A 32.73% weekly range and a -7.79% hour collide with a fixed daily loss limit
Should an intraday mean-reversion desk cut its position size so its fixed daily loss limit still binds at the intended loss, or keep its size and accept that a single -7.79% hour can stop it out?
The market facts are real; the decision maker is illustrative.
The week and the position
In the week of 24 June 2019, BTC ranged 32.73%, from a high of 13,970.00 late on Wednesday 26 June to a low of 10,525.10 late on Thursday 27 June, and closed 10,854.10, down 0.48%. Wednesday was the best day at 10.76%; Thursday the worst at -13.50%. The largest one-hour move was -7.79% at 20:00 UTC on Wednesday. Dollar volume ran 3.8 times its prior 12-week average, and ETH fell further than BTC. An illustrative mean-reversion desk buys dips under a fixed daily loss limit.
Judgment
The limit was sized for calmer weeks, and this week's -7.79% hour plus a -13.50% day would carry a dip-buying book through it before any reversion can occur. Keep the limit unchanged and cut size until it binds at the intended loss under this week's hourly and daily moves, accepting smaller rebound profits. The close sat at 9.6% of the week's range, in the lower third, while the week returned -0.48%.
What would change it
If hourly moves and realised volatility, now at 155.9% annualised, contract towards prior weeks' levels for several consecutive weeks, the original size can resume under the same limit. If the desk's own fills show stop-outs inside -7.79% hours, size must fall further.
What sets this case apart
A fixed loss limit is not fixed risk when intraday moves reach -7.79% in an hour; the decision is what position size makes that limit bind at the intended loss.