The last few trades are vivid and the last few hundred are not. After three wins a method feels reliable, after three losses it feels broken, and decisions about size, strategy and confidence swing accordingly, although three trades say almost nothing about an edge.
It is closely tied to streaks. A run of results in either direction is expected to happen by chance in any long sequence, but it feels like information about the future. Raising size after a winning run, or abandoning a sound setup after a losing run, are the two common results.
The remedy is to judge a method by a defined sample and not by its latest results. Decide in advance how many trades you need before you will evaluate a setup, say fifty or a hundred, and review it by the average result in R over the whole batch. Changes are then made on a schedule, not in the heat of the last session.
It also helps to keep a record that makes older results as visible as new ones. A weekly or monthly summary that always compares the current period with the full history keeps the recent period in proportion.
A rule of thumb is to ask whether you would make the same decision if the last three trades had not happened.