not because they’re lying. But because they have a structural reason to be wrong in a specific direction.

The alarm.

I want to describe one thing about those sixty-five days of holding, because I think it captures more than I could explain abstractly.

In Singapore, when American trading hours open, it is deep night by local time. Around 02:00–03:00 AM. The strongest BTC price movement usually happens in the first hours of the US session.

I started setting an alarm every hour.

Not every two hours, not for the main sessions — every hour. When it rang, I would wake up, take my phone in the dark, check the position, check the price, check the funding rate. If the price was moving, I would stare at the chart for fifteen, twenty minutes, trying to figure out if it was a reversal or a continuation. Then I would set the alarm again.

This went on for weeks.

My husband slept next to me. I tried to move quietly so as not to wake him. Sometimes I went out of the bedroom and sat with the phone in the kitchen. At four in the morning. With a BTC chart in my hands. I couldn’t explain to him what was happening — he knew I was trading, but I didn’t show him how deep I was sitting in it now.

This is the moment when trading stopped being an activity and became a state. I was no longer “trading.” I was living inside an open position, as if inside a room with no exit.

I couldn’t talk. I couldn’t eat. I couldn’t drink. I was experiencing every price movement so intensely that my body stopped functioning normally.

What I had to do after January 31.

After the liquidation, I had to start a course of antidepressants.

I thought for a long time about whether to put this in this essay. It’s the most personal thing I have shared in this series. But this is exactly what the case is. A structural breakdown of $104k without this fact is incomplete.

I did not understand the meaning of life. I did not want anything. The home deposit I had been collecting for years was burned in a few hours on the evening of January 31. This is a very concrete kind of emptiness that cannot be described abstractly. It is an emptiness you feel in your body, not in your head.

It felt to me that no one could feel my pain. That this was some personal misfortune of mine, my personal stupidity. That I had to deal with it alone, and then, once I was “normal” again, I would be able to talk to people.

That period lasted several weeks. The antidepressants helped. The support of someone close to me helped. Time helped.

I’m not writing this to summon sympathy. I’m writing it because in the trading industry, this is not talked about. All stories of losses end at the loss number. What happens after — what months of silence set in, what medications you have to start, what relationships with people close to you fracture — no one writes about that.

If you, reading this, recognize yourself: know that this is not a moral failure. This is a known neurobiological reaction to a major financial loss, especially when the money had “a name.” It is treatable. People come out of it.

Why this is why I’m building The White Box.

After treatment, I felt a strange emptiness. Not bad. More like an empty room after renovation, when the furniture hasn’t been delivered yet.

And in that empty room I understood something permanently.

Not “don’t trade with leverage.” Not “set stop-losses.” Those are all technical lessons I already knew before Blur, before October, before BTCUSD inverse — and they didn’t stop me.

I understood: the market does not know what your money is for. And I understood something else, which is the reason I’m writing this at all.

When I was holding a position at $90k unrealized loss in December 2025, I was reading articles by crypto influencers with affiliate income. What I actually needed was something else entirely. I needed stories from people who had already been through this — with all the financial math, with all the personal devastation, with all the consequences.

Stories like that almost don’t exist. The trading industry produces case studies of winners. It produces obituaries of bankruptcies. It does not produce the genre I needed: the trader who lost, came out of it, and wrote it down honestly.

So I’m writing them myself. Mine first, so I can ask others to write theirs.

That is what The White Box is. It is the library I needed in December 2025, and didn’t have.

A short note on inverse contracts.

One technical thing this case taught me, that I should have understood before I opened the position: in an inverse contract, the contract size is denominated in USD, but your margin is in BTC.

This means: when BTC falls, your position loses money in USD, AND your margin loses purchasing power in USD. You are exposed to BTC decline twice. Once as the asset you are long. Once as the asset you are margined in.

I chose BTCUSD inverse for a reason that, in retrospect, was trivial: I had BTC sitting on the exchange, and using it as margin without converting saved about $50-100 of conversion fees on a typical position.

The savings were less than 0.1% of the final loss.

When choosing contract type, the question is never about the fee. The question is: what exposure am I taking, and have I priced it correctly? Coin-margined contracts in a down move are not the same trade as linear contracts in a down move. They are a more leveraged version of the same trade.

I did not know that. Now I do.

Six structural mistakes that make this trade reproducible.

Mistake 1: I chose inverse contracts for the wrong reason. The reason was fee savings on conversion. The real cost was unrecognized double exposure to BTC decline. Coin-margined exposure is structurally more leveraged in a down move than the position size suggests.

Mistake 2: I switched from setup-based to identity-based trading without noticing. After three winning swings, the October 24 entry had no setup. It was a re-press of the same button. This is the moment trading discipline silently dies. If you re-enter the same instrument in the same direction within 48 hours of a flat exit, you are no longer trading the chart — you are managing an identity.

Mistake 3: I averaged down twenty-four times without revising the thesis once. Each addition lowered the average price and raised the total risk. The math improved with each add. The thesis was disconfirmed every time. Averaging down on an unchanged thesis is sunk cost fallacy with a calculator.

Mistake 4: I partially capitulated, then re-added within three days. The November 19-21 exits were correct. The November 24-27 re-additions nullified them. Re-add after partial capitulation is the most expensive trade in a losing campaign — it pays nothing and dismantles the only management decision you had left.

Mistake 5: I held for 65 days on a hope structured to look like a thesis. “BTC will return to $100k” had no exit triggers, no time bounds, no falsification criteria. A thesis without triggers is a hope. Hope held with leverage is a forced liquidation waiting for its trigger.

Mistake 6: I ignored four parallel macro signals that all pointed down. By January 31, four independent negative macros were active simultaneously. Any one of them required a position review. The four together were not a “risk” — they were an event already in progress. When multiple uncorrelated macro factors all point the same direction, you are not exposed to risk anymore. You are exposed to a trend that already started.

⚠️ Quick interruption.

This is Case 03 of three: Blur (2023), BTC October (2025), BTCUSD January (2026). The full series tracks the same trader across three different mistake patterns over three years. The structural similarities are more interesting than the differences.

Subscribe to The White Box to get the next case as soon as it goes live.

👉 Subscribe

How this case compares to the previous two.

Blur (2023) BTC October (2025) BTCUSD January (2026) Duration 213 days 5 days 143 days Instrument BLUR spot BTCUSDT linear BTCUSD inverse Setup before Winning SHORT on BLUR Winning SHORTs on BNB 3 swing wins on BTCUSD Size vs capital ~125% ~100% ~50% (but inverse-margined) Entry style One LONG entry Two LONG entries 11 averaging-down adds Funding signal Mostly negative 16/17 negative (I paid) 254/315 positive (I received) Exit trigger Capitulation, Sunday morning Forced on one tweet Forced on four macros Loss −$87,989 −$44,535 −$104,047 Character of pain Slow grind Fast break Slow grind + macro break

The biggest of the three. And the one with the most repeats.

Blur was one bad decision held for 213 days. October was one bad decision force-closed in 5. January was 11 additions, one partial capitulation, one re-add, 65 days of hope-holding, and four macros that converged on a Saturday.

If you are wondering which of these patterns is most common in everyday crypto trading — it’s this one. The Blur pattern is dramatic. The October pattern is fast. The January pattern is the one that most people are living through right now without knowing it.

Why The White Box exists, and why this case completes the trilogy.

After Blur, I had roughly $30,000 left. After October 10, I had roughly $1,300. After January 31, I had... worse.

Three trades. Together they destroyed approximately $237,000 against $80,000 of net deposited capital.

The math is uncomfortable. But the pattern is more uncomfortable than the math, because the pattern is consistent across radically different conditions — different instruments, different market regimes, different time horizons. Same trader. Same architecture of decisions.

That’s what The White Box is for. Real losses, decomposed into structural patterns, turned into rules. I publish my own first, so I can ask others to share theirs.

Case 04 is the April 17, 2026 BTCUSDT SHORT that was squeezed in the Iran ceasefire rally. It’s the inverse twin of this one — same trader, different direction, different mistake set. Coming next.

👉 Subscribe to get it

Other ways to follow:

• 📁 Full case studies, glossary, reading order: thewhitebox.net

• 🗂️ Structured version of this case: thewhitebox.net/#january31

• 💬 Telegram channel (RU): @thewhiteboxcrypto

— Sani

Case 02 of The White Box.

—$44,197.13.

That’s what I lost in a single forced liquidation on October 10, 2025. Not over months. In 60 seconds, at 21:13:04 UTC, twenty-three minutes after Donald Trump posted on Truth Social about 100% tariffs on China, and twenty-three minutes into the largest deleveraging event in crypto history: $19.13 billion liquidated in 24 hours, 1.6 million accounts wiped, nine times the previous record for any single-day crypto liquidation.

This is the second case study in The White Box series.

The first one — Seven Months in Blur — was a slow grind over 213 days.

This one is the opposite: a fast break in five days, with the same behavioral pattern underneath.

I’m going to break this trade down the same way: hour by hour, signal by signal, lie by lie.

Why I’m writing this one now

After Blur, I had roughly $30,000 left in my Bybit account.

The case study you just read — if you read it — ends in September 2023. Most readers assume the worst is over. The lesson has been paid for. The book is closed.

It wasn’t closed.

By June 2024, I had $679 left. By April 2026, $1,287. Between Blur and now, I executed more catastrophic single-position trades that, combined with Blur, destroyed approximately $155,000 against $80,000 of net deposited capital.

This essay is the middle one. October 10, 2025 — a long BTC position opened at $124,664.93 at the all-time-high, force-liquidated at $112,198.50 inside the cascade that broke crypto for the second half of last year.

I am writing this not because the trade is interesting on its own. I am writing this because the trade has the exact same structural signature as Blur. Different size, different instrument, different speed, different macro backdrop — same architecture. Once you see the pattern once, you cannot unsee it.

As you may have already understood from the Blur essay, I suffer from one specific thing: the desire to participate in the big race at its peak. FOMO is a genuinely terrifying psychological force, and one that can lead to enormous mistakes.

This essay exists because in October 2025, despite having lived through every single sentence of the Blur case study, I executed the same pattern again. Faster. With a different instrument. In a different chemistry of the market. With the same outcome.

Living through Blur did not save me from October. That is the hard fact I have to start this one with. Writing Blur down afterwards made the mistake legible to me — but it did not, by itself, change my body’s response in the moment when the next race started.

I am writing this one now because two essays about the same pattern in the same hands is the beginning of a structural account. One essay is a confession. Two is a method. And the method is what The White Box is for.

September 21 – October 4, 2025.

The setup that wasn’t a setup.

The eighteen days before the trade matter. The catastrophic position did not appear out of nowhere. It appeared on top of a winning streak, the same way Blur appeared on top of a small winning short.

Here is the receipt for those eighteen days. All shorts. All BNB and small BTC. All winners:

• September 21: closed BNB short for +$7,927

• September 22: three more BNB shorts closed for +$2,869

• September 23: two BNB shorts closed for +$4,549

• September 30: closed BNB short for +$13,293

• October 1–2: small BTC shorts for +$1,195

• October 7: three more BNB shorts closed for +$3,594

Roughly +$33,500 realized in eighteen days, on the short side.

This is what my screen looked like the morning of October 5, when I started buying BTC at $125,000.

I want to be specific about what this kind of winning streak does. It does not give you money. It gives you a chemical state — the same one I described in the Blur essay, eighteen months earlier. Conviction without analysis. The feeling that “I read this market.” The feeling that “the short regime is over, time to flip.”

Neither of those is a trading thesis. Both of them are emotional states dressed up as trades.

I had two more pieces of context in my head that morning, both of them dangerous:

First — BTC had just printed an all-time high above $126,000 the day before. I did not want to be on the sidelines watching the breakout.

Second — open interest on crypto perpetuals had climbed to a record $217 billion across all exchanges. Analysts had been flagging this number for days as a setup for a cascade. I read those threads. I scrolled past them.

What I saw on the screen that morning was Bitcoin going up. What I saw on Twitter and Telegram was the entire crypto industry in a positive, happy mood. I won’t pretend I was somehow outside of it — I was inside it. To be honest, I was in a pretty bright state myself.

And the entry decision sounded, in my head, like this: high margin plus leverage, and if BTC moves up just a few more points from here, I get a real profit.

That sentence is FOMO with the receipt attached. There is no analysis in it. There is no question about size, no question about timing, no question about what happens if BTC moves down instead of up. There is only the assumption of continuation, scaled up by leverage.

And underneath all of that — the same body that had been through Blur was the one placing this order.

October 5, 2025, 06:15:07 UTC. The 11-second entry.

At 06:15:07 UTC, fifty limit orders to buy BTC at $125,000 began filling on my Bybit account. By 06:15:18 — eleven seconds later — I held 2.998 BTC long. Notional value: $374,625.

This is one of those moments I want to describe precisely, because the mechanics of the entry already tell you something is wrong before the price moves an inch.

A controlled entry into a high-conviction trade looks like: one or two orders, sized to the thesis, executed at predefined levels. Fifty orders filling in eleven seconds is not a controlled entry. It’s a batched ladder of limit orders that was pre-staged and triggered at once. Which means: the decision was not “buy here, at this level, this size.” The decision was “I want to be long, fill me anywhere in this zone, do it now.”

I was not entering a trade. I was buying a feeling.

Within an hour after entry, BTC was already below $124,000. The position was already down. The morning of October 5 was the only morning in the five days the position existed where I could have closed it for a manageable loss. I did not.

October 5, 21:11 UTC. The averaging down.

Fifteen hours after the initial entry, I added another 0.464 BTC at $122,500.

Average entry now $124,664.93. Position size: 3.462 BTC. Notional: $431,590.

My free capital at that moment was approximately $40,000 to $50,000. Which means my position was about ten times my capital. Which means a 10% move against me would erase the account.

This is the exact size mistake from Blur. In Blur, the position was 125% of my free capital. In October, it was approximately 100%. The number is different, but the rule it broke is the same: never enter a position you cannot psychologically afford to close at a 20% drawdown.

The averaging down was not a thesis revision. The thesis hadn’t been revised — it had been reinforced by lower price, which is something different. Reinforced by lower price is what you do when the position is already controlling you and you are looking for permission to make it bigger.

There is one piece of context I want to add about this entire holding window that the trade history does not show: I was on a business trip. In Hong Kong. From the day I opened the position to the day it was liquidated.

Which means: my attention was split. My evenings were not fully mine. My mornings were not mine. My focus on the position was scattered across hotel rooms, meetings, and a time zone seven hours ahead of where my brain usually lives.

You cannot, structurally, place this size of a bet from that state. Not because it’s “irresponsible” in some abstract moral sense — but because a position this large requires uninterrupted access to your own judgment, and a business trip is by design the thing that removes that access. I did not have a plan for being away. I had a plan for being right.

October 6, 2025. The thing I didn’t know yet.

On October 6, Binance posted a quiet, technical announcement. It was not a viral tweet. It was the kind of update that traders who follow exchange-specific risk pages read, and almost no one else.

The announcement said: Binance will update the pricing oracle for USDe, bnSOL, and wBETH. Implementation date: October 14.

If you do not know what those three tokens are, the short version is: synthetic stablecoin (USDe) and two wrapped staking derivatives (bnSOL, wBETH). All three were accepted as collateral on Binance’s Unified Account system. All three were priced, for the purposes of collateral valuation, using Binance’s internal orderbook data — not external oracles like Chainlink.

The announcement created an eight-day window in which the oracle behavior on these three assets was documented, public, and about to change. October 6 to October 14.

I did not read this announcement.

Someone did.

That’s all I want to say about this for now. We’ll come back to it.

October 6–9. The funding receipt.

What happened over the next four days was that the market told me, sixteen times out of seventeen, that I was on the wrong side. The mechanism was funding payments.

On a Bitcoin perpetual, every eight hours the exchange settles a payment between longs and shorts. When the funding rate is positive, longs pay shorts. A positive funding rate usually means: the long side of the market is overloaded — too many people sitting on the same side of the trade.

Sixteen outflows. One small inflow. Total paid: $494.53.

This is not a large number compared to the eventual loss. But that is not what the funding rate is measuring. Funding is a vote count. The exchange takes a poll every eight hours and reports the answer to me: which side of the market is bigger? If I am the side paying, I am the side overloaded. Sixteen consecutive payments outward is the market telling me, sixteen times in a row, that I am standing in a crowd that is about to be moved.

What happened, in my own head, over those four days was that BTC started ranging. It went sideways. And I waited. The thought that kept running through my head was simple: it will probably go back up.

What I was not registering is that the waiting had a price. That I wasn’t passively holding — I was actively paying. Sixteen funding events in a row, every eight hours, mathematically precise, automatically deducted. I was paying the exchange for the privilege of waiting for a thesis to reverse, and I was not even seeing the receipt.

This is the part of the trade I find hardest to forgive myself for. Not the entry. The entry is one bad morning. Not the size. The size is one bad decision. The funding payments are sixteen separate refusals on my part to look at what was on the screen.

October 10, ~20:00 UTC. The post.

The morning of October 10 looked like the previous mornings. By 16:00 UTC, BTC was at $118,957.80 and my unrealized loss was around –$19,800. Uncomfortable but survivable.

Around 20:00 UTC, Donald Trump posted on Truth Social.

The post announced 100% tariffs on all Chinese imports, effective November 1. Combined tariffs on China would rise to 130%. This came on top of months of escalating US–China tension: China had stopped buying US soybeans in May 2025 (lost market: $12.8 billion), and had imposed rare-earth export controls in late September. Trump had been calling these moves “economically hostile” for weeks.

The post itself was the trigger. But the charge — the energy that the trigger would release — had been accumulating in two parallel structures all week.

The first charge was macro: equities had been rolling over since mid-week, and crypto open interest was at a record. Any meaningful headline was going to produce a meaningful move.

The second charge was the one I did not know about yet. The Binance oracle window — October 6 to October 14 — was now four days in.

October 10, 20:50 UTC. The cascade.

According to forensic analysis from Amberdata later, the cascade entered its violent phase at 20:50 UTC. BTC began moving through $115,000 within minutes. Then $113,000.

Liquidation engines on every major exchange began processing forced closes in size. The sell pressure from forced longs created the next round of liquidations, which created the next. A single Hyperliquid trader — later nicknamed the “Hyperliquid Whale” — had opened a large short minutes before Trump’s post and would close it for approximately $200 million in profit before the cascade ended. Whether that was insider knowledge or coincidence has not been resolved, and may never be.

I was in Hong Kong. By that hour local time it was already very late — Hong Kong is seven hours ahead of UTC, which means the cascade phase began at around 4:50 AM local time for me. I was asleep.

And the last thought I had before I closed my eyes that night was: I still have room to the liquidation price.

That sentence is the most dangerous one in this entire essay. It is the sentence that, in retrospect, contains all five mistakes folded into one. “I still have room” treats the liquidation price as a destination — something distant, something you can navigate toward or away from, something you have time to react to.

The liquidation price is not a destination. It is the moment the math stops working. And it does not care whether you’re awake.

I went to sleep.

October 10, 21:13:04 UTC. The execution.

At 21:13:04 UTC, my position was force-closed.

Market sell of 3.462 BTC at $112,198.50. Closing fee: $388.43. Realized loss: –$44,197.13.

The price did not stop there. By the local low, BTC traded at approximately $103,000–$105,000. If the liquidation engine had not closed me, my loss would have reached roughly $74,000. The forced close, paradoxically, contained the damage.

This is a strange thing to be grateful for. It is also the thing that taught me — viscerally, not intellectually — what the liquidation engine actually is. It is not the exchange punishing you. It is the exchange refusing to keep extending credit on a position whose math has stopped working. It is, in a real sense, the last line of defense against your own thesis.

I learned about the liquidation the way you learn about anything you slept through: from a notification on my phone and an email in the inbox. By Hong Kong morning, it had already been six hours.

The email is a routine, automatic message. Bybit’s liquidation engine sends one every time it closes a position. It contains the trade details, the closing price, the realized P&L. It is written without emotion, because there is no person at the other end of it. It is generated by an automated system, dispassionate, factual.

I cannot describe to you how much I hate that email.

Every time I look at it, I keep feeling that someone sat down and wrote it specifically for me. That someone made the decision to type those exact words and send them, in that exact wording, to me personally. I know, intellectually, that this is not how it works. I know it is a template. I know my email address was one of 1.6 million in that 24-hour window. None of that knowledge changes the feeling.

The email reads like a verdict. And the verdict is addressed to me by name.

In the same twenty-four hours, the market liquidated $19.13 billion across all exchanges. 1.6 million accounts. Nine times the previous record. Eighty-eight percent of the liquidations — $16.7 billion — were longs.

I was a small fraction of one statistic in that number.

October 10, 21:18–22:15 UTC. What was happening on a different exchange.

I was on Bybit. Most of the public story of the cascade — the part that would dominate crypto Twitter for the next four months — happened on Binance.

I want to give this section its own space, because it changes how you read the rest of this essay.

Here is the timeline of what happened on Binance during my own liquidation hour:

• 21:18–21:51 UTC. Binance’s internal asset transfer system slowed down. Transfers between spot, earn, and futures accounts stalled. Some users saw zero balances due to backend timeouts. This started five minutes after my own forced close.

• 21:36–22:15 UTC. USDe, the synthetic stablecoin issued by Ethena Labs, depegged on Binance — and only on Binance. It fell to $0.65 while remaining at $1.00 on Coinbase, Kraken, Bybit, and every other major exchange. wBETH and bnSOL — the two staking derivatives also covered by Binance’s pending oracle update — depegged simultaneously.

• During that window, an estimated $346 million of USDe was liquidated as collateral on Binance, because Binance’s Unified Account valued the depegged USDe using its own internal orderbook, not external oracles.

Forensic analysis from Rena Labs later documented 28 trading anomalies in the orderbook before the USDe depeg — large-volume order volleys, suspicious spoofing-like patterns, what looked like positioning ahead of a known event.

The sequence reads like this: a targeted dump of approximately $60–90 million of USDe on Binance’s internal market crashed its on-exchange price to $0.65. Binance’s collateral valuation system used that internal price. Thousands of leveraged positions on Binance, collateralized by USDe (or by other tokens whose prices were also dislocated in the same window), saw their collateral values fall by 30%+ in minutes. Those positions were force-closed by Binance’s risk engine. The forced selling pushed prices on every exchange — Bybit included — further down.

In forensic terms, a $60 million attack produced $19.3 billion in destruction — an amplification of approximately 322 times. For comparison, the largest previous oracle-manipulation attack on record (Harvest Finance, 2020) involved $24 million in direct losses. The October 2025 event exceeded that by orders of magnitude.

This is what’s being called “the Binance angle” of 10/10.

Four months later, on January 31, 2026, OKX founder Star Xu posted publicly: “No complexity. No accident. 10/10 was caused by irresponsible marketing campaigns by certain companies.” He was naming Binance, and specifically Binance’s September 2025 promotional campaign offering 12% APY on USDe — which had encouraged tens of thousands of traders to swap USDT and USDC into USDe and use it as collateral, accepted on the same terms as traditional stablecoins. Xu called the result a “leverage loop” with “hedge-fund-level structural risks” hidden inside what users believed was stablecoin exposure.

CZ called the accusation FUD. Ethena’s founder Guy Young pushed back too, saying the USDe depeg happened after BTC had already bottomed, and therefore could not have caused the cascade. Chainlink’s Zach Rynes echoed this. Binance’s own post-mortem attributed the chaos to a “liquidity vacuum” — market makers withdrawing bids during extreme volatility — rather than to anything specific to their products.

The dispute is unresolved. There has not been a third-party forensic audit.

What is not disputed:

1. USDe fell to $0.65 only on Binance, while staying at $1.00 everywhere else.

2. The Binance oracle update for USDe, bnSOL, and wBETH had been announced on October 6 and was scheduled for October 14. The crash happened on October 10 — in the middle of that eight-day pre-implementation window.

3. The amplification ratio — the gap between the size of the trigger (~$60M USDe dump) and the size of the destruction ($19.3B in liquidations) — is the largest on record.

4. Binance’s own infrastructure failed during the event: API timeouts, internal transfer delays, asset-balance display errors.

What this means for my trade. I was on Bybit, not Binance. My BTC long was force-closed because BTC moved against me, not because of any oracle manipulation on Binance directly. But the velocity of the BTC move from $118,957 to ~$103,000 — the part of the move that turned a $19,800 unrealized loss into a $44,197 forced close — was accelerated by the Binance cascade. Every forced sell on Binance pressed the BTC price across all venues. The cascade that ate my position was, in origin, a Binance cascade.

This does not exonerate the trade. The position was structurally wrong — wrong size, wrong direction, wrong timing — and it would have been liquidated in any meaningful drawdown of BTC. What the Binance angle changes is the framing: the drawdown that took me out wasn’t ambient market noise. It had an architecture. It had a window. It had someone, on some exchange, who knew exactly which pricing oracles were about to change and had eight days to position for it.

There is a version of this essay where I lean on that fact and call myself a victim of a coordinated event. I’m not going to write that version. The size of my position was my own decision. The funding receipts were my own to ignore. The Binance angle is context, not exoneration.

But the context matters. If you trade leveraged perpetuals, the version of the story where “Bitcoin just dropped” is incomplete. Bitcoin was dropped. And the system that did the dropping is still online, still has the same Unified Account collateral architecture, and is still running yield campaigns on synthetic dollars.

October 11. The morning that was a Saturday.

October 11, 2025 was a Saturday. A day off. I was in Hong Kong. And I was lost.

The amount of money in that liquidation is, for me personally, very significant.

A colleague of mine — she was also in Hong Kong for the trip — wrote that morning: Let’s get breakfast. I met her for coffee. I tried to hold a positive register at the table, in the moment. But I wasn’t hiding how much this hurt.

She finished her coffee and went on with her day. And I went to walk Hong Kong by myself.

I did not know what I was supposed to do with the morning. I did not know how to react. I did not know whether I was supposed to cry, or to tell myself, gently, it’s okay, these things happen.

I walked. That was all I had.

Meanwhile, on the screens I wasn’t watching, Trump and Vice President Vance had softened the rhetoric overnight. “Don’t worry about China, everything will be alright!” Trump posted on Truth Social. BTC was bouncing from the $103k–$105k zone back toward $113,000.

I want to name what this kind of move does to someone who has just been force-liquidated.

The bounce is not relief. The bounce is the harder lesson. Because the bounce gives you a number — a number that says: if you had held one more day, the engine would not have triggered, and you would have been mostly fine. That number is a lie. The engine triggered because the math demanded it. The bounce is what happens on the other side of the math, after the engine has done its work. The two are not the same event.

But the lie is convincing. The lie is convincing because there is a small, irrational voice somewhere in your nervous system that wants the engine to be wrong. That voice has the rest of the day to talk to you. And I had nothing to do with the rest of the day except walk.

October 11, afternoon. The second mistake of the same day.

The day did not end with the liquidation.

I wanted, somehow, to take care of myself. To distract from what was sitting in my chest. So I did what people do when they want to feel like an adult in control of something: I went shopping.

The plan was simple. After the liquidation, I had a small amount left on Bybit. I would withdraw it to my crypto.com account, because crypto.com gives you a card you can spend with directly. A few hundred dollars to walk around with. A small, tangible thing to put in the world that was still mine.

But — as we say in Russian — черт попутал. The devil tripped me up.

In my Bybit address book, I picked the wrong line. Not crypto.com. Coinhako.

The problem with Coinhako is that they had closed my account several months earlier, because I hold a Russian passport. The account existed in the address book as a relic — a registered destination from a previous era, before the geopolitical environment around banking and crypto rails made it impossible.

So I accidentally sent 5,000 USDT to a closed account.

I want you to imagine the moment of watching the transaction hash confirm on-chain and then realizing that you have not sent the money to the place where you live. You have sent it to the place that has formally rejected your right to receive anything there.

A complete disaster. Self-administered. On the same day. On top of the liquidation.

I want to be honest about the emotional state this produced, because the case-study framing makes it sound like a clean second event. It wasn’t. It landed on top of a body that hadn’t processed the first event yet. By the time the wrong-address transaction confirmed, I had pretty much finished myself off for the day.

Naively, I assumed it would just come back. That someone at Coinhako would see a closed-account deposit and send a reverse transaction within the hour. The refund took two months. And it took two months for a reason that is structural, not bureaucratic: blockchain transactions are immutable. The funds sat in a closed-account ledger entry, and reversing them required Coinhako to manually build an off-chain return process for a user they had already off-boarded — a flow that did not, by default, exist.

The lesson here is the one nobody teaches you when you first hear the phrase blockchain is fast and transparent. That phrase is true for the canonical path: right wallet, right network, right coin, active counterparty. The moment you deviate from the canonical path in any direction — wrong network, wrong address, wrong coin, blocked counterparty — the speed-and-transparency promise evaporates completely, and what replaces it is a multi-month manual recovery process running on top of the blockchain, not inside it.

I learned this on October 11, 2025, on top of a fresh $44,000 liquidation, alone in Hong Kong, on a Saturday afternoon.

This is in the past now. But it is still very uncomfortable for me to remember this day.

October 12, 15:17 UTC. The return.

Thirty-eight hours after the liquidation, I opened BUY 0.238 BTC at $113,170.

I want to write something defensive here. I want to write that I was “catching the bounce,” or that “the move was overdone,” or some other phrase traders use to make decisions sound like analysis. That’s not what was happening.

What was happening was: I was sitting in front of a destroyed account, and the impulse to “make it back” was overwhelming. The same instrument, the same direction, eight days after the first entry, two days after the forced close. The mechanism that had just ruined me was, in the same week, what I returned to.

Two days later, on October 14 at 07:01 UTC, I closed it at $112,016.70. Another –$338.64.

This was not a trade. This was emotion paying tuition.

I want to name this precisely, because it is the part of the post-liquidation story that almost no one writes about. The revenge trade is not “an additional mistake.” The revenge trade is the proof that the original mistake was structural. If the entry into BTC at $125,000 had been a one-off error of judgment, the impulse to return would not exist. The impulse to return is the same impulse that produced the entry. It just shows up after the liquidation has stripped away every other rationalization.

You can lie to yourself about why you opened a position. You cannot lie to yourself about why you opened the same position thirty-eight hours after being liquidated out of it.

The pattern.

When I lay October 10 next to Blur, what comes up is not two stories but one. The architecture is identical.

Blur (2023): I shorted BLUR for one day and made $1,699. The next day, I opened a long at $1.35 for $99,464 — on free capital of approximately $80,000. The position bled for 213 days. I closed it at $0.176 for –$87,989.87. Funding worked against me almost the entire time.

BTC (2025): I shorted BNB and BTC for eighteen days and made $33,500. The next day after the run, I opened a long at $124,664.93 for $431,590 — on free capital of approximately $45,000. The position was force-closed in five days at $112,198.50 for –$44,197.13. Funding worked against me sixteen times out of seventeen.

Same architecture. Different surface.

• Both times: a winning short sequence created the chemical state of “understanding the instrument.”

• Both times: the next position was a long, on the same or a related instrument, in the opposite direction.

• Both times: the position size was close to the entire free capital.

• Both times: funding worked against me for the entire holding period.

• Both times: I did not revise the thesis.

• Both times: the exit was not a decision — it was an event.

What I keep coming back to is that this pattern is not a “trading mistake.” It is a behavior with a clean structural signature, repeated under wildly different market conditions, with the same result.

The market did not destroy me twice. I destroyed myself twice. The market provided the moments.

⚠️ Quick interruption.

Case 03 is in the editing stage right now:The position destroyed in a single forced close, with the same structural pattern as Blur and October 10, but on the short side.

Subscribe to get it as soon as it goes live.

👉 Subscribe to The White Box

Five mistakes that make this trade reproducible

Mistake #1: I switched direction after a winning short run, without a chart signal. Eighteen days of profit on shorts ended in a long. There was no setup change. There was no thesis transition. There was the chemical aftertaste of three weeks of being right. The reversal happened in my body, not on the screen.

Mistake #2: I opened a position approximately equal to my entire free capital. $431,590 of notional on roughly $45,000 of capital. Effective leverage near 10x. This wasn’t sizing — it was a binary bet. At that leverage, a 10% move against me ended the account, and BTC moves 10% in a quiet week. In a record-OI week, it does it in a day.

Mistake #3: I ignored sixteen consecutive funding payments going outward. Each payment was a polite, quantitative, public message that I was on the overloaded side of the market. Three payments in a row should have triggered a thesis review. Sixteen in a row is the funding rate begging me to look up.

Mistake #4: I held a position into a known headline-risk window with no plan. Trump–China was the largest active macro story in markets. Open interest on perps was at $217B — the highest ever recorded. I had not asked myself the question what is my plan if a tweet drops? before opening. Not asking it doesn’t make the tweet less likely. It just makes me less prepared when it arrives.

Mistake #5: I returned to the same instrument 38 hours after forced liquidation. A revenge trade on the same instrument, same direction, before any thesis review. This was not “trying to recover capital.” This was the same broken model trying to run again. The additional –$338.64 is not the cost of a trade — it is the cost of a body that hadn’t yet processed what just happened.

Why The White Box exists, and why you should stick around

After Blur, I had roughly $30,000 left. After October 10, $679. After April 17, 2026, $1,287. Three single-position trades destroyed approximately $155,000 against $80,000 of net deposited capital.

Three trades. One structural pattern, manifested three times.

That’s what The White Box is for. Every case I publish breaks down one trade — mine or someone else’s — at the level of detail you just read. Full timestamps. Full math. Full macro context. And then the structural pattern extracted, named, and turned into a rule you can apply to your own positions.

I publish my own first, so I can ask others to share theirs.

👉 Subscribe to The White Box

Other ways to follow:

• 📁 Full case studies, glossary, and reading order: thewhitebox.net

• 🗂️ Structured version of this case with phase tables, the full funding receipt, and macro chronology: thewhitebox.net/#october10

• 💬 Telegram channel for announcements and short reads between cases: @thewhiteboxcrypto

• 📨 To submit your trade for analysis: reach me through any of the above.

The market is the same for all of us. The cognitive traps are the same for all of us. The exchange architecture is the same for all of us. The least we can do is share what they look like up close.

— Sani, founder of The White Box

A few months ago a friend asked me why one altcoin he’d been holding had been delisted from his exchange while another, much smaller one was still trading.

“Is it just liquidity?” he asked.

Not really. The answer is more interesting than that — and more useful. Because once you understand how a listing decision actually gets made, the next altcoin you almost buy will look different. So will the exchange you trade it on. And so will the role you’re being asked to play in someone else’s exit.

This piece is the long version of that answer.

It covers what an exchange’s listing committee actually checks. Why the same listing on Coinbase and on MEXC means structurally different things. Why retail loses money on listings at a higher rate than at almost any other moment in a token’s life. Why 53% of every cryptocurrency ever listed on a tracker is now dead. Ten case studies of tokens that went from huge gains to zero, including one — BLUR — that I lived personally in 2023. Ten influencer-promoted coins that ended up worthless. And a practical 10-minute filter that would have caught most of them.

The piece runs about 15 minutes. It’s the first in a series on how the exchange industry works from the inside.

📖 Reading note: every linked term below opens its full definition + real-world example in The White Box glossary (44 terms, free, no signup). Click any underlined word to go deeper.

1. What an exchange actually looks at

Every serious listing decision starts with the same handful of questions. The depth of the answers — and how willing the exchange is to accept gaps — is what separates one platform’s policy from another’s.

Tokenomics — the first filter

Tokenomics is the economy of the token. Supply schedule, distribution between team and investors and community, vesting periods, burn mechanisms, the actual utility that gives the token a reason to exist.

When a listing committee opens the project’s documentation, the first table they read is the one that says who owns what and when they can sell.

• Supply distribution. If 40% of supply sits with the team and early investors with no vesting cliff, that’s a time bomb. After the listing, those 40% reach the market, and the listing itself becomes the exit event.

• Vesting schedule. When are the unlocks? If a major unlock is scheduled six months after the CEX listing, that’s a structural signal that the listing might be exit liquidity for venture capital, not the start of a long-term holding journey.

• Wallet transparency. Hidden team wallets are an immediate red flag at every tier-1 exchange. Coinbase has stated this publicly: undisclosed founder allocations end the conversation.

• Token utility. Does the token do something real inside the ecosystem? Or is it purely speculative, with utility “coming soon”?

• Inflation and emission. Endless emission without a burn mechanism is a long-term death spiral. The math doesn’t lie, even when the marketing does.

Liquidity and market making

Before a token goes live, the exchange typically requires the project to commit market-making capital — often in the range of $1–5 million. Without this, spreads on day one would be wide enough to trigger user complaints about manipulation. A serious listing is, in part, a logistics operation.

White label — a category most retail traders miss

A white-label exchange runs on another platform’s infrastructure: matching engine, KYC stack, wallet system. The user trades on “Exchange X,” but the engine underneath belongs to “Exchange Y.”

This matters because the white-label partner can have a listing policy that differs from the parent platform’s main brand. The technology is the same; the risk policy is not. A token rejected by the parent brand can appear on the partner venue.

Knowing whether you’re on a white-label exchange changes how you read its listings — and how you assess custody risk.

2. Two philosophies — two different worlds

Two distinct schools of thought have coexisted in this industry for years. Neither is “better.” They are different bets on what users want, and they map directly onto what’s loosely called the exchange tiers — tier-1, tier-2, tier-3, classified by volume, regulation, and operational history.

Philosophy A: Slow, regulated, narrow

Coinbase, Kraken, Gemini, Bitstamp, and (more cautiously) Binance.US.

The pitch: long due diligence, US/EU regulatory standing, deep institutional liquidity, conservative listing pace. Coinbase has listed roughly 330 coins across its entire history. Each one went through months of legal, compliance, and technical review.

The cost: you don’t get exposure to early-stage tokens. By the time something appears on Coinbase, it’s often already had its biggest move on smaller venues.

Philosophy B: Fast, wide, fewer filters

MEXC, Gate.io, LBank, Bitget, HTX.

The pitch: thousands of listings, fast-track for memecoins, exposure to projects you literally won’t find anywhere else. MEXC alone has listed over 3,000 tokens and at peak added 214 new tokens in a single month in 2025. Many appear in pre-market trading before the formal launch.

The cost: due diligence depth varies dramatically per listing. A memecoin can go from anonymous Telegram to live trading in days. The risk of buying a token that will be delisted in 90 days is materially higher.

Neither philosophy is wrong. They serve different traders. The mistake is assuming the safety profile is the same.

3. Why a listing is the #1 hype event for retail

Listings work like a stimulant. Here’s why.

The psychology of “moonshots in the moment.” When a token lists on a major CEX, price often makes 2–10x in the first hours. Occasionally 100x. A trader watching this on Twitter develops one reflex: the next one is mine. FOMO is the single strongest driving mechanism in this market.

Most people miss that the 100x was visible only to early holders who received the token in a private round at $0.001. The retail buyer enters at the listing price or higher and frequently exits at a loss within days.

The illusion of validation. A CEX listing reads in the back of the brain as the exchange checked it, so it’s safe. This cognitive bias works on experienced traders too. The reality: the exchange’s check protects the exchange — from regulators, from reputational risk — but does not guarantee the trader’s profit.

Social proof. The hype around a listing is collective excitement in Telegram, Twitter Spaces, Discord. When thousands of people talk about the same token at the same time, the brain reads it as this is an important event. In fact, what happened is one new row in a listing announcement.

The marketing machine. The exchange is itself interested in the hype. Listings drive volume, new sign-ups, and commissions. So launches come packaged with launchpools, airdrops, and trading competitions. This isn’t sinister — it’s how the business works. But the trader needs to remember: the exchange’s marketing budget and the trader’s P&L are not the same KPI.

4. The math of survival

The numbers are the sobering part of this industry.

Per CoinGecko’s December 2025 report:

• 53.2% of all cryptocurrencies ever listed on GeckoTerminal are dead. Defined as: trading volume under $1,000 over a three-month rolling window, plus inactive social channels and an unreachable team.

• Q1 2025 alone added 1.8 million dead tokens to the count.

• Of the dead: 93% had low volume; 65.5% were abandoned; 58% had inactive Twitter and websites.

• Average lifespan of a “dead” project: 2.21 years.

This is not the tail of the distribution. This is the majority outcome.

When you evaluate any new project, the question is not “will it succeed?” The question is “what’s the structural reason it would not become a dead coin within two years?”

Most projects fail the second question.

5. Ten cases from x’s to zero

These aren’t stories about “bad projects.” They’re patterns that repeat every cycle.

1. LUNA / UST. Peak $119.18 (April 5, 2022). By June 6 — $0.00008. Supply grew from 343M to 6.53 trillion in a week. Hyperinflation of +1,908,651%. Context: the Fed raised rates from 0.25% to 1.00% on May 4, 2022.

2. FTT — the exchange token domino. Peak $84.18 (September 2021). After CZ’s tweet about selling FTT — from $22 to $5 in 24 hours. Today $0.28. −99.6%.

3. SQUID — the textbook rug pull. Peak $2,861.80 (November 1, 2021). +23,000,000% in 6 days. In 5 minutes the developers drained $3.38M in liquidity. An anti-dump function in the smart contract made selling impossible by design. The canonical example.

4. BitConnect (BCC) — the classic Ponzi scheme. Peak $463 (December 2017). Collapse in January 2018 after cease-and-desist orders from Texas and North Carolina. 1% daily return through a mythical “trading bot.” Founder charged by the SEC with a $2.4B fraud.

5. SafeMoon. Peak $0.0000141 (April 2021). +20,000x in a month. Today $0.00000003. SEC charged the founders with fraud in November 2023.

6. ICP — the most expensive listing in history. Listed on Coinbase May 10, 2021, opening at $700+. Top-10 by market cap immediately. Today $4–5. −99.3%. Not a scam — just tokenomics with massive unlocks.

7. AXS (Axie Infinity) — the P2E dream. Peak $164.90 (November 2021). Today ~$2. −98%. The Ronin Bridge hack in March 2022 — $625M stolen by Lazarus Group, the same actor that would steal $1.5B from Bybit in February 2025.

8. STEPN / GMT — the move-to-earn bubble. Peak $4.11 (April 2022). Today ~$0.05. −98%. Retail bought NFT sneakers for $1,000+ to farm GST. The mechanic held up exactly as long as new shoe buyers kept arriving.

9. BLUR — my own case. Peak $1.35 (February 2023). Floor $0.18 (September 2023). −86%. An NFT marketplace trying to take share from OpenSea through token incentives. The vampire attack worked tactically, but the token price didn’t survive it. I held a long position for 213 days. Zero days in profit. Realized loss: $87,989.87. The full breakdown lives on The White Box as Case 01 — see the full breakdown.

10. Friend.tech (FRIEND) — fresh social-fi case. Peak $4.50 (May 2024). In August 2024 the founders “stepped back,” transferring smart-contract ownership. Effectively a soft rug. Today ~$0.01. A clean example of how abandonment works slower than a hard rug — but no less effectively.

6. Famous names ≠ truthful information

The most dangerous category of listings in the last two years are tokens promoted by public figures.

The pattern is almost always the same: an influencer posts, the token pumps to peak in minutes or hours, insiders (often with pre-launch access to the contract) extract millions through one-sided liquidity pools, the price collapses 90–99%. Retail, trusting the famous name, is left with empty wallets.

Popularity ≠ truth. A large audience amplifies not the quality of information, but the speed of its distribution. In crypto, that translates directly into the speed at which retail loses capital.

Ten cases of influencer coins that ended up worthless:

1. $LIBRA — Javier Milei (President of Argentina). February 14, 2025. Milei’s post on X drove market cap to $4.5B within hours. Insiders extracted $107M through one-sided liquidity pools on Meteora. Drop of −94% in hours. Per Nansen, 86% of traders lost a combined $251M. Parliamentary investigation, account freezes, Milei deleted the tweet.

2. $HAWK — Hailey Welch (”Hawk Tuah girl”). December 4, 2024. Solana launch. Market cap climbed to $490–500M and collapsed to $60M in 20 minutes. 96% of supply in clustered wallets (a textbook signal that would have been caught by basic wash trading analysis). Welch was paid $125,000 to promote it.

3. $TRUMP — Donald Trump. January 17, 2025, three days before inauguration. Official memecoin of the US president. Peak market cap around $14B, steady decline since. 80% of supply held by the Trump entities on a scheduled unlock plan.

4. $MELANIA — Melania Trump. January 19, 2025. 48 hours after $TRUMP. Same playbook, same design team. Direct cannibalization of attention and capital. −90%+ from peak.

5. $EMAX (EthereumMax) — Kim Kardashian. June 2021. Instagram promo to 225M followers for $250,000. Peak at $0.00000092 in May 2021. Today a fraction of one four-billionth of a cent. −99%+. SEC fined Kardashian $1.26M in October 2022 for failing to disclose the paid nature of the post.

6. $CTR (Centra Tech) — Floyd Mayweather + DJ Khaled. 2017–2018. ICO raised $32M. Instagram promo by Mayweather and Khaled. False Visa/Mastercard partnership claims. SEC fraud charges; founders received prison time.

7. CryptoZoo — Logan Paul. 2021–2022. NFT “collect hybrid animals” game from YouTuber Logan Paul. The game never launched. Coffeezilla’s investigation showed developers received payments and disappeared. Paul announced a “refund” program a year later, after public backlash.

8. Save the Kids — FaZe Clan. June 2021. “Charitable” token promoted by FaZe members. In reality a pump-and-dump with vesting that allowed promoters to sell at peak hype. FaZe Clan removed the implicated members. Token went to zero in days.

9. $JENNER — Caitlyn Jenner. May 2024. Pump.fun token Jenner aggressively promoted on X. It later emerged she was working with a known crypto promoter linked to prior scam projects. −95%+ within days.

10. $MOTHER — Iggy Azalea. May 2024. Meme token on Solana from rapper Iggy Azalea. Peak market cap around $250M. Despite Azalea’s genuine involvement, the tokenomics and dependency on one person’s attention made the decline inevitable.

What all ten have in common:

• One contract, one post, one peak. Hype concentrates in a single moment and has no renewal mechanism.

• Supply concentration. 80–96% of tokens sit with the team and insiders. Not the community.

• No vesting cliff. Insiders can sell immediately.

• Paid promotion without disclosure. The audience doesn’t know the influencer was paid, not sharing a genuine recommendation.

• Tier-3 venues or DEX only. Tier-1 exchanges don’t list these tokens, because the filters in Section 1 catch them immediately.

This isn’t a call to distrust influencers. It’s a call to separate the source of attention from the source of truth. A large audience knows how to distribute information. It doesn’t make that information accurate, verified, or aligned with the audience’s interests.

7. What this means in practice

Most of what’s described above, nobody actually checks. I didn’t check it either when I bought my first altcoins. So this section is not a “hard checklist of 20 items,” but a realistic minimum: what I do now when I see the next round of hype around a new coin.

When you see hype — the 10-minute minimum

You don’t need to become an on-chain analyst. You need three things, which take less time than a cup of coffee.

1. Check top holders on bubblemaps.io. Enter the ticker, see visually how concentrated supply is. If 80%+ sits in a single cluster of related wallets, that’s all you need to know. This is how $HAWK (96% in a cluster) and $LIBRA (82%) were exposed.

2. Find the vesting schedule on tokenunlocks.app or cryptorank.io. When is the next major unlock? If it’s within 30 days, you’re being sold someone’s locked allocation, and you’re buying exit liquidity. If there’s no vesting at all, insiders can dump from the first second.

3. Search the ticker on YouTube + Coffeezilla, ZachXBT on X. If either has already covered the project, that’s your last chance to read the breakdown before your purchase. Takes 2 minutes.

That’s it. Not perfect — but in 90% of cases one of the three checks is enough to stop or proceed consciously.

Signals that should make you pause

Any one alone is a reason to stop and think. Several together is a reason not to open the trade:

• Anonymous team

• Influencer hype + X account created less than 60 days ago

• 15%+ APY promised without a clear source of returns (the classic Ponzi signature)

• Pre-sale “fills in seconds” (those are sniper bots, not real demand)

• Audit from an unknown firm, or “audit coming soon”

• Project lives only on tier-3 exchanges or DEX for 18+ months

• Volume pattern that looks like wash trading on tier-2/3 venues

If you decide to enter — three things that matter

Let’s say the checks pass, the risk is understood, you’re going in. Three practical things separate a position from a casino bet:

• Position sizing. Altcoin — no more than 2% of portfolio. Any coin other than BTC/ETH — no more than 5%. This isn’t a “cautious” approach, this is basic survival math: one −80% position at 10% of portfolio takes 8% of your capital. A few of those in a row and you’re out of the game.

• Stop-loss BEFORE entry. Not “I’ll watch it.” A concrete price, placed as a limit order immediately after opening. In a market that moves 20% per hour, a manual decision arrives too late.

• Write down why you’re buying. One sentence: “I’m buying X at price Y because [reason]. I’ll exit on [condition Z] or after [N] days.” If you can’t write it, you don’t have a thesis — you have an emotion. Emotions don’t work in markets.

This won’t make every trade a winner. It will eliminate the catastrophic losses that are psychologically hard to come back from. Most traders who lose in crypto don’t lose a series of small mistakes. They lose one big one, on which they violate all three rules at once — exactly the pattern I broke down in Case 01: Seven months in BLUR.

Why this piece exists

The White Box is post-trade analytics for crypto traders. The project’s premise is simple: we share real losses, broken down with dates, sizes, and the specific mistakes that produced them — so that other traders can shorten the feedback loop between making a mistake and learning from it.

This article is the framework. Case 01 — my own $87,989.87 in BLUR over 213 days — is the application of the framework to a single trade.

If you’ve been carrying a position like this, write back. Trade analyses go up anonymized or not, your call. We publish the math, the dates, and the mistakes. Not the names.

Continue on the site

→ Full interactive article with table of contents and reverse-linked glossary

→ Case 01: Seven months in BLUR — 213 days, $87,989.87, four mistakes

→ Full glossary of 44 terms — every concept used in The White Box analyses

→ Telegram channel — short reads, announcements, and early access to new cases

— Sani, founder of The White Box

That’s what I lost on a single trade in 2023. Not in a flash crash. Not in a liquidation cascade. Over 213 days of consciously holding a position that was wrong from day one. Zero days in profit. One exit window I refused to take. A Sunday morning in September that I still feel in my hands.

This is the most expensive trade in my career, broken down hour by hour, news cycle by news cycle, lie-to-myself by lie-to-myself.

It is also Case 01 of The White Box — and the reason I’m starting the project with my own worst trade is that I think the crypto industry has a structural blind spot. We have an entire genre of content built around how someone caught the next 100x. We have almost nothing about losses that are analyzed. Confessed, sometimes. Analyzed, rarely.

I’m going to fix that, starting here.

🔍 New here?

The White Box is post-trade analytics for crypto traders. I take real trades — mine, and submissions from other traders — and break them down with the full math, the exact timestamps, the macro events that hit during the hold, and the cognitive patterns that produced the decision. Subscribe to get every case study in your inbox, plus the framework I’m building from each one.

👉 Subscribe (free)

Why I’m writing this now

The industry is saturated with how-someone-made-it stories. We get the chart, the entry, the exit, the lifestyle photo. What we don’t get is a clean look at how the same brains, the same accounts, the same tools also produce the trades that lose you a house deposit in 213 days.

I want to build the library that’s missing. A reference of real losses, with the math shown, the dates pinned, the macro context laid out next to the emotional state. Not a confession booth. A specimen collection.

I can’t promise that publishing this analysis will make me stop making these mistakes. They’re built into how human cognition meets leverage. But I can promise this: naming the mistakes precisely makes them rarer. And the way you get other traders to share their pain is by going first.

So I’m going first.

Three positions in my own account destroyed approximately $155,000 between 2023 and 2026.

Blur is the first one. The other two — long BTC at $124,664 on October 10, 2025, and short BTC at $69,793 on April 17, 2026 — are being written up next. More on those at the end of this piece, and in the cases that follow.

February 14, 2023. The setup.

To understand why I entered, you have to remember what the NFT sector looked like that month.

OpenSea, the long-running monopolist, was losing share. Blur — a marketplace built for professional flippers, with zero fees, aggregated listings, and instant bidding — had already overtaken OpenSea in trading volume by early 2023. This was not “another marketplace.” This was the takeover of the segment.

On February 14, Blur ran its Season 1 airdrop: 360 million tokens distributed to active users. Some wallets received 128,000 BLUR each. In the first hours after Coinbase listed it, BLUR traded above $5. By the end of the same day, below $1. The recipients were dumping.

I watched the cycle from the side. And I saw in it what I wanted to see: the token was oversold, the fundamentals were strong, the platform actually worked, and a bounce was inevitable.

February 19, 2023. The trade before the trade.

Here’s what most people miss when they tell a story like this: the catastrophic trade almost never happens in isolation. It happens right after a winning trade.

It was a Saturday. My husband was getting ready — we were going out somewhere. While he was organizing the morning, I had some time alone with the screen, and I made a long position on a different token that printed money. Around ten thousand dollars, give or take. I don’t remember the exact figure now, which itself is a tell — that win didn’t matter compared to what came next.

What it gave me was not money. It gave me a chemical state. Euphoria. The conviction that I could read the tape. The conviction that the market was about to keep going, and that I was sitting in the right chair at the right moment. And it was going. Which made the feeling worse, because the market kept confirming me in real time. My body chemistry was firing. I felt sharp. I felt right.

The trade I’m about to tell you about is what happened when I tried to scale that feeling.

February 19, 2023, 06:32 UTC. The first short.

I opened a small short. 50,000 BLUR at $1.366. I closed it the same day at $1.40. +$1,699.

This trade matters, but not for the size. It matters because it gave me a false confirmation that I understood Blur’s price action. I had now made money on BLUR twice — once on the warmup token, once on this short. Three for three on the morning.

I wasn’t focused. My husband was almost ready. We were about to walk out the door. I wanted one more — one bigger one — before we left. Pure emotion. No thesis review. Just the chemical state of a body that had been right twice in a row and wanted to be right a third time, faster.

So I clicked.

The long entry. The decision I wish had a pause.

The same day, a few hours later, I opened a long position on 73,677 BLUR at $1.350.

The notional value of that position was $99,464. My free USDT on Bybit at that moment was around $80,000, plus about $7,000 in BIT tokens. The position was larger than my entire portfolio. With leverage, obviously.

I want to be specific about what happened in my head when I clicked, because I think it’s the most universally recognizable moment in this whole essay. Other traders will read this and recognize themselves.

I did not set a stop-loss. I did not even consider one. It wasn’t that I considered one and rejected it — it never crossed my mind. There was no internal voice asking “what if this goes against me.” There was no pause before clicking. There was no five-minute window where I sat back and asked myself whether I had actually done research on this token, or whether I was just riding the emotional wave from the morning’s wins.

I had heard, somewhere — in some Telegram chat, on some Twitter feed — that Blur was a “good project” that would “do Xs.” That was the depth of my research. That was the entire substrate of a hundred-thousand-dollar position.

And the most terrifying part isn’t that I didn’t research. It’s that I didn’t wait. There was no pause. The reasoning, to the extent I had any, was: the token is still active, it’s still popular, if I don’t move now it might dump further. So move now.

This is the part of trading that doesn’t get talked about enough. People analyze entries as if they’re decisions. Sometimes they aren’t. Sometimes they’re the trailing edge of a chemical state that started two trades ago.

Side note: this exact pattern — catastrophic position opened immediately after a winning one, with no pause and no stop — is what produced both of my next two disasters too. By the time we get to October 2025 and April 2026, the structure is identical. Same body, same morning, different token. Subscribers will get the full breakdowns as they go up. — Subscribe to follow the series.

February 19–22. The first 72 hours.

The price moved in exactly the opposite direction of my thesis.

• February 19 close: $1.32 (–2.4% from entry)

• February 20: $1.25 (–7%)

• February 21: $1.18 (–13%)

• February 22: $0.98 (–27%)

In three days the position was down $27,000. And right here — at the breaking of the psychological $1 level — I broke a rule I knew theoretically: I did not exit when the thesis failed to confirm in the first 24 hours.

What happened in my head instead was the first lie. And it had a very specific structure: the market will reverse. The price will reverse. The loss is too big to close at.

That last sentence is the trap. Not “the loss is too big” — that’s just emotion. The trap is the word “at.” “Too big to close at“ implies that there exists a future moment when the same loss won’t be too big to close. And the only way that future moment exists is if the price comes back. Which means I had quietly converted my trading position into a prediction about the future, and made my own future emotional comfort depend on that prediction being true.

This is sunk cost fallacy, but it never arrives in your head wearing its own name. It arrives as a reasonable-sounding argument about fundamentals.

March 10–13. SVB and USDC.

On March 10, 2023, Silicon Valley Bank collapsed. On March 11, Signature Bank followed. SVB held $3.3 billion of Circle’s reserves backing USDC. The market figured this out over the weekend, and USDC depegged to $0.88.

Crypto entered systemic panic. Altcoins fell two to three times harder than BTC. BLUR dropped from $0.78 to $0.55 across those days.

My position was now down $58,500. Three-quarters of my original $80k portfolio, gone in display value.

This is roughly one month into the trade. I am sitting in an enormous loss. And I am not allowing myself to close, because I am still — still — assuming the market will reverse.

I want to be honest about this part: it is genuinely stupid. You can say you’ve never seen a more idiotic way to trade. You’d be right. But this is a real, lived case from my own practice, and the reason I’m describing it in detail is that I think it’s the most common form of losing trade — far more common than the dramatic blow-up. The slow-bleed bag-hold, where the trader is fully aware the position is wrong but is paralyzed by the size of the loss.

I knew I still had room to liquidation. The liquidation price was far away. So the displayed minus number scared me, but not enough to act. I told myself: I cannot close at this loss. I wasn’t waiting for profit anymore. I was waiting for less minus.

And I didn’t tell anyone. Not friends, not other traders, not my husband in any detail. I was ashamed. And shame is what closes off the optionality you actually need in moments like this — because there were other things I could have done. I could have hedged the position. I could have opened an offsetting short on a correlated asset. I could have at least sized down. But every cell of my attention was locked on the underwater position itself, on the question of “how do I get out of this drowning,” which is the wrong question, because drowning is not a problem you solve by staring at the water.

By March 13, the U.S. Treasury announced it would protect SVB depositors. USDC repegged. Altcoins bounced. BLUR recovered to $0.67. My loss narrowed from –$58k to roughly –$50k.

And that, in some way, was worse than if BLUR had stayed at $0.55. Because the bounce gave me a new argument: the fundamentals are holding, the panic is past, now we catch up.

April 14–19. The only spike of hope.

On April 12, Ethereum executed the Shanghai hardfork — stakers could finally withdraw staked ETH. The market had braced for a massive sell-off. It didn’t come. ETH rallied.

In parallel, Blur announced its Season 2 airdrop. On April 14, BLUR began moving: $0.59 → April 17 $0.71 → April 19 $0.78.

That’s a 32% rally in five days. My position moved from –$55k to –$42k.

And here is what I did with that information. Instead of treating it as an exit window, I treated it as proof the trend was reversing. I told myself: look, the minus is getting smaller. If it went from –54 to –42, it can go from –42 to –30, and from –30 to –15, and from –15 to zero. All I have to do is wait.

This is exactly the same internal mechanism as in February, just with smaller numbers. The loss is getting smaller, so I should wait longer. I was now in the second month of the trade and I had still not asked myself the only question that mattered: if I had clean cash right now, would I open this position at the current price? If the answer is no, you are not holding an investment. You are bag-holding. And the longer you hold, the more expensive your refusal to ask that question becomes.

On April 20, BLUR turned over. By April 24: $0.59. By May 1: $0.68. By May 14: $0.48.

The window closed. I did not use it.

⚠️ Quick interruption.

If you’re reading this far, you already understand what this kind of breakdown does — both as analysis and as protection. Case 02 is in the editing stage right now: October 10, 2025 — long BTC liquidated at the peak of crypto’s largest deleveraging event in history ($19B wiped in a single cascade). Case 03, after that: April 17, 2026 — short BTC squeezed during the Iran ceasefire rally. Both are bigger losses than Blur. Both follow the exact same structural pattern you just read about.

You can subscribe to get them as soon as they go live. Free.

👉 Subscribe to The White Box

May–July. Numbness.

This period is harder to describe because it had no events. Just slow attrition. $0.48 → $0.45 → $0.40 → $0.35 → $0.31.

But my life outside the screen had events. I was leaving one company and joining another. Job transitions take a specific kind of emotional energy — they demand focus, optimism, the ability to walk into a new room and present a version of yourself that is forward-leaning, capable, hopeful. And I was carrying this position the entire time. Like a stone in my pocket that I couldn’t put down because putting it down would have meant closing the trade and admitting the number.

The number was the thing. To earn that amount of money — the kind of money I was watching evaporate in display — would take many months of work. I was going to walk into a new role and start from a balance that didn’t reflect what I had earned. It reflected what I had given back, voluntarily, through one stupid clicked button on a Saturday morning. There was anger in it. There was apathy in it. And there was the very specific weight of knowing you carry this thing alone, because you haven’t told anyone.

What kept me in the position through May, June, July was a sentence I heard everywhere in crypto Twitter that summer: “the bull run is coming.” People were predicting it. Many were saying Blur specifically was a quality project. I kept believing. I kept hoping the price would be better.

On June 15, BlackRock filed for a spot Bitcoin ETF. The crypto narrative turned. Wall Street had arrived. BTC began to rally. ETH began to rally. Major altcoins began to rally.

BLUR did not.

It went from $0.40 to $0.36 to $0.32. While the entire market was greening, my token stayed red. This was the signal I could no longer ignore: the problem is not the market. The problem is the token.

But “knowing the trade is lost” and “closing the trade” are two different acts, separated by an emotional layer most people underestimate. I had been knowing the trade was lost for months. I was still in it.

The hope was no longer about reversal. The hope had shrunk to a smaller, sadder shape: maybe the negative balance will be a little less when I finally close. Blur kept making announcements. Each announcement gave me 24 hours of artificial hope. Then the price kept doing what the price was doing.

August. The acceptance.

By August, the hope was gone. Fully gone.

I understood it was a complete write-off. I had no more clever ideas, no more strategies, no more reframes. There was no more “maybe if I hedge” or “maybe if I wait for an announcement.” There was just the position, sitting there at $0.20, and me, watching it.

I was waiting for liquidation. That was the plan. The plan was: do nothing, let the engine close it for me, accept the cleanup. Because I genuinely felt there was nothing left to save.

This is a state I want to name precisely, because I don’t think it’s discussed enough. It’s not panic. It’s not denial. It’s not even depression in the clinical sense. It’s a kind of strategic exhaustion — when the cognitive cost of evaluating new options exceeds your remaining capacity, and you simply hand the decision to the machine.

Liquidation, in this state, feels like relief.

September 10, 2023. The Sunday.

Sunday again. A day off. Morning.

I understood there was no point in waiting anymore. I washed my face. I made coffee. I sat down at the screen, and I closed the position myself.

73,677 BLUR at $0.176. Realized loss: –$87,989.87.

I didn’t want liquidation. I wanted to take the action. I understood the size of the loss and I wanted my hands on it, not the engine’s.

What happened next is the part of the trade that doesn’t have numbers.

My legs and hands went numb. I literally couldn’t speak for a stretch of time — I don’t know how long, maybe a minute, maybe more. The apartment around me was empty. There was no one there. But I felt a specific kind of deaf silence settle on the room, the kind of silence that has weight, that you can almost hear.

Then I stood up.

And I said to myself: “This is the best and harshest lesson you’ve ever taken. This is all I’ll remember from this period.”

That’s true. Almost three years later, that’s what I remember.

Four mistakes that make this trade reproducible

Two weeks after I closed, BTC began the BlackRock rally that would carry it to $48k by January 2024 and to $73k by March. The crypto winter ended. BLUR did not go anywhere. At the time I’m writing this (May 2026), BLUR trades below $0.10. My 73,677 tokens would be worth less than $7,000 today.

This is not a story about bad timing. It’s a story about four specific mistakes — each one recognizable to anyone who has traded altcoins. The structural breakdown with the rule derived from each mistake lives on the case page. Here, in short:

Mistake #1: I opened a position larger than my entire portfolio. This wasn’t trading. It was a binary bet. The moment I sized that way, I had already eliminated stop-loss as an option, because any stop at that size would have been catastrophic.

Mistake #2: I mistook a counter-trend bet for a contrarian one. Contrarianism works against majority emotion, not against majority facts. The facts at entry were screaming: airdrop recipients were dumping, no vertical buyer was visible. That’s a slowly burning fuse, not a contrarian setup.

Mistake #3: I refused to exit in April. When the market handed me an exit at –$42k instead of the –$88k it eventually became, I reframed it as confirmation of my thesis. This is the most expensive of the four — because the first three I could plausibly blame on inexperience. This one I can only blame on my refusal to admit the first three.

Mistake #4: I held a position for 213 days without re-validating the thesis. I never sat down and asked: if I had clean cash right now, would I open this position at current price? If the answer is no, you are bag-holding, not investing.

Why The White Box exists, and why you should stick around

After I closed Blur, I had roughly $30k left. By June 2024, $679. I went on to enter two more catastrophic single-position trades — long BTC at the ATH on October 10, 2025 (during the $19 billion liquidation cascade), and short BTC during the April 17, 2026 Iran ceasefire short squeeze. Together, those three positions destroyed approximately $155,000 against $80,000 of net deposited capital.

Three trades. One structural pattern, manifested three times.

That’s what The White Box is for. Every case I publish breaks down one trade — mine or someone else’s — at the level of detail you just read. Full timestamps. Full math. Full macro context. And then the structural pattern extracted, named, and turned into a rule you can apply to your own positions.

👉 Subscribe to The White Box

Other ways to follow:

• 🌐 Full case studies, glossary, and reading order: thewhitebox.net

• 📊 Structured version of this case with phase tables and macro chronology: thewhitebox.net/cases/blur-2023

• 💬 Telegram channel for announcements and short reads between cases: @thewhiteboxcrypto

• ✉️ To submit your trade for analysis: reach me through any of the above.

The market is the same for all of us. The cognitive traps are the same for all of us. The least we can do is share what they look like up close.

— Sani, founder of The White Box