Every experimental signal is judged on what actually happened: did it fire before a real funding round, and by how many days? Below is every resolved prediction — the lead is measured, not claimed.
Every resolved prediction the signal, and the raise it preceded
| Company | Signal | Fired | Raise announced | Lead |
|---|---|---|---|---|
| Modal | TLS certificate transparency | 2026-07-23 | 2026-10-14 | 83d early |
| General Intuition | TLS certificate transparency | 2026-08-09 | 2026-09-30 | 52d early |
| Legora | TLS certificate transparency | 2026-09-02 | 2026-10-21 | 49d early |
| Harvey | investor_commentary | 2026-07-24 | 2026-09-09 | 47d early |
| Mistral AI | TLS certificate transparency | 2026-08-16 | 2026-09-08 | 23d early |
| Nscale | TLS certificate transparency | 2026-09-03 | 2026-09-25 | 22d early |
| Supabase | TLS certificate transparency | 2026-09-03 | 2026-09-22 | 19d early |
| Anthropic late-stage | SEC Form D — secondary demand | 2026-07-24 | 2026-07-30 | 6d early |
| Etched | SEC Form D — secondary demand | 2026-08-13 | 2026-08-18 | 5d early |
| Gallatin AI | Federal contract awards | 2026-10-08 | 2026-10-09 | 1d early |
Only private venture companies are credited. Lead = calendar days from when the signal fired to when the raise was announced. Directional and capped at 365 days — a signal only counts if the raise followed it within a credible window. Misses count as fully as hits in the ledger's rates.