Scaling into weakness is legitimate when it is planned: the levels, the total size, and the invalidation are all decided before the first entry, and the full position is within normal risk.
The rescue version has none of that. Size was not planned, the addition is justified by the loss itself, and the invalidation moves to accommodate the new average. That is escalation-of-commitment with an order ticket, and it is how single positions end accounts.
The test is documentary. If the addition appears in the written plan with its level and size, it is strategy. If it was invented after the position went against you, it is a rescue.
Related: escalation-of-commitment, martingale-mindset, sunk-cost-fallacy, doubling-down