TWO DIFFERENT EVIDENCE QUESTIONS
The original editorial verdict alongside the stricter plan, platform and test-context checks for comparable test evidence.
Provisional editorial metascores are published under editorial-1. The “0 eligible inputs” here refers to stricter exact-tier/platform/test-window checks, not a zero VPN rating. An aggregation of service-level editorial opinions does not establish comparable tests.
35 records from 5 publications and 4 provisional families: 30 overall ratings and 5 category-only records. Passing all stricter matched-test context checks: 0.
TechRadar and Tom’s Guide share the Future family. RTINGS’ Private Browsing category is not an overall editorial verdict. PrivadoVPN’s PCWorld article mixes free and premium coverage.
20 observations in 3 separate cohorts. Eligible performance inputs: 0.
Missing per-test context and provisional status keep these out of scoring. Their download metrics, routes and baselines are not interchangeable. Read each cohort.
Policies, court filings, assessor reports and ownership events retain their own source and scope. They are not votes or a numeric privacy score.
An audit count, absent case file or country group cannot establish safety. Inspect the evidence matrix.
This stricter record checklist remains version readiness-1. It requires exact test context. The separately published editorial-1 policy instead aggregates overall service opinions with equal family shares, at least two families and visible exclusions. Neither policy turns article dates into test dates. A matched-test performance/security synthesis still needs its own declared method and sufficient inputs.
Keep a finite rating, its bounds and category. A missing rating is not zero.
Overall editorial verdicts, category scores and customer ratings are separate lanes.
A mixed free/premium verdict is not a separately scored paid result.
Match the tested subscription to a named tier; a provider’s current cheapest offer cannot fill this gap.
Record the tested device/OS. An available app or a publisher’s general method is not a per-provider test record.
Record test start/end dates separately from article publication, editing and collection dates.
Review publisher families, reused data and syndication before treating inputs as independent.
Identify the actual connection protocol for a performance observation.
Record the number of measurements behind a performance statistic.
Identify the tested client version rather than assuming today’s app.
Record the tested endpoints/routes. Normalizing a baseline does not make different routes equivalent.
Keep provisional observations outside score calculations until their limitations are resolved.
3 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
 Top10VPNtop10vpn family | 8.3 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2026-10-14 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports uneven long-distance speeds, a dated interface and missing advanced options such as multi-hop. |
 TechRadarfuture family | 3.9 / 5Overall editorial rating | See publisher for tested planTested tier / platform: not captured / not captured | Published: 2022-10-24 Article update: 2026-02-26 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It finds the feature set less innovative, the device limit awkward and port forwarding unavailable. |
 PCWorldpcworld family | 4 / 5Overall editorial rating | Paid serviceTested tier / platform: not captured / not captured | Published: 2026-01-22 Article update: 2026-01-22 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports variable server performance and fewer extra security features than some competitors. |
4 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
| TechRadarfuture family | 4.1 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-10-24 Article update: 2026-04-22 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It finds the service expensive for its feature set; standard Lightway performance is weaker, and port forwarding is absent. |
 Tom’s Guidefuture family | 4.45 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-10-24 Article update: 2026-04-22 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | Its fastest results depend on Windows-only Lightway Turbo. The review also flags missing iOS controls and no multi-hop feature. |
| Top10VPNtop10vpn family | 9.4 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2026-09-22 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It considers the service relatively expensive and less feature-rich for advanced users, with neither multi-hop nor port forwarding. |
 RTINGSrtings family | Private Browsing: 6.2 / 10 Category ratings only; no overall score imported. | Paid service; see source for tested tierTested tier / platform: not captured / not captured | Published: 2025-03-31 Article update: 2026-07-13 Test window: Not captured | Separate review questions · Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | Its Private Browsing assessment is less favorable because of policy and business-practice concerns. That category is separate from technical security and overall buying value. |
3 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
| Top10VPNtop10vpn family | 8.3 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2026-07-10 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It warns that the kill switch is not enabled by default and finds international streaming and app design less convincing. Its separate free-tier verdict is much weaker on speed. |
| TechRadarfuture family | 4 / 5Overall editorial rating | See publisher for tested planTested tier / platform: not captured / not captured | Published: 2022-10-24 Article update: 2025-12-08 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | Its captured negatives flag small kill-switch issues, dated no-logs audit coverage and unclear virtual-location disclosure. Beginners may find the desktop apps complicated. |
| PCWorldpcworld family | 4 / 5Overall editorial rating | Paid serviceTested tier / platform: not captured / not captured | Published: 2026-03-17 Article update: 2026-03-17 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports inconsistent speeds and a learning curve for beginners. Its unlimited-data free version is assessed separately and was slow in the review. |
4 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
| TechRadarfuture family | 4.5 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-10-24 Article update: 2026-11-04 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | Features differ between operating systems, and keyboard-only navigation is limited. These are platform-specific judgments, not a promise for every device. |
| Tom’s Guidefuture family | 4.45 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-10-28 Article update: 2026-10-20 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | Renewal increases and paid upgrades weaken the value proposition. It also finds the mobile layout cramped and some macOS features missing. |
| Top10VPNtop10vpn family | 9.2 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2026-10-22 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It finds the apps less approachable than some competitors and warns about renewal costs, paid feature upgrades and briefly retained connection timestamps. |
| RTINGSrtings family | Private Browsing: 6.4 / 10 Category ratings only; no overall score imported. | Paid service; see source for tested tierTested tier / platform: not captured / not captured | Published: 2025-03-31 Article update: 2026-07-13 Test window: Not captured | Separate review questions · Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | Its weaker Private Browsing verdict reflects concerns about policy clarity, registration privacy and business practices. Passing technical checks does not settle those questions. |
3 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
| Top10VPNtop10vpn family | 7.3 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2026-08-27 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It finds limited server coverage, no independent audit and few advanced privacy features. The article’s stronger enthusiasm for the free version should not be read as a paid-plan verdict. |
| TechRadarfuture family | 3.7 / 5Overall editorial rating | See publisher for tested planTested tier / platform: not captured / not captured | Published: 2022-04-13 Article update: 2024-12-18 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It finds the network small, the mobile apps less polished and streaming access inconsistent. Its separate free-plan comments do not establish what the paid subscription includes. |
| PCWorldpcworld family | 4 / 5Overall editorial rating The article discusses both free and premium tiers; this is its overall verdict, not a separately scored paid-only result. Its older Swiss location description predates the 2026 operator transfer. | Mixed free/premium articleTested tier / platform: not captured / not captured | Published: 2024-09-27 Article update: 2025-06-18 Test window: Not captured | Separate paid and free · Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It identifies the small network, limited advanced features and lack of an independent audit as drawbacks. |
4 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
| TechRadarfuture family | 3.95 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-10-24 Article update: 2026-07-03 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | Its negatives include uneven support responsiveness and weak phishing protection. It sees less recent innovation than in some rivals. |
| Tom’s Guidefuture family | 4.05 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-11-07 Article update: 2026-08-20 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports less impressive speeds and some streaming difficulties, with a learning curve in the apps. Its captured feature list also flags an Android split-tunneling gap. |
| Top10VPNtop10vpn family | 8.7 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2026-10-14 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It finds international streaming less consistent and the many settings demanding for beginners. Its favorable privacy verdict differs from RTINGS’ assessment. |
| RTINGSrtings family | Private Browsing: 5.6 / 10 Category ratings only; no overall score imported. | Paid service; see source for tested tierTested tier / platform: not captured / not captured | Published: 2025-03-31 Article update: 2026-07-13 Test window: Not captured | Separate review questions · Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports unencrypted traffic escaping after a system restart in its kill-switch test. Policy, audit-scope and ownership concerns also lower its Private Browsing assessment. |
3 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
| Top10VPNtop10vpn family | 7.4 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2026-08-19 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports congestion on distant connections, no in-app WireGuard support and limited external evidence for the logging policy. |
| TechRadarfuture family | 2.5 / 5Overall editorial rating | See publisher for tested planTested tier / platform: not captured / not captured | Published: 2022-04-14 Article update: 2023-08-15 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports DNS exposure with some protocols and no security audit. WireGuard requires manual configuration rather than the reviewed apps. |
| PCWorldpcworld family | 3.5 / 5Overall editorial rating | Paid serviceTested tier / platform: not captured / not captured | Published: 2024-03-25 Article update: 2025-06-19 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It sees the service falling behind on newer features, with a small server network and no WireGuard support in the reviewed app. |
4 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
| TechRadarfuture family | 4.2 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-10-24 Article update: 2026-10-23 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports weaker OpenVPN speeds, streaming trouble with NetShield and some kill-switch compatibility issues. Protocol and settings matter to these results. |
| Tom’s Guidefuture family | 4.35 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-11-02 Article update: 2026-10-28 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It finds standard pricing less competitive and the apps and help material potentially daunting for beginners. Free-tier praise does not describe paid-plan features. |
| Top10VPNtop10vpn family | 9.1 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2026-10-14 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports variable download performance and some playback frame-rate drops. The article also reviews the free tier, whose streaming and torrenting restrictions are different. |
| RTINGSrtings family | Private Browsing: 7.5 / 10 Category ratings only; no overall score imported. | Paid service; see source for tested tierTested tier / platform: not captured / not captured | Published: 2025-03-31 Article update: 2026-10-14 Test window: Not captured | Separate review questions · Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It is less positive about policy readability and subscription-retention practices. Its Private Browsing score combines policy judgments with technical findings; it is not an overall rating. |
3 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
| Top10VPNtop10vpn family | 6.5 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2024-11-25 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It is critical of upgrade prompts, paid extras, inconsistent streaming and privacy assurances. Its tests also raise a concern about kill-switch reliability. |
| TechRadarfuture family | 4 / 5Overall editorial rating | See publisher for tested planTested tier / platform: not captured / not captured | Published: 2022-10-24 Article update: 2025-03-07 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports uneven app behavior, particularly on Android, and only average connection performance. Some highlighted extras require an additional payment. |
| PCWorldpcworld family | 3.5 / 5Overall editorial rating | Paid serviceTested tier / platform: not captured / not captured | Published: 2025-05-19 Article update: 2025-05-19 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | Its verdict is held back by the historical privacy record, uneven streaming access and persistent sales prompts for additional features. |
4 records · 3 provisional source families
These verdicts can emphasize different strengths and limitations. Compare their caveats and original categories; the table below preserves their original scales. Eligible provider-level opinions are normalized only in the separate editorial metascore calculation. Shared ownership does not prove shared test data, but it prevents treating every logo as another independent vote.
Scroll horizontally for scope, actual test context, exclusion reasons and differing caveats.
| Publication / family | Original rating / category | Plan / platform | Article dates ≠ test dates | Checks still outstanding | Caveat in the source verdict |
|---|---|---|---|---|---|
| TechRadarfuture family | 4.25 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-10-24 Article update: 2026-10-06 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | The review warns about high automatic-renewal charges and finds stronger streaming alternatives in its comparison set. |
| Tom’s Guidefuture family | 4.45 / 5Overall editorial rating | Not capturedTested tier / platform: not captured / not captured | Published: 2022-10-28 Article update: 2026-04-10 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It flags steep renewal increases and a kill-switch issue under stress testing, while saying the IP address did not leak in that test. Port forwarding is absent. |
| Top10VPNtop10vpn family | 9.0 / 10Overall editorial rating · paid service | Paid service; exact subscription not capturedTested tier / platform: not captured / not captured | Published: Not captured Article update: 2026-09-14 Test window: Not captured | Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | The reviewer finds day-to-day usability friction and some performance gaps compared with its leading alternatives. Introductory value is not the renewal bill. |
| RTINGSrtings family | Private Browsing: 5.4 / 10 Category ratings only; no overall score imported. | Paid service; see source for tested tierTested tier / platform: not captured / not captured | Published: 2025-03-31 Article update: 2026-07-13 Test window: Not captured | Separate review questions · Exact tested tier · Actual tested platform · Actual test window · Test-data lineage | It reports kill-switch leaks after crashes or restarts and raises concerns about retained connection metadata. These findings explain its less favorable Private Browsing verdict. |
The counts and record-level reasons above are generated from the same curated review, performance and exact-tier data used elsewhere on this site. No paid rating is inherited by a Free tier, no category is promoted to an overall verdict, and no article update is substituted for a test date.
Download checklist results · Original observation fields · Read the published editorial method and its limits.
