Is increasing position size after a winning streak rational?

A profitable streak might indicate that a strategy is performing well under current conditions, but it can also occur through normal variance.

Increasing risk because of recent wins may create overconfidence, while keeping risk constant can sometimes ignore meaningful evidence that a strategy's edge has changed.

Would you ever adjust risk based on a rolling sample of expectancy and drawdown, or do you believe position size should remain fixed unless the strategy is formally revalidated? What evidence would justify a change?