Recent census · resolved subset
( DECODER )
An independent record.Of public claims.traced through
AN INDEPENDENT AUDIT OF @ARAVIND’S GEOPOLITICAL CALLS. WHAT HE SAID. WHAT HAPPENED. WHAT THE EVIDENCE SUPPORTS.
OUR OWN FORECASTS, WITH DATES AND PROBABILITIES. OPEN TO THE SAME SCRUTINY.
Explore all forecastsThe shape of the record.
EVERY CLAIM COUNTS / ALL AUDIT SETSOUR FORECASTS / NEXT DEADLINES
What happens next?
Our own calls, made before the outcome. A probability, a deadline, and a record we can be held to.
01 / The headline numbers
Four numbers to remember
Each figure below describes a specific sample, not the whole archive.
Posted after the news broke
India protest calls that landed
Audience growth
02 / The audit
All unique audited claims. Only hits, partials and misses count as resolved.
% overall strict hit rate = hits ÷ resolved claims. vague, unresolved or after-the-fact claims are excluded. Inspect these claims ↗
Every audited call
The entire library, including vague and unresolved claims. These are excluded from the resolved hit rate.
When calls were made
Audits cluster in 2026
Audited calls per month. Darker is more.
03 / The sample
Pick your sample, change the story
Historical audit subsets, which may overlap. Strict counts hits only; half credit counts partials as half a hit.
04 / What makes a call look good
Revisited, public and vague wins
Resurfacing effect
Follow-up and outcome of each call.
Where calls land
Box size is calls. Color is hit rate.
Was it already news?
Hit rate by how public the idea was.
Did it have a deadline?
Hit rate by deadline type.
05 / Timing
First post follows the first report
Hours after the first public report
The pattern
Hours behind, usually quoting an old call.
06 / Reach
The audience grew anyway
Growth came in steps, around big news.
Median views per post, by month
Likes follow views
Same curve, steeper.
07 / Our calls
Each has a deadline and a probability. We score them all.
Probability by deadline
Scoreboard
08 / Go deeper