How should you size the hedge leg in a pair trade?

The correct size depends on what you are trying to neutralize. If the goal is broad beta removal, the hedge should be sized against benchmark sensitivity. If the goal is to soften event-driven tape noise, scenario sizing may matter more than a historical regression.

Most hedge sizing methods are just different ways of translating risk into notional.

A theoretically perfect hedge ratio is not useful if the hedge leg is expensive to trade, hard to rebalance, or likely to move the book. The practical size is always the theoretical size filtered through execution reality.

Hedge size drifts as volatility changes, event probability moves, or one leg becomes more sensitive than the other. That is why sizing and rebalancing belong in the same research loop instead of being treated as separate implementation chores.