Fading one-hour drops into a -9.43% day on a desk with a daily loss limit
Should the mean-reversion desk keep buying ETH's one-hour selloffs at its usual size, or cut trade size after a week where Saturday alone fell 9.43%?
The market facts are real; the decision maker is illustrative.
The week and the position
In the week of 26 January 2026, ETH fell 19.41%, from 2816.90 to 2270.15, closing in the lower third of its 37.07% range, with realised volatility at 78.1% and volume 1.32 times its prior 12-week average. Monday gained 4.03%, but Saturday lost 9.43% and contained the largest one-hour move, -3.44%. An illustrative intraday mean-reversion desk buys dips under a fixed daily loss limit. ETH also underperformed BTC by 8.22 percentage points while hourly correlation stayed close at 0.86.
Judgment
Mean reversion paid on Monday and broke on Saturday: a -3.44% hour inside a -9.43% day means fresh dip-buying can lose several times before the day ends, and the fixed daily loss limit forces exits at the worst prices. Cut per-trade size so one day of repeated losses stays inside the limit, and fade only one-hour drops that hold above the prior hour's low rather than every drop. Accept fewer trades and lower return this week; the limit itself stays unchanged.
What would change it
If daily ranges shrink so no single day loses more than roughly the size a reduced trade can absorb, size can return to normal. If ETH closes mid-range rather than in the lower third, or the ETH-BTC ratio stops falling, dip-buying regains its premise.
What sets this case apart
The loss limit, not the direction call, sets the trade size; the case turns on one day's compounding of small hourly losses.