A 95th-percentile week tests whether a breakout signal is signal or one-off
Should the research team keep the breakout signal live at full weight after this week, or downgrade it to provisional until it replicates outside the 64-week sample?
The market facts are real; the decision maker is illustrative.
The week and the position
In the week of 6 May 2019, BTC rose 20.68%, from 5,773.18 to 6,967.31, over a 33.86% range, closing in the upper third (close_location_pct 70.9) on dollar volume of 2.04 billion, 2.05 times its prior 12-week average. The low came first (Monday, 5,619.14), the high on Sunday (7,521.78), and Saturday alone gained 11.0%. Three of four prior weeks were up. The market still sat 20.87% below its 52-week high, and the week ranked in the 95th return percentile of 64 weeks.
Judgment
The signal's confirming conditions aligned: a low before the high, a close in the upper third, and volume well above the 12-week average. But one day supplied 11.0 of the 20.68 points, the week sits at the 95th percentile of a 64-week sample, and Sunday closed down 1.54% with a -2.65% hour. Keep the signal live, but at base weight, not boosted, and treat this week's return as an outlier observation rather than a new baseline for sizing.
What would change it
If the following week closes back below the 5,773.18 open on average-or-lower volume, the signal is suspended until it fires again. If a second week closes above 6,967.31 with volume above its 12-week average, the signal is confirmed at full weight.
What sets this case apart
The question is whether the signal's own confirmations and a single dominant day outweigh a 95th-percentile return, not whether the price will rise.