What an event study is
Align a discrete sky event to the next market bar and measure a target metric at a horizon. · 10 min
The Lab’s unit of work
- Sample the sky daily over a date range.
- Evaluate a
SkyCondition(aspect, angle window, moon phase, retrograde, feature threshold). - Extract entry events (first day the flag becomes true), optionally with a minimum gap.
- Align each event to the next market bar and compute a target metric at each horizon.
Method
The result is a historical association on that sample: mean, median, win rate, paths. It is not a forecast that “the next square will lift the index”.
Key points
every-samplemode keeps overlapping days and is labelled optimistic.- Horizons are 1 / 3 / 5 / 10 / 20 bars by default; each reports its own n.
- Fewer than 20 events at the primary horizon is INSUFFICIENT SAMPLE, never a confidence slogan.
Loading seed experiment 1…
References
- MacKinlay, Event Studies in Economics and Finance — MacKinlay A.C., Journal of Economic Literature 35(1), 1997, 13–39
- RAINBOW research methodology — docs/RESEARCH-METHODOLOGY.md