A 10.89% hour inside a 19.17% week tests a fixed daily loss limit
Should the mean-reversion desk cut its trade size so the fixed daily loss limit survives another hour like Thursday's, or keep its sizing and accept that limit hits end the day early?
The market facts are real; the decision maker is illustrative.
The week and the position
In the week of 9 April 2018, BTC rose 19.17%, from 7,011.04 to 8,355.00, after a weekly low of 6,611.00 late on Monday and a 10.89% hour on Thursday 12 April, the week's best day at 13.94%. The close sat at 95.9% of the 27.51% range, in the upper third, with realised volatility at 103.7% annualised. Volume was below the prior 12-week average, and the market remained 28.74% below its 52-week high after four weeks that alternated direction. An illustrative intraday mean-reversion desk trades this market under a fixed daily loss limit.
Judgment
Mean reversion sells strength and buys weakness, which puts the desk on the losing side of Thursday's hour until prices revert, if they revert that day. A loss limit sized for calmer weeks converts one such hour into a forced stop before the reversion can pay. Cut trade size so the limit survives a 10.89% hour at full exposure, and stop fading the rally while the close holds in the upper third with the low preceding the high. Keep the limit itself unchanged, and accept missing part of a rising market.
What would change it
If hourly moves fall back towards the range of the four prior weeks and the close settles into the middle of the weekly range, normal size can resume. If the desk's own records show it held no position exposed to Thursday's hour, the size cut is unnecessary.
What sets this case apart
The decision is sizing against one hour of 10.89%, not a view on direction: the loss limit is what the week attacked.