In excess of any more stretch of time, There is certainly hardly a statistically significant autocorrelation in significant frequency returns. If there was, then the above might be applicable which might dampen the impact. $begingroup$ I estimate day-to-day pnl on the CDS place using the spread c