Every call published before market open, verified against actual closes. No cherry-picking. No hindsight edits.
From 7-day to 30-day windows, performance remains consistently strong in the 89–91% range.
Strongest in large-cap and mega-cap names most actively traded by institutions.
We publish misses because trust comes from transparency, not perfection.
Misses concentrate in highly volatile mega-caps during earnings windows. Confidence scoring flagged 73% of failed calls with lower confidence.
Benchmarked against common industry performance ranges.
A prediction is counted as successful if the target-date close lands inside the predicted high-low band, OR if the directional call is correct.
This standard is stricter than direction-only benchmarking used by many quantitative teams.
- All predictions generated before their target dates
- Actual prices sourced from verified market close data
- Every prediction in the period is included — no cherry-picking
- Duplicates removed; N/A (delisted/halted) excluded from win rate
- Raw records are available for independent audit
- 14 proprietary research modules
- Multi-layer reasoning framework
- 7-day, 14-day & 30-day horizons
- 226 unique stocks tracked
- Enhanced prediction system with deeper signals
- Improved analytical architecture
- Expanded stock coverage & real-time data
- Built on Model 1 learnings — iterating on outcomes
Founder pricing: $25.99/month · Locked through 2027