Intelligence
How PREVICT earns conviction
Live engine readiness, independent agreement and the evidence gates behind every forecast.
Forecast Council
Eligible sources
0
Independent views
0
Shadow P(up)
—
Disagreement
Not measurable
No governed consensus yet
Not enough eligible, fresh evidence for a Council read.
SPY · 1M · would abstain
What prevents promotion
- · Needs a second independent eligible source
- · Shadow track record is still building
Edge Engines
- RESEARCH
1 · Opening-range continuation
Research effect documented; point-in-time replication has not started.
Conditional intraday continuation after a qualified opening range.
0/200 resolved · Not Council eligible
Evidence
B
Deployability
D
- RESEARCH
2 · Liquidity-conditioned reversal
Research effect documented; liquidity feature plane is not available.
Short-term reversal conditioned on liquidity and market quality.
0/200 resolved · Not Council eligible
Evidence
B
Deployability
D
- CANDIDATE
3 · Earnings underreaction
Live research reads remain visible; delayed or missing inputs reduce timeliness and coverage.
Post-earnings drift candidates from point-in-time results, guidance and reaction.
0/200 resolved · Not Council eligible
Evidence
B
Deployability
C
- RESEARCH
4 · Cross-sectional momentum
The feature plane is available; the independent reducer and walk-forward protocol are not frozen.
Medium-term trend and relative-strength hypotheses across the equity universe.
0/200 resolved · Not Council eligible
Evidence
A
Deployability
C
- CANDIDATE
5 · Volatility dislocation
The production chain adapter is configured; its entitled read-only session is unavailable.
Implied volatility versus forecast volatility and event-implied movement.
0/200 resolved · Not Council eligible
Evidence
A
Deployability
D
- RESEARCH
6 · Quality-moat value
Method specified; vintage-aware fundamentals and replication are not complete.
Long-horizon quality, value and panic-accumulation research sleeve.
0/200 resolved · Not Council eligible
Evidence
B
Deployability
D
Promotion law
1 · Produce
Candidate and shadow engines may publish research reads, but cannot influence official forecasts.
2 · Resolve
Every point-in-time output resolves against the same horizon and cost-aware baselines.
3 · Earn weight
Only positive measured skill, freshness and independence earn Council weight.