Why can a real website still run its own checks when a browser test shows a perfect score? Drawing on 2026 research into server-side fingerprinting, this guide explains the difference between attribute hashes, visitor identifiers, and account verification. It offers practical record-keeping methods for operations teams and shows how to maintain traceable business environments in PurpleMark.
A score of 100 on a test page should be read in light of that tool's scoring criteria. It describes the items covered by that particular check. Which signals a business website uses and how it makes verification decisions require separate investigation.
When reviewing an antidetect browser, divide acceptance into three parts: understand the test's criteria, record environment changes, and confirm actual use in the authorized business workflow. With that context, a score becomes useful information.
What new clues does the 2026 research provide?
In Understanding Server-side Commercial Fingerprinting, Elisa Luo and her coauthors studied how commercial fingerprinting services identify visitors by combining browser attributes, cookies, and IP information. In controlled experiments, the researchers observed visitor identifiers returned by the service, rather than comparing only client-side attribute hashes. Original paper
For experiments that changed one attribute at a time, the paper reports that, among the 26 attributes tested, changing 13 individually changed the client fingerprint hash while leaving the server-side visitor identifier unchanged. This proportion describes only the paper's attribute experiments. It is not a website identification rate, a share of users, or a pass rate for any browser.
The specific point is that a changed attribute summary and a server deciding that two visits came from different visitors are separate outcomes. The study tested particular commercial services in depth. Other websites and services may use different methods.
Check 1: Confirm what the score measures
When you see a test result, find the scoring explanation before looking at individual items. Do not put percentages from different tools on one leaderboard without checking what they mean.
| Test-page result | Question to confirm | Record to keep |
|---|---|---|
| Overall score or percentage | Does a high score mean consistency, coverage, or something else? | Tool name and scoring explanation |
| Uniqueness result | Which sample set was used, and when was it collected? | Sample description and test time |
| Canvas or another hash | Which output does the summary represent? | Attribute, hash, and environment version |
| Visitor identifier | Was it calculated locally or returned by a server? | Definition in the tool documentation |
| Anomaly notice | Which specific field triggered it? | Original field value, exact notice, and retest conditions |
For example, “a graphics output changed” and “the current parameters conflict with one another” are different observations. Keep the original field values so the team can decide whether to check graphics settings, the browser version, or the test tool's own documentation.
Check 2: Keep client-side results separate from business verification
A client-side test can help identify parameters currently exposed to a page. A business website may also use other context when handling a visit. In research revised in September 2026, Castle likewise describes fingerprint information as one signal in a broader assessment and discusses the combined role of network, behavior, and server-side information. Castle research
Keep two linked records. One should capture environment test fields. The other should note the business action's time, page, and outcome. For example, if a test page shows no anomaly but a business dashboard asks for verification again, record both results accurately and follow the platform's verification process.
Do not infer a platform's internal algorithm from this observation alone. Its undisclosed parameter weights, thresholds, and correlation rules generally cannot be reverse-engineered from a single page prompt.
A practical acceptance record can use this format:
Environment name / identifier:
Business account and owner:
Browser version and test date:
Test tool and score definition:
Fields reviewed and observed values:
Business page, action, and result:
Changes since the previous check:
Follow-up action and owner:
There is no need to include passwords or complete session data. Information that identifies the environment, explains the observation, and helps schedule a retest is enough for routine collaboration.
Check 3: Retest to answer one question at a time
If you update the browser engine, change the language, switch proxies, and clear the session all at once, a changed result will not tell you which change was related.
Instead, start with one question, such as “After this update, does the resolution field match the plan?” Keep the previous environment record, make the relevant adjustment, then check that field with the same tool. If you also need to check the business page, record its access conditions and result separately.
In the PurpleMark client, environment names, groups, account assignments, and fingerprint settings help tie these checks to a specific business workflow. Before making an adjustment, confirm the target environment and its owner, then review the relevant parameters. If the network connection is involved, also check the proxy assigned to that environment. This keeps each test record tied to the work it describes.
For example, a team may create separate environments for two authorized brands. Brand A needs its language settings checked after an update, while Brand B continues routine dashboard work. Assign the test to the correct environment and record it under the right brand. The retest is then easier to trace and hand off.
How should a team make decisions from test results?
It is more practical to turn “should we keep using this?” into questions that can be checked than to rely on a score threshold.
| Decision question | Evidence needed |
|---|---|
| Is the environment configured as planned? | Comparison of planned parameters with observed fields |
| Did the current version change anything? | Before-and-after records and relevant retests |
| Is the correct account open? | Mapping between environment, account, and owner |
| How should a verification request be handled? | The platform prompt and its normal verification process |
| Is another adjustment needed? | A specific anomalous field or business observation |
Teams beginning to organize these records can download the PurpleMark client, complete acceptance for one real business environment, and then apply the same record-keeping method to other environments.
Frequently asked questions
Two tools give different scores. Which one should I trust?
Compare what each tool measures and how it defines its score, then check the output for the same field. Different test coverage can produce results that answer different questions.
Does a changed hash mean I have successfully changed my identity?
A changed hash means the tested output or input combination changed. The server-side study above shows that a client hash and a server-side visitor identifier can produce different outcomes, so interpret them separately.
Should I adjust fingerprint settings every day?
Schedule maintenance according to business needs and actual changes. Keep configuration and change records for existing environments. After an update or a specific anomaly, retest the relevant question.


