Skip to main content
Validraft

What you can validate

News sentiment & NLP

Transformer-scored news as a tradable signal.

News-sentiment strategies turn headlines and filings into signals via NLP scoring (FinBERT). We validate sentiment drift, overreaction reversals, and news-volume spikes with point-in-time timestamps and stale-news filters.

  • US equities

20 example backtests

  1. 01

    FinBERT headline-sentiment drift on US equities

    Sample reportComing soon
  2. 02

    News-volume spike momentum study

    Sample reportComing soon
  3. 03

    Negative-sentiment overreaction reversal

    Sample reportComing soon
  4. 04

    Earnings-call tone vs subsequent drift

    Sample reportComing soon
  5. 05

    Press-release sentiment continuation

    Sample reportComing soon
  6. 06

    Sentiment dispersion as a risk filter

    Sample reportComing soon
  7. 07

    Sector-level news-sentiment rotation

    Sample reportComing soon
  8. 08

    Sentiment-confirmed breakout entries

    Sample reportComing soon
  9. 09

    Macro-news sentiment overlay on index futures

    Sample reportComing soon
  10. 10

    Same-day sentiment vs next-day return

    Sample reportComing soon
  11. 11

    Sentiment-weighted small-cap momentum

    Sample reportComing soon
  12. 12

    News-shock volatility expansion study

    Sample reportComing soon
  13. 13

    Stale-news fade (already-priced) detector

    Sample reportComing soon
  14. 14

    Analyst-language sentiment factor

    Sample reportComing soon
  15. 15

    Sentiment trend (rolling 5-day) signal

    Sample reportComing soon
  16. 16

    Headline-count anomaly screen

    Sample reportComing soon
  17. 17

    Sentiment-conditioned mean reversion

    Sample reportComing soon
  18. 18

    Cross-source sentiment agreement filter

    Sample reportComing soon
  19. 19

    Sentiment-momentum top-decile basket

    Sample reportComing soon
  20. 20

    Event-tagged sentiment reaction by category

    Sample reportComing soon

Have a news & sentiment idea of your own?

Send us the rules in plain English. We confirm scope and feasibility, then validate it on the same institutional infrastructure.