Leak or Bad Luck? Using Control Limits on a Losing Stretch

[Shadow Trader]

Every strategy with a win rate below one hundred percent produces losing runs, so a bad week alone proves very little. The challenge is separating a genuine leak, meaning a repeatable behavior or flaw that costs money, from ordinary variance.

One approach borrows from quality control. Using the historical hit rate and trade count, a trader can estimate how often a run as poor as the current one would occur by chance. A streak that falls within the range seen in a few reshuffled copies of the same history is consistent with luck. One far outside that range deserves a closer look.

Variance has no pattern, whereas a leak usually does. Tags help here: if losses concentrate in a single setup, a single hour, or trades taken right after a prior loss, the cluster is a clue. Losses spread evenly across every tag look more like noise.

Execution data is the other check. If average slippage, spread paid, or deviation from planned entries has drifted, the cause may be costs rather than the idea. Process errors such as ignored rules are leaks even when the outcome happens to be positive.

Both mistakes are expensive: fixing what is not broken, and ignoring what is.