A 98th-percentile +30.06% week leaves the mean-reversion desk fading into strength
Should the desk keep fading intraday ETH overshoots at its usual size this week, or suspend its short-into-strength trades and trade only pullbacks in the direction of the weekly rise?
The market facts are real; the decision maker is illustrative.
The week and the position
In the week of 20 July 2020, ETH rose 30.06% from 239.12 to 311.00, a return at the 98th percentile of 127 weeks, closing in the upper third of a 36.19% range and 1.65% below its 52-week high. The best day gained 9.35% while the worst lost only 1.3%; volume ran 1.66 times the prior 12-week average, and 8-hour funding averaged 0.0365%, peaking at 0.0944%. An illustrative intraday mean-reversion desk fades overshoots under a fixed daily loss limit. ETH outperformed BTC by 22.21 percentage points, and hourly correlation between the two was loose at 0.68.
Judgment
The desk's edge depends on dips that get bought back; this week offered almost none. The worst day lost 1.3% against a 9.35% best day, so every fade of strength faced a rising tape, though positive funding of 0.0365% on average, 0.0944% at peak and 0.01% at low per eight hours credited the short leg. Suspend the short-into-strength trades and keep only reversion entries taken against the weekly direction, at reduced size, accepting days with no trades. Do not convert to trend-following at the 98th percentile of 127 weeks.
What would change it
If a later week produces two or more down days exceeding the 1.3% fall of 20 July, or the largest hourly move turns negative and funding falls to 0.01% or below, the full short-fade programme returns. If price closes below the week's low of 234.05, the trend premise ends immediately.
What sets this case apart
The conflict is between a reversion strategy and a one-sided week, not between two directional forecasts; the funding credit to the short leg argues for restraint on price grounds alone.