Eleven minutes and forty seconds, start to finish. That is how long the repair took on July 18, after eight months of screenshots in which my portfolio pie sat obediently at 50-25-15-10, month after month, exactly on target. The pie was lying the entire time. I had been photographing the wrong tab, and nothin' about the lie announced itself cuz the chart kept all of its colors.
$23,700 total on paper, split into 50 percent bitcoin, 25 percent ethereum, 15 percent stablecoin, 10 percent cash — the allocation I set in November after selling a car and refusing to speculate with the proceeds. A portfolio that size does not get a financial advisor. It gets a spreadsheet, a calendar reminder, and whatev the app's allocation view claims is true. I gave the view my trust for free, and it spent eight months spending it.
The actual mix, counted by hand in July: 71-14-9-6. Bitcoin ran thru spring, every weekly buy fed the biggest slice, and a December app update had quietly changed which tab opens by default — the target pie instead of the current pie. Eight monthly screenshots of a promise instead of a fact. The correction took less time than boiling rice, and the lesson cost zero dollars, which is the only kind I respect.
the changelog line I never read
December 4, app version 6.2, pushed overnight. One line in the release notes said refined allocation view, and I skimmed past it at 7 am with a train to catch. The allocation screen kept its colors, its donut, its familiar spot on the screen, and swapped its default toggle from CURRENT to TARGET without so much as a tooltip. Nothin' else moved. A chart answering a different question, in the same spot, with the same colors.
january first, same perfect donut
January 1, 9 am, coffee. I took my monthly screenshot of the allocation pie for the folder named PORTFOLIO-MONTHLY. The donut sat at 50-25-15-10 in four neat wedges. I archived it and closed the app, satisfied, and the satisfaction is the part I keep comin' back to. Seven months in a row, the pie would go on showing the plan, and every month I would read the plan as a report, which is the sentence I now say out loud when an app shows me a chart I did not verify.
seven screenshots, one folder
January thru July: seven images, each showing identical weights, each filed under its month. Each image identical to the last. Identical numbers should have been the tell, cuz a live portfolio never sits still that precisely — fees alone would nick the second decimal place, every single month, without asking permission. I missed the tell cuz identical numbers feel like discipline, and discipline is exactly what I believed I was looking at. The folder now reads as seven months of documentation with the fact edited out, and it stays in the archive on purpose, as a museum of one specific mistake.
71-14-9-6, counted by hand
July 18, 9:10 am, calculator app, no charts. Bitcoin: $16,830, which is 71 percent. Ethereum: $3,320, which is 14. Stablecoin: $2,130, approximately 9. Cash sitting as USD: $1,420, the last 6. Sum: $23,700 exactly — my money, wearing a stranger's allocation. Four divisions on a phone calculator beat eight months of donut charts, and the calculator has no target tab to confuse anybody, which is now a feature I actively look for in tools.
the eleven-minute fix
9:41 am: sold $6,140 of bitcoin. 9:46 am: bought $1,190 of ethereum with part of the proceeds, leaving the remainder in stablecoin on purpose. Fees across both trades: $14.70. Confirmed at 9:53 am, and the new mix read 45-19-30-6 by my own arithmetic. I refused to force the whole thing home in one mornin' — the remaining drift walks back toward target through two scheduled August buys, cuz a correction should not become a market call just because the calculator is warm. Eleven minutes of clicking for the part that mattered, then a walk around the block.
the uncomfortable part: drift paid
Here is what the rebalancing articles skip. The drift made money. Eight months of letting bitcoin run turned every $150 weekly buy into a bigger slice of the year's best performer, and dragging the weight back down cost me approximately $900 of spring upside, by my own estimate. I did it anyways, the same way I wear a seatbelt on a commute that has never once crashed. The pie is not a profit machine. It is the agreement I signed in November, when the market was calm and my judgment was cheap.
the alarm now says count the pieces
Quarterly, first Saturday, 9 am: an alarm labeled COUNT THE PIECES, next one set for October 3. The rule under it changed too. Count from the balances screen first, four divisions, by hand, and only then open the allocation tab to compare, cuz the comparison means nothin' til the counting is mine. Screenshot both, current first, and name the file with the numbers I computed rather than the numbers the app drew. The tab is for display. The calculator is for truth. Apps get updated at midnight, and my trust now gets renewed quarterly, in writing, by me.
july's screenshot, archived
The July image sits in the folder with its seven ancestors, except this one is a pair: the donut on the wrong tab, and beside it the four divisions from my calculator with the same date underneath. Eight months of a chart describing the plan while the money did somethin' else, closed out by eleven minutes at a kitchen table with a phone calculator and a cup of gone-cold coffee. The alarm is set. The folder is labeled. That is the whole routine.
