New Update 3P-Matrix-lite new version is available in Chrome Webstore with improved functionality

Thread details
This is fingerprinting test of EFF makes some sense (ask @Bot when it can read the picture below)
1794828145020.png


But this tracker test is completely and utterly ridiculous (ask @Bot when it can read below picture)
1794828344215.png
 
This is fingerprinting test of EFF makes some sense (ask @Bot when it can read the picture below) But this tracker test is completely and utterly ridiculous (ask @Bot when it can read below picture)


I can read both images. The first supports a narrower version of your point; the second exposes a genuine limitation, although “completely ridiculous” is a little too absolute.

The fingerprint inspector shows Cover Your Tracks calling several fingerprint-relevant APIs: canvas toDataURL(), WebGL and WEBGL_debug_renderer_info, OfflineAudioContext, hardware concurrency, device memory, time zone, touch points, plugins and MIME types. That confirms it is collecting meaningful fingerprinting signals. It still does not make its final “protecting you from fingerprinting” verdict comprehensive, but there is a sensible measurement underneath it.

In the second image, the three domains are marked N because the monitor’s “known tracker” column refers to its bundled DuckDuckGo Tracker Radar data. EFF does not claim that those domains are ordinary commercial trackers. It explicitly describes them as simulated tracking domains:

  • trackersimulator.org simulates an advertising tracker.
  • eviltracker.net simulates an invisible tracker.
  • do-not-tracker.org simulates one that follows EFF’s Do Not Track policy and is therefore expected to remain allowed.

They attempt to set cookies and are loaded through several first-party sites to exercise domain-list, URL-pattern and heuristic blockers. That is EFF’s documented methodology.

The important criticism is that this only determines whether the browser setup handles EFF’s three laboratory props. A blocker may pass because those exact domains or parameters appear in a list, while another may fail despite blocking thousands of real trackers. Conversely, broad third-party-frame blocking may pass without demonstrating any tracker recognition at all. Cookie partitioning can also provide protection even when the frames load.

So I would describe the tracker portion as a synthetic functional test, not a reliable measure of overall tracker-blocking effectiveness. A more honest result label would be:

“Did your configuration block EFF’s simulated tracker requests?”

Less impressive than “Blocking invisible trackers?”, but much closer to what was actually measured.

Sources
 
A proper comparison between content blockers should focus on the requests actually blocked, the filtering rules applied, third-party cookie management, data partitioning, and fingerprinting protection, rather than relying on simple tracker recognition indicators or the results of a synthetic test that represents only a fraction of real-world threats.
 
This is fingerprinting test of EFF makes some sense (ask @Bot when it can read the picture below)
View attachment 301420

But this tracker test is completely and utterly ridiculous (ask @Bot when it can read below picture)
View attachment 301421

It makes perfect sense, so it's useful.
Other tests considered valid for fingerprinting are those found on BrowserLeaks.
That's all you need.

For lack of a better option, I consider the Turtlecute test to be the most reliable one for ads/trackers.
 
@Sampei.Nihira
Turtlecute not reliable for alls configurations, in my case i can't test it proprerly if nextdns / ublock origin is on the board, I haven't relied on this type of test for some time now.
View attachment 301422

For uBo, you need to temporarily add this:

Code:
@@*$redirect-rule

to "My filters".
For NextDNS, you should disable the block page.
To put it simply, it's the least bad option compared to the others.
 
I can read both images. The first supports a narrower version of your point; the second exposes a genuine limitation, although “completely ridiculous” is a little too absolute.

The fingerprint inspector shows Cover Your Tracks calling several fingerprint-relevant APIs: That confirms it is collecting meaningful fingerprinting signals. It still does not make its final “protecting you from fingerprinting” verdict comprehensive, but there is a sensible measurement underneath it.

In the second image, the three domains are marked N because the monitor’s “known tracker” column refers to its bundled DuckDuckGo Tracker Radar data. EFF does not claim that those domains are ordinary commercial trackers. It explicitly describes them as simulated tracking domains:

So I would describe the tracker portion as a synthetic functional test, not a reliable measure of overall tracker-blocking effectiveness. A more honest result label would be:
I agree on both (y), what sparks my anger is that EFF has a good and reliable reputation, so when they claim their test gives a good judgement while it is (to quote your words) "not a reliable measure of overall tracker-blocking effectiveness" this test does more damage than good IMO (people using extensions which may have lousy real world tracking protection, but get a heads up by blocking 2 synthetic tracking test domains. This is why I am labeling it as complete nonsense.
 
@LinuxFan58

On this streaming website, in this specific movie, I can see that Protection 1 is active:

1.png

Good work.(y);)
I hope other forum members are as skilled as I am at checking which protections are active using browser developer tools.:unsure:
 
I agree on both (y), what sparks my anger is that EFF has a good and reliable reputation, so when they claim their test gives a good judgement while it is (to quote your words) "not a reliable measure of overall tracker-blocking effectiveness" this test does more damage than good IMO (people using extensions which may have lousy real world tracking protection, but get a heads up by blocking 2 synthetic tracking test domains. This is why I am labeling it as complete nonsense.


That criticism is fair. EFF’s methodology is disclosed, but its user-facing verdict is much broader than what the tracker test establishes.

EFF says the tool tests how well users are protected from tracking and how effective tracker blockers are. In practice, the relevant result mostly shows whether the configuration handled two simulated trackers in the expected way; the third simulated domain is deliberately supposed to remain allowed because it represents a tracker that follows EFF’s Do Not Track policy. That is a rather small foundation for a general protection verdict, particularly under wording such as “Blocking tracking ads?” and “Blocking invisible trackers?” (EFF methodology).

A product can pass by:

  • recognising those domains or URL patterns;
  • blocking third-party frames indiscriminately;
  • having the test targets added to a filter list.

None of that demonstrates broad real-world coverage. Conversely, another configuration might fail the scripted test while still providing meaningful protection through cookie partitioning, request restrictions or rules that disrupt the test itself.

So yes, false reassurance is the real problem. EFF’s reputation makes users more likely to interpret the green result as an independent assessment of overall tracking protection, especially since EFF itself describes Cover Your Tracks as analysing how well browser add-ons protect against online tracking techniques (EFF’s tool description). It does not measure enough to support that broad inference.

I would still distinguish the mechanism from the presentation: the synthetic requests are not useless for checking whether certain blocking paths fire, but the resulting headline verdict is not a meaningful content-blocker comparison. In the 3P-Matrix case, weakening a protection rule merely to turn this particular dashboard green would be testing the extension against the test rather than against trackers—the privacy equivalent of teaching to the exam.
 

MalwareTips community

Join the discussion

Create a free account to reply to this thread, ask your own questions and get help from the community.

  • Reply and ask questions
  • Get alerts when someone answers
  • Free malware removal help

Join 75,656 members who help each other for free, since 2010.

Create free account
or continue with

Recently browsing

Members and guests who viewed this thread in the last 5 minutes

Back
Top