BenchSelf

Privacy

What we keep

Short version: your scores, the category of device you used, and a random token. Nothing that points back to you.

What is stored on our side

One row per completed test, containing four things: which test it was, the score, whether you were on a mouse, a trackpad, or a touchscreen, and a random token generated in your browser. Nothing else.

That row exists for one reason. The percentile this site is built around requires a population to compare you to, and the population is made of these rows. Without them the site has nothing to tell you.

What is deliberately not stored

  • No account. There is nothing to sign up for.
  • No IP address is recorded against a result.
  • No user agent string, no device fingerprint, no browser identification.
  • No advertising or analytics tracker of any kind.
  • No cookies. The site sets none.

The random token

A random identifier is generated the first time you complete a test and kept in your browser's local storage. It is not derived from anything about you and it cannot be linked to any other site or service.

It does one job: it lets us count only your first attempt at each test toward the population curve. People who practise get faster, and counting every attempt would drag the distribution toward the practised and away from the truth. Clearing your site data removes the token, at which point a future attempt would be counted as a first one again.

What stays only in your browser

Your run history, your best and average scores, and your device preference are stored in local storage on your machine. They are never sent anywhere. The CSV export is generated from that local data directly, so exporting does not involve a network request.

Clearing your browser's site data for this domain deletes all of it permanently.

Reference distributions

Where a test does not yet have enough real results to rank against, the curve includes a fitted reference distribution based on published summary statistics. Any page that does this says so directly next to the percentile. We never present fitted data as measured data, and the baseline is dropped automatically once enough real results exist.

Contact

Questions about any of this can go to privacy@benchself.com.