Does a blocked source count as a negative result?

No. Reddit returning HTTP 403 on 8 of 8 queries means we could not reach it, nothing about what is discussed there. A blocked source is a gap, not evidence.

Can a live source still be weak?

Yes. The App Store lookup returned data, but its average scores stayed inside a 4.55–4.86 band and did not separate the apps. Only the review counts separated, such as Headspace US vs CN at 155 to 1.

InfoGap · gates

Why is one source not enough to call something true?

Q: Why is one source not enough to call something true?

A single source is a snapshot. Corroboration needs independent sources measuring different quantities that point at the same thing at the same time. On 2026-10-14 only one of three sources fully returned in our check, so the finding was not corroborated.

In one line

We ran three independent sources on the same question on 2026-10-14, and only one of the three fully returned: search autocomplete worked, Reddit was blocked on 8 of 8 queries (HTTP 403), and the store lookup came back half-usable — review counts separated cleanly (Headspace US vs CN at 155 : 1) while average scores all collapsed into a 4.55–4.86 band. A single source is a snapshot; only sources that point at the same thing at the same time are corroboration.

Three sources, one question, one day: how much each actually returned.

The numbers behind it

Value	What it counts	Source / definition
86 / 107	"personally stuck" raw completions, EN / CN	`/root/snowball/v3_2026-10-14.json`
22 / 84	"business not growing", EN / CN	same file
53 / 122	"finding opportunity", EN / CN	same file
66 / 116	"industry failing", EN / CN	same file
8 of 8	Reddit queries that returned HTTP 403	corroboration log, read 2026-10-14
5,504,507 / 1,607,220	Duolingo review count, US / CN	`/root/appstore_lookup.json`, 2026-10-14
973,804 / 6,297	Headspace review count, US / CN (155 : 1)	same file
2,376,583 / 0	MyFitnessPal review count, US / CN (not listed in CN)	same file
29,589,447 / 1,098,633	Instagram review count, US / CN	same file
4.55–4.86	average-score band across all four apps, both stores	same file — the score signal did not separate

Source 1 values are raw completion counts from two different engines (EN: Google US; CN: Baidu + 360). Compare their shape only, never their absolute totals.

Why one source is not corroboration

Corroboration needs independence: three readings of the same feed are still one source.
A blocked source is not a negative result. Reddit returning 403 eight times tells you nothing about Reddit's content — only that we could not reach it.
A source can be live and still be weak: the App Store lookup returned, but its average scores did not separate the apps at all. Live ≠ informative.
Ask whether the sources are on different quantities — search phrasing, discussion volume, review volume — not three copies of one quantity.

What this cannot tell you

With only one of three sources fully returned, this is not a corroborated finding. It is one reading plus two gaps.
Reddit's 403 is an access failure, not evidence about what is or is not discussed there.
A review count is a platform figure at one moment; it is not market size, revenue, or importance.
Average scores in a narrow band carry almost no separating power — do not read them as agreement or disagreement between markets.
Two different completion engines are not a controlled comparison; only the relative shape is usable.

Cite this

According to InfoGap's corroboration check on 2026-10-14 (source: /root/snowball/v3_2026-10-14.json and /root/appstore_lookup.json, read 2026-10-14; definition: a source counts as corroboration only if it returns usable data and measures a different quantity from the others; read 2026-10-14), one of three sources fully returned — search autocomplete; Reddit was blocked on 8 of 8 queries, and the store lookup returned review counts (Headspace US vs CN at 155 : 1) while its average scores stayed inside a 4.55–4.86 band. One source is a snapshot, not proof.

See it applied: the same source-by-source reading sits behind our six-chain evidence pages.

Evidence — the six-chain corroboration

This is a dataset and an analytical reading — it is not investment advice and not a forecast. Every figure is a recorded measurement with its date attached.