Benchmark — does the watch keep time?

A source-linked retrospective benchmark with a placebo control: we checked whether 3,445 documented life events from 570 famous lives fell inside the Vimshottari dasha periods the engine ranked strongest, and compared that against two controls.

Result

Real charts: 80.2% of events inside the top-30 ranked periods.
Placebo charts: 81.9% — random birth data scored slightly higher.
Random placement: 79.4%.

The real-minus-placebo difference falls inside the confidence interval, so this benchmark shows no chart-specific skill. Any earlier reading of these numbers as evidence of accuracy was wrong.

What this is not

This is a retrospective event-alignment benchmark, not a measure of prediction accuracy. The "blind test" set was inspected during method development and is labelled as a retrospective evaluation set. No causal claim, no individual reliability claim, no health/finance/legal guidance.

See the full methodology · Download the spec (JSON) · Download the results (JSON)