"Would you buy those positions that you bought a long time ago at their prices today? Are they undervalued today? If not, then you might consider selling them."
Those are the questions I ask every time I look at my portfolio. Nothing is out of play. All you can do is evaluate the probabilities and the stochastics (real stochastics, not that "technical analysis" BS) to decide if a transaction is worth the costs and limitations, and if taking or holding a position is worth the exposure.