Research workspace
Edge Search
Test what happened after hourly price moves, funding settlements or daily high–low overreactions. Compare the outcome with normal market observations and inspect every event behind the result.
How the comparison works
Price events use the hourly close and enter at the next open. Funding events use exchange-reported settlement timestamps and enter at the first hourly open strictly after settlement. The feed does not show when the rate reached a trader. A funding event is compared only with another observed settlement in the same UTC hour of day.
Outcomes measure the change in price of a modelled long position, regardless of event direction. Both groups subtract the same round-trip cost. Funding payments, actual fills and margin are outside this model.
Within each sample, the earliest complete event windows take priority. Each eligible event is paired with the closest earlier non-event at the same UTC hour within 14 days. Windows with gaps or shared candles are excluded. The ledger shows unmatched events and exclusions.
The percentiles describe observed returns. They are not confidence intervals. Both samples are visible, and repeated tests on these dates can overfit them.
Daily high–low overreaction mode uses a separate fixed 30-day, 1-standard-deviation rule and compares next-day momentum and counter-move outcomes with normal-range days. The candle direction proxy and venue data differ from the original study.
These tools explain calculations and historical evidence under the assumptions shown. They are not forecasts, trading advice or proof of future performance.