2,376
daily scores
For twenty years he wrote the documents that prove airplane wings stay attached to airplanes. There was exactly one variable the discipline never measured. The engineer. This is what the same math says when the instrument is turned inward.
From why to i · the wave continues
Adapted from Insider Threat · Volume II, Chapter 12 — The Spar That Failed
Airplane wings don't usually break because of one terrible moment. They wear out from thousands of ordinary flexes — every bump of turbulence bends the metal a little, and engineers have an eighty-year-old math for adding those little bends up into one number: how much of the structure's life has been spent.
This page points that exact math at a person. One number a day — how are you, 0 to 10 — kept for twelve and a half years. Every swing between a good stretch and a hard one is one flex of the structure. The math adds the flexes up the same way it does for a wing.
Two things fall out. A running total of wear, called D. And the surprise of the whole record: most of the wear didn't come from the worst days — it came from the everyday swings. Scroll for the story, or jump straight to trying it on your own scores. Nothing you enter leaves your browser.
In November 2013 the daily ritual began: a single emotional-wellness score from 0 to 10. Zero was the floor. Ten was bullet-proof. The scoring took less time than finishing brushing his teeth.
By May 2026 the recorder held 2,376 daily entries across forty-eight behavioral columns — sleep, exercise, a rumination index, a self-worth score, blood pressure, a gratitude tally, the journal. It is the same kind of data an aircraft carries on its flight-data recorder. The aircraft samples at hundreds of hertz. He sampled once a day.
But the principle is identical. A structure that carries variable load deposits damage. Damage that is not measured accumulates anyway. Damage that is measured can be turned into a number on a panel.
The procedure is ASTM E1049-85 — the same cycle-counting used on military wing spars under MIL-STD-1530D. It takes the messy record of every load a structure has ever seen and returns one number: how much fatigue life has been spent. Six steps turn a wellness score into a damage index.
If you have never seen a fatigue calculation, here is the whole idea in four pictures — no engineering required.
The structure doesn't care how you feel — it cares how hard it's being pushed. So we flip the 0–10 score into a load: a floor day is maximum stress, a bulletproof day almost none. σ = 10 − W.
Sitting still does nothing. Damage comes from the distance between a high and the low that follows — a peak of joy dropping to a valley of grief, like a fighter pulling up then pushing over. We count every swing.
A thousand gentle ups-and-downs are nearly free; a handful of violent ones spend most of the life. The curve relating swing size to how many a structure survives is the S-N curve — its tolerance.
Each swing spends a thin slice of the total. Stack the slices and you get one number, D. Empty is 0; a full tank — the calibrated end of life — is 1.0. The precise version of all this is right below.
The constant A is a calibration handle. Here it is set so cumulative damage equals exactly 1.0 at a single objectively physiological event — the documented stroke of October 23, 2024. Everything below 1 is pre-failure; 1 is the event; above 1 is Extended Service Life.
D = 1.339 — the running total of wear. The scale is set so 1.0 equals one real, documented medical event. Below 1 is the run-up; past 1 is borrowed time that turned out, after inspection and repair, to be livable — what aircraft people call Extended Service Life.
The two rates are the speedometers: how much life was spent in the last 30 days (the sprint) and the last year (the marathon pace). The yearly rate is the one worth watching — it started climbing three years before the event did.
The curve crossed 0.3 in late 2021 — the start of an eight-resignation stretch. It crossed 0.7 in mid-2023, the worst year of the record. It reached 1.0 at 03:13 on October 23, 2024. The wing did not come off. The airframe was grounded, inspected, repaired, and returned to service. The curve continues past 1.0 — at a slope roughly a third of its chronic peak.
Curve computed directly from all 2,376 daily scores in the Hindsight 80/20 workbook — the same rainflow → Walker → S-N → Miner pipeline the live tool runs below. It reproduces the reported figures: D = 1.34 today, exactly 1.0 at the stroke, 583 cycles.
38% of twelve and a half years of fatigue damage lives in a single bin — high-amplitude swings around a tired, mid-range baseline.
The next bin adds 12%. The top three hold more than half of everything spent on the structure. The rare floor day is dramatic and earns its own line in the journal — but it deposits less damage than the engineer would have guessed.
The damage is in the swing, not the catastrophe.
In aircraft analysis the exceedance curve answers a single question: how often does the structure see this kind of load? A fighter has roughly nine g of headroom above cruise and three g of cellar below. The human is upside-down from that.
Baseline median wellness sits at 7 — the structural cruise altitude. The ceiling is 10: three points of headroom, where the rare elated peak lives. The floor is 0: seven points of cellar. The pull-ups are rarer than the push-overs, and the push-overs go deeper. The airframe issued has three g of joy and seven g of grief, and it does not appear to be negotiable.
| n_e | Wellness | Recurrence |
|---|---|---|
| +3.0 | W = 10 | ~1 / 21 days |
| +2.0 | W ≥ 9 | ~1 / 9 days |
| +0.5 | W ≥ 7.5 | ~1 / 6 days |
| −2.0 | W ≤ 5 | ~1 / 8 days |
| −4.0 | W ≤ 3 | ~1 / 13 days |
| −7.0 | W = 0 | ~1 / 33 days |
A parallel regression — 107 behaviors, 17 studies, 2,261 days, screened at p ≤ 0.05 — ranks what is actually associated with wellness movement in this subject's own data. Cardio, formal meditation, sleep duration, and most supplements did not survive the screen. The signal is in the inner game.
Four behaviors are doubly validated — both regression-significant and visibly lived on the high-wellness days, not merely reached for as coping on the hard ones:
PLAY · ART JOURNAL · LET GO · I-MATTER
These act on the amplitude of the swing — the lever the damage analysis says matters most. Vocation, geography, and who is in the room move the mean; those are bigger structural changes. The amplitude moves are the ones you can make this week, and they show up first on the panel.
Coded from a year of journal entries against Plutchik's wheel. The positive pole is wide but mild — joy is common, ecstasy almost absent. The negative pole is narrow but reaches its intense form readily: basic disgust scores zero, but loathing, its intense variant, posts thirty-three.
An oscillation between joy and loathing produces exactly the high-amplitude cycles the damage Pareto flags as most fatiguing. Two completely independent analyses — rainflow on the stress signal, a count on the journal — describe the same asymmetric spectrum.
Two instruments. One panel. They agree.
Paste a column of daily 0–10 wellness scores, load a sample, or enter a few by hand. The full pipeline — rainflow → Walker → S-N → Miner → exceedance — runs entirely in your browser. Nothing is uploaded or stored.
A run of daily 0–10 scores — even a month shows the shape. A year unlocks the chronic gauge, the one that mattered most in this record.
Your own D(t) curve, both rate gauges, and the map of where your damage lives — the same readout as the twelve-year record above.
Nothing. The math runs on this page, in your browser. Close the tab and it is gone — no upload, no account, no storage.
Want the record to persist? Hindsight Daily is this same pipeline as a phone app — one number a day, saved on your device, with a monthly service-log report. Fourteen days free.
A synthetic ~3-year record with the same character as the reference data — a positive baseline, occasional floor days, and joy peaks. Generated fresh each time.
These are the four knobs of the S-N / Walker model, exactly as in the workbook. α sets how much a tired baseline matters versus the size of the swing; m sets how steeply big swings outweigh small ones; S_th is the swing size below which a day does no lasting damage; A just sets the scale. A is calibrated to the reference 12.5-year record — a gentler spectrum than his accumulates damage more slowly, so a small D is good news.
No. It is an engineering read of a signal — a gauge of accumulated load, not a diagnosis, a screening score, or a prediction. It has been calibrated on exactly one person. Testing whether it means anything in a clinical sample is precisely what the proposed NIMH Phase I study exists to do.
By itself — nothing certain. The 1.0 mark is a calibration pin, set here to one documented event on one record. Without your own anchor event, the number worth watching is the yearly rate: whether the pace of wear is rising, steady, or easing. The rate doesn't care where 1.0 sits.
They matter — they are just rarer, and they total less than intuition predicts. What this record shows is that the everyday swing between good stretches and hard ones deposited most of the wear. The hopeful part: swing amplitude is something a person can actually work on this week.
The pipeline runs on anything with a few reversals — a month shows the shape. The 30-day gauge needs a month; the chronic gauge needs a year. Gaps are fine: anything over a week simply splits the record, and no swing is counted across a gap.
It carries less than a flight-data recorder and more than nothing — and it takes less time than finishing brushing your teeth, which is why the record survived twelve years. The method doesn't need precision from any single day; it reads the pattern of thousands.
The Excel companion workbook runs this same pipeline with every formula exposed — drop in your own record and trace any single day all the way to D. The full story is in Insider Threat, Volume II, Chapter 12. Both are linked below.
The structure that was being inspected was inspecting itself — with its own instruments, from inside the cockpit, while still in flight.
One number a day, on your phone — the same pipeline as this page, running entirely on your device. v2.1 adds the Mission layer: pick a mission spectrum, name it in your own words, and a daily −1/0/+1 tap tracks whether the days build it. Fourteen days free, then one purchase. Nothing is ever uploaded.
Start your record → The bookThe full story lives in Volume II, Chapter 12 — The Spar That Failed. Where the engineering vocabulary finally says the part the older ones could not.
Get the book → The labEvery person is a material under load. Find yours, then watch how it carries the spectrum it is given.
Open Mind My Material →The framework — why measure a life in damage units →
Everything on this page is Task V — wear on a structure whose properties are assumed. The companion instrument measures the properties themselves — i5.3.2.1, the modal survey: how fast do you get back up? →