A -9.69% hour and a +9.6%/-10.51% day pair test a fixed daily loss limit
Should an intraday mean-reversion desk keep trading ETH at its normal size behind a fixed daily loss limit, or cut size so that one -9.69% hour cannot exhaust the limit before a reversal arrives?
The market facts are real; the decision maker is illustrative.
The week and the position
In the week of 10 May 2021, ETH fell 8.71% to close 3,582.10, in the lower third of a 30.8% range, after four straight up weeks (3.91%, 3.79%, 27.11%, 33.05%). Realised annualised volatility reached 158.4% and dollar volume was 2.82 times the prior 12-week average. The largest one-hour move was -9.69% at 23:00 UTC Wednesday, and the day after a +9.6% best day (14 May) came a -10.51% worst day (15 May). The desk enters on weakness expecting a snapback.
Judgment
Mean reversion is a plausible edge here: the close sits in the lower third of the range and the desk's premise is a snapback from stretched moves. But a -9.69% hour and a -10.51% day mean the fixed loss limit is the binding constraint, not the forecast. Keep the strategy and limit, and cut per-trade size so that a move like 12 May's 23:00 UTC hour consumes only a small part of the daily limit, leaving room for the reversal. Stay out of the late UTC session, where both the week's largest hour and its low occurred. Accept fewer trades and lower profit while volatility stays near 158.4%.
What would change it
If hourly moves and daily ranges shrink back towards the calmer prior weeks, or if volume falls back towards its 12-week average, size can be restored towards normal. If the loss limit is hit on a single hour like 12 May's -9.69% before any bounce, size must be cut again.
What sets this case apart
The decision is about sizing inside a fixed limit during a volatility regime shift, not about forecasting ETH's direction after a four-week rally.