Two-strategy comparison

Calendar Effect vs Event Repricing

Shared available sample: 24 months, 2024-11–2026-10. Missing observations are excluded, never filled with zero.

Monthly correlation

−0.46

Requires 12 paired months and non-zero variance

Both positive

11

Shared months

Opposite return signs

6

One positive, one negative

Zero-return months A / B

6 / 1

Zero return does not establish an empty position

Maximum month-end drawdown: Calendar Effect −6.75%; Event Repricing −0.57%. Both below a prior peak: 0 months. Only Calendar Effect below peak: 8; only Event Repricing: 1.

Complementary monthly returns do not establish a trading hedge or future diversification. Drawdowns use the arithmetic cumulative returns of the shared sample; gaps omit unknown returns. Month-end data hide intra-month losses. Backtest and mixed months are included and labelled; a short sample is uncertain. Metric definitions.