What did Radar see on Wish?

Chrome (Chromium) scan of 2026-11-13 17:37:16 UTC.

Headline for the Before choice window

We observed 17 signal families across 32 scripts, with 15 data-flow-linked transmissions, during the observation window before any consent choice, measured in Chrome (Chromium) 153 on linux from tr-workstation.

  • Measured in Chrome (Chromium) 153 on linux from tr-workstation.
  • Across 3 valid runs, 17 of 18 observed signal families appeared in every run and 1 appeared in some runs only.

Read this first

  • This report describes one scan: Chrome (Chromium) 153.0.8010.12, region tr-workstation, method version NR-v0.3.0.
  • Scans of another browser, region or method version are separate and never merged into it.
  • Counts describe what Radar observed in this scan; they are not a rating.
  • 17 limitations are recorded for this scan. Read the limitations

Choose a window

Counts in this scan

Radar loaded the page four ways, each time in a fresh browser profile. Pick a window; every number below follows it.

Showing Before choice: 3 of 3 runs valid.

  • Signal families17all before choicerange 17–18 over 3 valid runsView evidence
  • API surfaces99all before choicerange 96–105 over 3 valid runsView evidence
  • Scripts32all before choicerange 18–59 over 3 valid runsView evidence
  • Destinations37all before choice20 third-partyrange 22–67 over 3 valid runsView evidence
  • Data-flow-linked transmissions15all before choice111 correlated onlyrange 1–42 over 3 valid runsView evidence

Means over valid runs, rounded to whole numbers. Counts describe what Radar observed in this browser, region and method version; they are not a rating.

The four windows side by side

One group per count: the same measure in each window. A window without a valid run shows why and is never drawn as zero.

Means over valid runs, rounded for the chart. Reject all and Accept all count events after the choice only.

How these four counts were measured

Each window starts from a fresh browser profile. You are looking at Before choice.

Stage 1 of 5: Fresh profile

Every situation starts in a new browser profile of its own and loads the same entry page, so no cookie, storage or consent state carries over from one situation to the next.

One fresh browser profile branches into four situations, each run in a new profile of its own. Before choice: no consent action, the whole observation window is counted. Reject all: the most comprehensive rejection control is used, and only events after the action settled are counted. Accept all: the most comprehensive acceptance control is used, and only events after the action settled are counted. GPC enabled: the Global Privacy Control signal is sent with no consent action, and the whole window is counted. Events before a consent action are reported separately and never added.

For Before choice

Notes about this scan (2)

What to know before you read the numbers: the window, other browsers and any run that did not count.

  • Also measured in Firefox
    Radar has a separate Firefox scan of this platform. Numbers are never merged across browsers. Open the Firefox report
  • Before choice: whole observation window
    Every event from navigation until the 10 s observation window ends. No consent action is taken in these runs, so all events are before choice.

Runs in this scenario (3)

Each repeat is a separate visit in a fresh profile. Only valid runs enter the counts.

  • 1st of 3 repeatsValidThe run completed and enters comparisons.
  • 2nd of 3 repeatsValidThe run completed and enters comparisons.
  • 3rd of 3 repeatsValidThe run completed and enters comparisons.

From the evidence

Findings and questions

Sentences made of observed numbers, and questions answered only from what this scan observed.

What the numbers say

Questions, answered from the evidence

Did the scan observe canvas image reads?
Yes, in Before choice: 2 technique cards list the script, the stage and the evidence. Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe font probing?
No. It could be measured and was not observed in the 12 valid runs. Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Not observed
Did the scan observe audio or WebGL reads?
Yes, in Before choice: 2 technique cards list the script, the stage and the evidence. Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe WebRTC connection set-up?
Yes, in Before choice, Reject all, Accept all, GPC enabled: 4 technique cards list the script, the stage and the evidence. Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe identifier-like values written to storage?
No. It could be measured and was not observed in the 12 valid runs. Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Not observed
Did the scan observe third-party cookies?
Yes, in Before choice, Reject all, Accept all, GPC enabled: 4 technique cards list the script, the stage and the evidence. Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe candidate identifier values sent to another party?
Yes, in Before choice: 7 technique cards list the script, the stage and the evidence. Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe typed made-up values sent to another party?
No. It could be measured and was not observed in the 12 valid runs. Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Not observed
Did the scan observe hashed made-up contact details sent to another party?
Yes, in Accept all: 3 technique cards list the script, the stage and the evidence. Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to Google?
Yes: Before choice 8 requests, Reject all 10 requests, Accept all 40 requests, GPC enabled 8 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to TrustArc?
Yes: Before choice 9 requests, Reject all 6 requests, Accept all 6 requests, GPC enabled 9 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to Microsoft?
Yes: Before choice 9 requests, Reject all 6 requests, Accept all 12 requests, GPC enabled 1 request (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to ByteDance?
Yes: Accept all 16 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to The Trade Desk?
Yes: Accept all 14 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to Pinterest?
Yes: Accept all 6 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to Cloudflare?
Yes: Before choice 1 request, Reject all 1 request, Accept all 1 request, GPC enabled 1 request (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to Meta?
Yes: Accept all 4 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to Taboola?
Yes: Accept all 4 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to Magnite?
Yes: Accept all 2 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to X Corp?
Yes: Accept all 2 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed
Did the scan observe requests to Xandr (Microsoft)?
Yes: Accept all 2 requests (median over the valid runs). Not a guarantee: observed from one location on one date. Not observed is not proof of absence.
Observed

Not a guarantee: observed from one location on one date. Not observed is not proof of absence.

Descriptive score

Score of this scan

A versioned arithmetic over what was observed: it starts at 100 and every deduction is listed with its evidence.

Score
35 of 100Method NR-SCORE-1
Coverage
100%9 of 9 items measured; the floor is 70%
Percentile
2in a corpus of 288 scored platforms
Rank
282 of 288in the same corpus
Not a legal assessment. Not a guarantee. Observed from one location on one date.

Platform operators can reply to a score, ask for a correction or opt out by writing to hello@noxtica.com.

Write to hello@noxtica.com

Every deduction of the score: 100 minus 65 points
ItemObservedRulePoints taken offEvidence
Third-party requests before any choice45 third-party requestsBefore choice, median of 3 valid runs, rounded up1 point per 10 units, at most 85
External registrable domains contacted before any choice10 external registrable domainsBefore choice, median of 3 valid runs, rounded up2 points per unit, at most 1616capped at 16
Third-party cookies set before any choice9 third-party cookie keysBefore choice, median of 3 valid runs, rounded up2 points per unit, at most 88capped at 8
Fingerprint-like technique observations before any choice4 techniques (canvas, fonts, audio or WebGL, WebRTC)Before choice, median of 3 valid runs, rounded up4 points per unit, at most 1212capped at 12
Candidate identifiers shared with other registrable domains2 receiving registrable domainsBefore choice, median of 3 valid runs, rounded up2 points per unit, at most 84
Typed made-up values sent to other parties (unhashed forms)0 receiving partiesBefore choice, median of 3 valid runs, rounded up7 points per unit, at most 140No evidence id
Hashed contact details sent to other parties0 receiving partiesBefore choice, median of 3 valid runs, rounded up7 points per unit, at most 140No evidence id
External registrable domains contacted after Reject all9 external registrable domainsReject all, median of 3 valid runs, rounded up2 points per unit, at most 1010capped at 10
External registrable domains contacted with Global Privacy Control enabled9 external registrable domainsGPC enabled, median of 3 valid runs, rounded up2 points per unit, at most 1010capped at 10

Points taken off in total: 65. Score: 100 minus 65 = 35. An item that could not be measured takes nothing off and lowers the coverage.

  • Score method NR-SCORE-1: the score starts at 100 and each measured item takes the points of its table entry, listed here with the observed count and the evidence it cites.
  • Coverage is 100%: 9 of 9 scoring items were measured. An item that could not be measured deducts nothing.
  • 5 deductions reached the cap of its table entry.
  • Counts: median over the valid runs of the scenario, rounded up; a count taken from rows (third-party cookie keys, shared identifiers, typed values) counts a row when it was observed in at least half of the valid runs of the scenario.

The score method, with the table of deductions

What next

The report is four steps. This is the first; the others hold the detail behind these counts.