A 12.47% Sunday broke the fading rule that worked on Tuesday
Should a mean-reversion desk keep fading ETH's intraday drops at its usual size, or suspend those trades until a down day actually reverses?
The market facts are real; the decision maker is illustrative.
The week and the position
In the week of 31 March 2025, ETH fell 12.56% to close at 1,580.76, in the lower third of its 27.28% range. Tuesday gained 4.53%, the week's best day, but Sunday lost 12.47%, the worst day, with a -3.95% hour at 18:00 UTC. Realised volatility ran at 79.8% annualised, volume was below the prior 12-week average, and ETH underperformed BTC by 7.75 points, with three of the four previous weeks down.
Judgment
A reversion position taken into Tuesday's 4.53% gain would have met a 12.47% drop on Sunday, a loss far larger than the gain the rule targets. Suspend fades of multi-percent hourly declines and keep only reversion positions sized from the 3.95% worst hour rather than from the smaller moves seen midweek. Accept missing rebounds like Tuesday's 4.53% in exchange for keeping the daily loss limit intact, since the trend, the weak volume and the 9.0th-percentile weekly return all indicate drops are continuing rather than snapping back.
What would change it
Reinstate the full reversion sizing after a week in which ETH closes in the upper part of its range on volume above the 12-week average, or after two consecutive down days that end higher than they opened.
What sets this case apart
The desk's rule targets small intraday bounces like Tuesday's 4.53% gain and is exposed to a 12.47% daily drop; the trigger is the size of one down day, not of the weekly loss.