The ledger and the lesson

Every essay before this one theorises. This one is the dive itself, with the gauge readings attached. Between April and June I ran a real account through one melt-up and one rollover, and I kept the screenshots. What follows is reconstructed from them. The numbers are real. The voice is the story I have fabricated around the numbers, which is the only honest way to describe what I was actually thinking at the time.

I am publishing it because the whole site argues that a rule has to be written before the stakes arrive. This is the other half of the argument. This is what it looks like when the stakes arrive and the rule is tested in public, by me, on my own account.

It started with conviction in one name. On April 23 the first real lot went in: twenty shares of ServiceNow at $86.79. Not a starter position, a statement. ServiceNow was going to be the anchor of the whole portfolio.

Through early May the accumulation continued, methodical and a little impatient. May 4 brought a cluster of NOW buys around $91.25, twenty more shares split across five fills, the kind of laddering you do when you have decided and you are just getting the size on. May 8 and May 12 added fifteen more each near $90. The average cost was sitting comfortably in the low nineties.

Underneath the headline position, the rest of the book took shape: SoFi at an average of $16.02, a small Microsoft holding at $420.65, a growing Robinhood stake averaging $80.14, fourteen shares of Palantir at $137.04, and a small quantum-computing flyer in IonQ. A six-name book with one clear king.

Then on May 22 something shifted. NOW got bought again, but this time at $102.72. That was not accumulation anymore. That was chasing strength. The stock was already running, and the position was being added to into the move. This is the exact mistake the margin essay is about. I wrote the rule, then I reached into the margin anyway.

Everything caught fire at once. The four-hour charts all tell the same story: a near-vertical spike cresting around May 29 to June 1. ServiceNow ran to roughly $140. Microsoft touched $470. Palantir nearly $163. SoFi tagged $19. Robinhood pushed $96. For one stretch every single line was green.

The account caught the peak in real time. A +5.31% day, up about fifteen hundred dollars in a few hours. The week before that read +28.61%. Twenty-eight percent. In a week.

And on the quiet Sunday evening of May 31, a long look at the ServiceNow lots: 115 shares, up $4,058, up 37.73%. The high-water mark, glowing on the screen at 8:48pm. It felt less like a position and more like proof.

It felt like proof. It was weather.

The floor went out from under it the way it always does, suddenly, on a day that started ordinary. A −5.92% session, most of a week's gain gone red in an afternoon. ServiceNow, so recently at $140, printed $124.94 and kept falling. The euphoria inverted into something colder.

This was also the day the flyer got cut. IonQ, twenty shares, the speculative corner of the book, had been sitting at −$35.20. By the next screen it was simply gone. A small, decisive loss, taken without ceremony. The first real discipline of the whole journey: the willingness to be wrong cheaply.

Through the first half of June the charts rolled over hard and in unison. NOW 140 to 103. MSFT 470 to 392. PLTR 163 to 131. SoFi 19 to 16.50. The melt-up ran in reverse, frame by frame.

By midday on June 12 the damage was tallied position by position. ServiceNow was still standing, still green, up 8.02% on an average of $96.05, and now 52.57% of the entire portfolio. The king had given back a fortune in paper gains and was still the only thing holding the book up. Microsoft was underwater at −6.62%. Palantir, bought into strength, was paying for it at −4.11%. SoFi, the boring hold that never asked for attention, was quietly green. Robinhood, bought and added in fractions, was the quiet winner at +14.4%.

The full ledger, as it stood that afternoon:

And then the two screens that frame the whole thing. The same portfolio. The same afternoon. Two numbers that disagree completely, and both true.

The story is cleaner in the present tense, before I knew how it turned out. These are the entries as they stood on the day.

Here is what I actually learned, written down so I cannot pretend later that I knew it all along.

1. The melt-up was not skill. I felt like a genius at +28% in a week. I was the same person at −14% three weeks later. The market handed me the win and took most of it back, and my conviction was along for the ride both ways. This is the unpunished violation in reverse: a reward I did not earn, teaching me a lesson I should not learn.
2. ServiceNow is 52% of my book. That is not a portfolio, that is a bet with garnish. It saved me this month and it is the single thing that can wreck me next month. Concentration cuts in exactly two directions and I have only felt one of them so far.
3. Cutting IonQ was the best trade I made. Not because of the thirty-five dollars. Because it is the only time I let a position be wrong without negotiating with it.
4. The screen that matters is the one I am least tempted to look at. The three-month chart, not the daily P&L. The boring timeframe is the safety stop.

The story is not over. But the first chapter has a clear moral, and it is not that I am good at this. It is the same instruction the gauge gives a diver who is having too good a time at depth.

Zoom out, size down, and respect the part of the chart you cannot see yet.

Personal record, published as practice, not advice. No position is a recommendation, no figure is a forecast, and nothing here should be copied. Prices and percentages are as recorded on the dates shown; the account identifier and total balance are withheld on purpose. Carrying this into the skill pages on the markets side as it develops.