Met300 simulated women
The test at a glance
- What it tests
- Consistency: whether a voice that learned from its world's lived records reproduces that world's birth rates.
- Data
- The same 300 simulated women aged 18-45 in the world on UK 2019 tables (UN WPP 2024); the voice trained on 34,367 women's records from twelve other worlds on the same tables.
- Held out
- Not a held-out test; the 300 women's next year was never simulated or read, and the world they live in was never used for training.
- Compared against
- The UK 2019 birth rates the world was built on, and the prompted voice (5.24).
- Bar
- Total fertility within 0.20 AND mean age at childbearing within 1.0 year.
- Result
- 1.75 births per woman against 1.62, mean age 30.3 against 30.7: consistent.
A voice trained on the records of lives in its own world gave that world's birth rates: 1.75 births per woman against 1.62, and a mean age at childbearing of 30.3 against 30.7, inside both bars. Asked the same question about the same 300 women, a prompted voice had said 5.24. The voice learned from twelve other worlds built on the same UK 2019 tables, never this one, and was registered with its checksum before any of the 300 women was asked; each was asked once. In this world births depend on age alone, and the voice adds nothing beyond it: partnered and single women get the same answer.
Protocol, result and scorer