Defences
Why it matters Not all privacy tools fight fingerprinting the same way, and some make things worse. Knowing the two main strategies helps you choose a browser and settings that genuinely reduce what can be learned about you.
Two ways to hide in a crowd
Picture a costume party where a guest wants not to be recognised. One option is for everyone to wear the same plain mask and the same coat: nobody stands out. Another is for each guest to change a few details every hour, so nothing stays the same long enough to be remembered. Browsers use both ideas.
Uniformity: making many browsers report exactly the same values, so a fingerprint matches huge numbers of people and identifies none.
Randomisation (also called farbling): adding small, harmless random changes to some traits, so the fingerprint is different on each site or each session and cannot be used to link visits.
Uniformity in practice
Tor Browser is the clearest example (checked October 2026). It tries to make every user look alike: the same reported system details, a limited set of fonts, and "letterboxing", which rounds the page area to common sizes so your exact window size does not leak.
Safari takes a lighter version of the same approach (checked October 2026). For example, it shows pages only the fonts that come with the system, not ones you installed, so your font list matches everyone on the same system version.
Where it fails: uniformity only works if the crowd is large and you stay in it. Resize the window to an odd size, add unusual extensions, or change settings, and you step out of the crowd. A small crowd of identical users still identifies "one of these few".
Randomisation in practice
Brave, and Floati's own Shields, add tiny noise to traits such as canvas and audio results (checked October 2026). The noise is too small for you to see, but it changes the fingerprint code. Each site gets different noise, so shopa.example and shopb.example see different fingerprints and cannot match you between them.
Where it fails: some traits cannot be randomised without breaking pages, such as your time zone or the language you read. Noise must also be chosen carefully: if it changes on every single read, a script can read twice, notice the difference and ignore that trait. That is why good farbling keeps the noise steady within one site and session.
Other helpful layers
- Blocking known fingerprinting scripts: tracker blockers stop many of these scripts from loading at all. This works only for scripts on a block list.
- Reducing what browsers reveal: browser makers have been trimming details, such as how much the user agent string (the browser's self-description) gives away.
- Firefox adds fingerprinting protections in its stricter tracking-protection settings (checked October 2026).
A common misconception
"Installing five privacy extensions is five times safer." Extensions can change what pages see, and an unusual combination of extensions can itself be a rare trait. One well-chosen browser with built-in protection usually beats a pile of add-ons.
Try it Visit a reputable fingerprint test, such as the EFF's Cover Your Tracks, in two different browsers and compare the results. These tests run on their own servers and keep their own statistics, so read their privacy notes first.
Recap
- Uniformity makes everyone look the same; it fails when the crowd is small or you stand out.
- Randomisation (farbling) makes your fingerprint differ per site; it fails for traits that cannot be noised.
- Blocking known fingerprinting scripts helps, but only for scripts on a list.
- Piling on extensions can make you more unusual, not less.
This lesson is locked
Lessons open one at a time. Finish the previous lesson to unlock this one.