Local shuffles gave no clear positive magnitude gradient in Monte Cristo time-direction scores on days 1 to 30

We tested a magnitude gradient of future-minus-past nearby earthquake counts in a fixed USGS Monte Cristo Range catalog. We held elapsed day, 10 km area and preferred magnitude type fixed. The primary sample has 635 anchors at magnitude at least 2.5 on days 1 to 30. Neighbors lie within 10 minutes and 5 km. The within-group slope was 0.0420 events per magnitude unit. A one-sided magnitude shuffle gave p = 0.24505 and a central 95% reference range of -0.1218 to 0.1213. The range is not an effect confidence interval. Thus the primary test gave no clear positive gradient. A planned check that omitted the first minute gave slope 0.1365 and p = 0.008, but Holm adjustment across nine settings gave p = 0.072. The study extends a Ridgecrest test to a separate sequence. It does not show no physical magnitude effect or a difference between sequences. Sparse local groups, missed events, preferred magnitude scales, location error and the exchange assumption limit inference. The package gives the source snapshot, exact hashes, code, all shuffle draws, a figure and an independent count and variance audit.

Monte Cristo Range · magnitude gradient · time direction · conditional permutation · aftershocks · null result