Let me share my own position logic: I am genuinely bullish on storage and memory fundamentals—AI creates a rigid demand for memory, and big companies are investing tens of billions to expand capacity, so the logic is very solid. But being bullish doesn't mean blindly going all in. My approach is to keep a base position and gradually reduce it in batches on rebounds above my cost, rather than chasing to add positions below cost to lower the average price. Retail investors love to "buy more as prices fall, turning small positions into large ones," but by the time fundamentals haven't yet materialized, they get liquidated first. Low-frequency large bets mean heavily backing the right direction but always leaving an exit for each leg, not going all in betting you won't be wrong. Direction and survival are two different things.