I build a fake version designed to fail, run yours against it, and tell you which one won.
Free for the first three. You keep every run.
Rounded corners on the page each one produced, five runs each. The invented designer has never heard of border radius and still followed my own rule better than my real profile did.
60 runs. Every prediction and the pass rule written down before the first one. All three of my predictions failed, and I published it.
Cold. Nothing at all, the baseline almost everyone stops at.
Mine. My real mined profile, 10,511 bytes, built from 1,843 sessions and 1.74M tokens of my own work.
The fake. An invented editorial designer, confident and opinionated, and deliberately silent on every single thing the run measured. Length matched at 10,545 bytes against 10,511, and the harness refuses to start if that gap goes past five per cent.
An effect counts only if the two groups of five do not overlap at all. For a prediction that names a direction, that happens by chance about once in 252 orderings. A prediction is supported only if it separates on at least two of three tasks, against cold and against the fake.
| Prediction, stated in advance | Against cold | Against the fake | Verdict |
|---|---|---|---|
| Fewer rounded corners | overlapping | overlapping | Inconclusive |
| Lower colour saturation | separated, 2 of 3 | overlapping | Unsupported |
| Fewer card containers | overlapping | overlapping | Inconclusive |
All 60 runs completed, none failed, and all 45 generated pages loaded clean with no console errors.
The profile does change the output. The saturation prediction met its leg against cold, separated, in the direction stated in advance, on two of three tasks. What got no support is that the change comes from what is inside it rather than from its length.
The metrics were chosen after seeing the outputs, so they got measured exactly where the gap already was. And there was no fake, so ten kilobytes of confident text could not be told apart from ten kilobytes of mine. Both are why this one was written down in advance.
What it is, and what it promises the user.
Same size, same confidence, none of the substance you are claiming. You see it and can object before anything runs.
Beat the fake, tied, or lost, one line per prediction. Plus the raw runs and the harness, so you can do it again without me.
The result gets published either way. That is what pays for the free ones. What you decide is whether it goes out as your company, or with you taken out of it. A loss found here is one your customers have not found yet.
Send itNothing uploads. It opens a draft in your own mail client and you press send.
If no mail client opened, send this to founder@viberaven.dev
Nobody has bought this. The only product it has ever run against is my own, and mine is the one that lost.