The recommendation scoring database is currently empty. Run the nightly scoring generation scripts to precompute factor scores.
Twin Engine Methodology
Enter any TSE ticker. The algo extracts that company's "DNA" — its percentile rank across 5 dimensions — then scores every other company on the same dimensions to locate its mathematical market twin.
- Growth — Sales YoY%, Net Income YoY%, and 3yr Sales CAGR. Percentile rank vs full universe.
- Profitability — Operating margin (level + trend), ROE (level + trend). Trend = change over available annual filings.
- Momentum — 1yr and 3yr price returns vs universe.
- Capital Actions — Buyback, dividend increase, activist stake, TSE compliance (last 12 months).
- Balance Sheet — Equity ratio and net cash ratio.
Weight logic: Each dimension weight = max(0, ref_percentile − 50) normalised to 100%. If the reference company scores 90th percentile on Growth but 45th on Balance Sheet, Growth gets ~100% weight and Balance Sheet gets 0%. Equal weights apply if the reference is below average on all dimensions.
Data: EDINET annual filings, TDNet capital signals, TSE governance compliance, major shareholder reports. Prices from daily snapshots (45-day lag tolerance).