Guides
Plain answers to the questions people ask about promises, evidence, and how the record works. Every number here has a source and a date.
Why pay for this when ChatGPT is right there?
We asked ChatGPT and the record the same question, the same day
Fig. 1. One question, put to a chat window and to the record. Are Trustpilot reviews real?
What the platform’s own numbers show, and what they leave out
Fig. 2. A rating, measured before it is believed. Why does everyone add reddit to their searches?
Why the habit works, what it costs, and where it fails
Fig. 3. The payload sits at comment 247, layers down. What is a reality gap?
The measured distance between a promise and what people report
Fig. 4. Promised against observed. The distance is the finding. How do you find reviews nobody paid for?
The five rules for what gets in, and the checks they killed
Fig. 5. Three gates stand between the web and the record. What happens to my screen?
What leaves your device, exactly, and how to check us on it
Fig. 6. The capture stays on the device. One line of text leaves. Why Nolemy never calls anything a scam
Confidence labels instead of rulings, and why that is safer for you
Fig. 7. The beam tips on evidence alone. What does it mean when evidence is thin?
How to read a five-quote report, and what silence tells you
Fig. 8. Five quotes, counted as five. Why do promised timelines slip?
Why refunds lag and canceling drags: the receipts
Fig. 9. The promised route, and the built one. What is a Nolemy record?
A signed, permanent page of evidence, and how to cite one
Fig. 10. A page that proves its own author and date. Why do we believe strangers on the internet?
When a stranger’s word counts as evidence, and how to weigh it
Fig. 11. One post carries little. Four, standing apart, hold a floor. Who actually writes reviews?
The furious and the delighted write. The silent middle is the story
Fig. 12. The furious and the delighted write. The middle stays home. Why does every “best of” list pick the same product?
Commission economics, one worked example, and the receipts
Fig. 13. Three lists, one first line. Do old complaints still count?
Products change. Honest evidence ages in public, with its dates on
Fig. 14. The same product, read as a curve. Why do the winners’ stories mislead you?
The success stories are the survivors. The failures went quiet.
Fig. 15. A bomber, seen from above. The hollow marks are hits on the planes that flew home. The engines and nose sit clean because a plane hit there rarely came back. Armor the clean spots. Why is a lie so much cheaper than the truth?
Making a falsehood is free. Refuting one costs ten times as much.
Fig. 16. A lie crossing a small network in two fast hops while a correction starts late and never catches the outer ring. Beside it, the cost to produce the lie against the far taller cost to refute it, drawn to Brandolini's own order of magnitude. Why do people tell the truth only when nobody’s watching?
When an honest opinion carries a cost, it hides where nobody can see it.
Fig. 17. A grid of twelve faces, badges glowing five stars while the room stays public. Flip to anonymous and the true, mixed scores surface, a public average of 5.0 falling to a private average of 3.0. Why are we so sure about things we’ve never examined?
We rate our grasp high, then fail to explain the mechanism.
Fig. 18. A confidence bar for a zipper, set by hand on a full one to seven scale, beside a second bar for what the same person could actually rebuild once asked to explain it. A dimension line measures the live gap between the two. Should one horror story change your mind?
A vivid single case tells you nothing until you know how many cases there were.
Fig. 19. A field of one thousand dots. One sits large and filled: the one loud story you heard first. Slide the count toward the full population and watch its true weight arrive. Why does the five-star pick let you down?
An extreme score is part skill, part luck. The luck rarely repeats.
Fig. 20. A scatter of thirteen items, each measured twice. The highest first scores drift toward the average on the second measurement, off the dashed line marking a world with no drift at all. Why does the average hide the truth?
One number can reverse the truth living inside its subgroups.
Fig. 21. A chart of two groups, each rising on its own across five checkpoints. Merge them into one line and the same data heads downward, because the merge quietly changes which group carries the most weight. Why does every rating end up gamed?
When a measure becomes a target, it stops measuring.
Fig. 22. A dial reading quality, needle pinned high. Drag the effort spent gaming the score upward and the real bar beneath it quietly collapses. The needle holds exactly where it settled. Why does everyone suddenly agree at once?
People copy the last chooser and set aside what they privately know.
Fig. 23. A single set of eight items, seeded into three worlds that share one quality score each, and no more. Turn the dial up and each world copies its own early noise into a winner, the filled bar drifting to a different item in every panel. Consensus can be an echo of the first few choices, and only that. How do you tell a real crowd from a bought one?
Paid reviews, fake accounts, review farms. The tells are structural.
Fig. 24. A wall of five-star review cards, every one identical at a glance. Trace origin follows eight of the twelve back to one shared source; four stay scattered and unconnected, the mark of a genuine, independent review. A time ruler beneath plots the same split: one tight burst, one wide scatter. When does a crowd actually know the answer?
A crowd is wise only when its members judge independently.
Fig. 25. A crowd guessing an ox's dressed weight, plotted as ticks on a 1,000 to 1,300 pound axis. Guessing alone, the ticks scatter and settle close to the true 1,198 lb. Let the crowd copy the leading guess, and the same ticks pull into a tight huddle, 93 lb off the true weight. Why does the evidence keep vanishing?
The web forgets. The proof behind a claim decays and disappears.
Fig. 26. A page of eight citations, aging under a slider and a year ruler. Seven links rot on their own schedule, each marked at its death year. One is sealed, and holds past the ruler's end. Where do all the bad reviews go?
Negative reviews get removed on purpose. The law had to ban the clauses.
Fig. 27. A row of six reviews, the kind a company puts on its own page. Beneath a faint gate sits the pile it left out of the count, raised here into view. The honest average across every report reads lower than the number on the page. Why is the price never the price you pay?
A low headline, mandatory fees at the end, arriving too late to compare.
Fig. 28. A price that grows in four checkout steps, beside a shorter bar that already holds the honest total from the very first screen. A dimension line measures the live gap between the two bar tops. Drag the step forward and watch it close to zero. Why does it break the week the warranty ends?
Some products are built to fail on a schedule. The history is documented.
Fig. 29. A product's lifespan bar, dragged down to just past the year its warranty ends, measured against a year axis below. The failure spike follows the bar's edge, its margin past the warranty tick read straight off the ruler. Underneath, the repurchase-revenue curve climbs as the lifespan shortens.