A fixed daily loss limit met by a -3.01% hour: shrink size or loosen the rule?
Should the mean-reversion desk keep its daily loss limit unchanged and trade smaller, or raise the limit so one violent hour does not stop the day?
The market facts are real; the decision maker is illustrative.
The week and the position
In the week of 14 June 2021, ETH fell 10.59% to 2,243.47 inside a 29.37% range, at 85.1% annualised realised volatility, with a largest hourly move of -3.01% on Sunday. Wednesday was the worst day at -6.91%; Sunday was the best at +3.61%, yet the close sat only 33.8% of the way up the week's range. Volume was 0.5 times the prior 12-week average, ETH lagged BTC by 1.82 percentage points, and price was 48.22% below its 52-week high. An illustrative intraday mean-reversion desk buys dips under a fixed daily loss limit.
Judgment
A daily limit hit by a single -3.01% hour means the position is too large for this volatility, not that the limit is wrong. Keep the limit and cut size so one hour like Sunday's, or a day like Wednesday's -6.91%, consumes a set fraction of it. Dollar volume of 10.61 billion was 0.5 times the prior 12-week average, and the data records no reversion depth, so size must come from the hourly and daily moves seen. Accept missing the +3.61% rebound that came after the week's low.
What would change it
Restore the original size if ETH volume returns towards its 12-week average and the largest hourly move stays well below this week's -3.01% for a full week. Repeatedly negative 8-hour funding alongside a price recovery would also show the cuts came from volatility, not a persistent downtrend.
What sets this case apart
The limit is tested by single-hour moves arriving faster than any reversion under thin volume, not by the week's total loss.