Once you hold a position, a thesis feels like a possession. Evidence in its favour is read as meaningful; evidence against it is explained away as noise or a temporary move. Nickerson's 1998 review described this as one of the most pervasive patterns in human reasoning.
In trading it appears as moving a stop to avoid being wrong, reading only the analysts who agree with you, or remembering the times a setup worked and not the times it failed. It also appears in journals: a trader who writes mainly about the trades that confirm a favourite method will find that the method works.
Two habits reduce it. First, define the invalidation before you enter: the specific price or condition at which the idea is wrong. A rule fixed in advance cannot be argued with after the fact. Second, review every trade of a given setup, not selected examples, and compute the average result in R across all of them.
Deliberately writing the best case against a trade before taking it is a useful exercise, as is asking what you would expect to see if you were wrong and checking whether you are seeing it.
No process removes the bias entirely. The aim is to build a record and rules that make it harder for the bias to decide quietly.