moontower

Hedging on Implied Delta
What's My Bias?

When implied ≠ realized, your model-computed delta over- or under-hedges you. That bias has a regime where it's a feature, not a bug.

Core mechanic Higher vol → all deltas compress toward .50 (ITM lighter · OTM heavier)
OTM Options higher vol → heavier delta
If you hedge on implied and… Your Δ Prefers Why
Implied > Realized Heavy Chop Over-hedging scalps mean-reversion — more shares traded than "true" delta warrants
Implied < Realized Light Trend Under-hedging lets gamma run — fewer offsetting shares, profitable if direction persists
ITM Options higher vol → lighter delta
If you hedge on implied and… Your Δ Prefers Why
Implied > Realized Light Trend Under-hedging lets gamma run — fewer shares against a deep option, profitable if direction persists
Implied < Realized Heavy Chop Over-hedging scalps mean-reversion — more shares against a deep option, profitable if stock reverses

Heavy Δ → Chop
Over-hedging is profitable when the stock mean-reverts. You're scalping more aggressively than warranted.
Light Δ → Trend
Under-hedging is profitable when the stock trends. Less offset lets gamma profits compound.
Why "trend" is already in the framework
Daily close-to-close realized vol understates trending markets. Each daily return is small and same-signed — low RV. Sample less often (weekly) and the drift compounds — high RV. So "Implied > Realized" in a trending stock is partly a measurement artifact. The cheatsheet captures this: "Light Δ → prefers Trend" is exactly where daily RV misleads you.