Self-attribution keeps the ego intact and the feedback loop broken. If every winner proves your method and every loser proves the market was rigged that day, no amount of experience will improve you.
The usual signature is asymmetric vocabulary in a journal. Wins are described with verbs you own: I read the tape, I waited, I sized properly. Losses are described with things happening to you: it got stop-hunted, the news hit, the algos ran it.
A working review assigns both kinds of causes to both kinds of outcomes. Ask what luck contributed to the win and what you did to earn the loss. Grading execution separately from result, as in a behaviour-journal, forces the question.
Related: outcome-bias