A 91st-percentile week with 2.97x volume tests a signal outside its fitted range
Should the team keep its momentum signal live at full size after a week this extreme, or suspend live validation and run it in reduced size until volatility returns to normal?
The market facts are real; the decision maker is illustrative.
The week and the position
In the week of 4 January 2021, BTC rose 15.61% from 33,000.05 to 38,150.02, with a 49.13% range, a high of 41,950.00 on Friday 8 January 15:00 UTC, and a low of 28,130.00 on Monday 4 January 10:00 UTC. Dollar volume was 2.97 times the prior 12-week average, annualised realised vol reached 127.6%, and perpetual funding averaged 0.1199% per eight hours. The return sits in the 91st percentile of 151 weeks of history. ETH beat BTC by 12.59 percentage points with a 0.79 hourly correlation. An illustrative research team is deciding whether its signal survived the week.
Judgment
A week in the 91st percentile of 151, on 2.97 times normal volume and with positive funding in every eight-hour window, tells the team almost nothing about whether the signal works. A 15.61% rise rewards any long exposure taken at the open, so this week cannot separate the signal's contribution from the market's drift. Keep the signal live only at minimum size so execution and data quality are still exercised, and suspend any live pass/fail verdict. Attribute this week to the market regime, not to the signal. Re-open validation once a normal week arrives.
What would change it
If the next two weeks record returns inside the middle of the 151-week distribution and volume near 1.0 times the prior 12 weeks, full-size validation can resume. If funding stays at or above 0.2365% per eight hours while price holds near the 41,950.00 high, extend the suspension.
What sets this case apart
This is about whether a backtest edge can be judged during an extreme regime, not about the signal's direction or the week's profit.