Live signal validation

Does the Gap Score
predict anything?

We measure two things: whether mainstream media covers the story within 7 days (gap closure), and whether predicted price moves occur within 72 hours (signal strength). This page shows both — live, with case-level drill-down.

We're early. Sample sizes are small. Read the methodology and draw your own conclusions.

59 stories trackedUpdated Sep 12, 2026

Gap closure rate

Did mainstream media cover the story within 7 days of detection?
Showing: All verticals
Early signal
10% closure · n = 10
High gap score
score > 70
Confidence: low — rate may shift meaningfully as data accumulates (need n ≥ 20 to stabilise).
Building sample size
2 of 5 minimum stories measured
Mid gap score
score 40–70
Measurements begin 7 days after a story is tracked. Updates nightly.
Building sample size
2 of 5 minimum stories measured
Low gap score
score < 40
Measurements begin 7 days after a story is tracked. Updates nightly.

Related stock move at T+7d

Avg absolute % change for stories with ticker entities
High gap score
score > 70
n = 0
±5.39%
Mid gap score
score 40–70
n = 3
Low gap score
score < 40
n = 0

Individual story outcomes

0 of 0 measured gaps closed · showing 12 most recent
Canada Deals Across
Canadian Politics · detected Sep 11
71
at detection
pending (1d)
War Russia China
World · detected Sep 11
76
at detection
pending (1d)
Built Free App
AI Tech · detected Sep 11· AMZN
82
at detection
+0%
T+7d
pending (1d)
Ukraine Japan Taiwan
World · detected Sep 11
76
at detection
pending (2d)
Recession GDP Unemployment Data
Finance · detected Sep 11
64
at detection
pending (2d)
Tape 101 Liminal
AI Tech · detected Sep 10
76
at detection
pending (2d)
Tested Magnesium
Health · detected Sep 10
62
at detection
pending (2d)
Foreigners Fight Ukraine
World · detected Sep 10
83
at detection
pending (2d)
Canadian Party Quebec
Canadian Politics · detected Sep 10
76
at detection
pending (2d)
Crosspost Movies
World · detected Sep 10
64
at detection
pending (2d)
1660 Days Russia
World · detected Sep 10
62
at detection
pending (3d)
Ukraine Figured Hurt
World · detected Sep 9
70
at detection
pending (3d)

Sorted by detection date · misses included

What we don't claim

Correlation ≠ causation

A high Gap Score means social media discussed something before mainstream media covered it. It does not mean GapWatch caused the story to break, or that the social discussion was accurate.

📉
We show misses

High-gap stories sometimes go nowhere. The denominators in every stat include stories where the predicted outcome didn't materialise. We show the full count.

📊
Small samples

Stats with n < 20 are labelled "Early signal." Below n < 5, we don't publish a rate at all — just a progress count. The sample grows nightly as more stories cross their 7-day measurement window.

🔬
What we measure

Gap Score = ratio of social discussion to mainstream media coverage at detection time. Outcomes = what happened in the 7–30 days after. The relationship is empirical, not theoretical.

How are gap scores computed? Read the methodology →

Outcomes updated nightly · 2:00 AM UTC · last run Sep 12, 2026