Design a delta-hedging schedule for a long-vol or short-vol options position — discrete rebalance bands, expected gamma-scalp P&L, transaction-cost drag, and realized-vol breakeven. Use this skill when the user holds options and wants to manage directional exposure dynamically. Triggers: "delta hedge", "gamma scalp", "rebalance band", "realized vol breakeven", "how often should I hedge", "DvegaDtime", "gamma trading P&L", "is my long-vol position profitable". Defaults: hedge in shares, threshold |Δ| = 0.05 per 100 contracts, broker cost $0.005/share.
How this skill is triggered — by the user, by Claude, or both
Slash command
/option-risk-management:delta-hedgingThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Design and simulate a delta-hedging strategy for an option position. Decompose realized P&L into gamma scalp + vega + theta + transaction costs.
Design and simulate a delta-hedging strategy for an option position. Decompose realized P&L into gamma scalp + vega + theta + transaction costs.
Required inputs:
Use greeks-calculator for the per-leg Greeks. Net them:
Δ_net = Σ (sign · qty · 100 · Δ_leg)
Γ_net = Σ (sign · qty · 100 · Γ_leg) # raw gamma per $1 of spot
ν_net = Σ (sign · qty · 100 · ν_leg) # per vol point
Θ_net = Σ (sign · qty · 100 · Θ_leg) # per calendar day
Δ_net is the share-equivalent. To get flat, short Δ_net shares (or buy −Δ_net if negative).
Discrete delta-hedging incurs gamma scalping P&L PLUS transaction cost drag. The classic band is:
Δ_band = α · σ · S · √(T_period / 252)
where α is a tuning parameter (commonly α = 1, conservative). Rebalance whenever |Δ_net| exceeds the band. Smaller band → more scalp P&L AND more cost.
Practical defaults:
Realized variance pays gamma; implied variance is what you paid for it. Per share of underlying, per day:
expected_P&L_per_day = 0.5 · Γ · S² · (σ_realized² − σ_implied²) · (1/252)
Scaled by position size and re-evaluated daily.
If σ_realized > σ_implied → long-gamma trades profit. If σ_realized < σ_implied → short-gamma trades profit.
The realized vol at which gamma scalp P&L exactly offsets theta:
σ_RV_breakeven = √( σ_implied² − (Θ_net / (0.5 · Γ_net · S²)) · 252 )
Long-vol holders: realized vol must exceed σ_RV_breakeven to break even on the day. Short-vol holders: realized vol must stay BELOW it.
If (Θ_net / (0.5 · Γ_net · S²)) makes the radical negative, the breakeven is undefined — usually means a position with both signs of gamma across legs. Decompose by leg.
For a long-gamma position with band α·σ·S·√(dt), each rebalance trades ~Δ_band shares. Daily expected cost:
N_hedges_per_day ≈ 1 / α² # for Brownian motion
daily_TC = N_hedges · Δ_band · cost_per_share
This is rough — empirically validated with simulation. The narrower the band (smaller α), the more N_hedges grows, and the cost-of-trading ate the scalp P&L fast.
For more accuracy, run a Monte Carlo or use historical underlying paths to simulate the band strategy and report distribution of daily P&L.
import numpy as np
def sim_delta_hedge(S0, sigma_realized, T_days, alpha, gamma0, theta_per_day, cost_per_share, n_paths=1000):
dt = 1/252
pnls = []
for _ in range(n_paths):
S = S0
pnl = 0.0
for d in range(T_days):
# underlying move
ret = np.random.normal(0, sigma_realized*np.sqrt(dt))
S_new = S * np.exp(ret)
# gamma scalp (continuous approximation)
pnl += 0.5 * gamma0 * (S_new - S)**2
# theta
pnl += theta_per_day
# rebalance whenever |delta| exceeded band — assume on average alpha rebalances per day
band_size = alpha * sigma_realized * S * np.sqrt(dt)
pnl -= band_size * cost_per_share / alpha # rough
S = S_new
pnls.append(pnl)
return np.array(pnls)
This simulation is a first approximation — real Greeks change with spot and time, so for a 30+ DTE position re-compute Greeks daily.
For sizing the position, hand off to position-sizing. For per-leg Greeks intuition, see greeks-calculator.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub dongzhuoyao/finance-option-skills --plugin option-risk-management