Crowd Alpha
A policy game against an unknown population.
Every round narrows the field while adding a new dimension of evidence — from an open online contest to a two-day final at IIT Kanpur.
Each gate dissolves most of the field
One contest, three ways to produce a quantitative decision.
A policy game against an unknown population.
Probability & optimisation, expressed in code.
Query a hidden system. Infer the rule. Decide.
Angles to scale — 1% = 3.6°
| Module | Weight | Time |
|---|---|---|
| Crowd Alpha | 20 percent | 30 min |
| Algorithmic Quant | 40 percent | 60 min |
| Blackbox Lab | 40 percent | 60 min |
Percentile-normalised before weighting. All three modules must be attempted.
Authorship interviews — not personality rounds.
The interview verifies authorship — it does not rank personality.
Fewer cues. Greater depth. A sharper cut.
Exact modules and scoring remain confidential until the reveal.
Two days, one arena — the format stays sealed until the reveal.
The final format
The detailed format of the Systems Arena — contests, weights and the day-wise programme — is kept sealed for now. It is revealed to verified finalists and partners as the championship approaches.
What is public: two days · IIT Kanpur · 13–14 March 2027 · 50 finalists