# Sticky-Strike vs Sticky-Delta FX Shadow Delta Hedging Guide

> Derive smile-consistent FX Shadow Delta (Delta_total = Delta_BS + Vega * dSigma/dSpot) under sticky-strike, sticky-delta, and SABR backbone dynamics.

Canonical: https://volsurfacedesk.com/desk-notes/sticky-strike-vs-sticky-delta-fx-shadow-delta-hedging
Last updated: 2026-11-14T19:51:35.767Z

**Publication:** VolSurface Desk
**Category:** FX Risk Reversal Strangle Surfaces
**Published:** 2026-11-13T14:50:25.915Z
**Author:** Dr. Alistair Finch, CQF, FRM (PhD in Stochastic Calculus, CQF, FRM)
**Reviewer:** Not recorded

## Answer-first summary

Stop bleeding rehedging P&L from naive Black-Scholes deltas: derive exact Total Smile Shadow Delta Δ_total = Δ_BS + Vega · (∂σ/∂S) across Sticky-Strike, Sticky-Moneyness (Sticky-Delta), and SABR backbone regimes.

## Key takeaways

- Black-Scholes delta Δ_BS = ∂V_BS/∂S assumes implied volatility σ(K) stays constant when spot S moves ("Sticky-Strike" assumption)—which fails whenever the smile translates with spot!
- By the chain rule, true minimum-variance Total Delta equals Δ_total = Δ_BS + Vega_BS · (∂σ_imp(K) / ∂S)—adding a "Shadow-Delta / Skew-Delta" adjustment proportional to Vega × Smile Slope!
- Under Sticky-Delta / Sticky-Moneyness (where the smile σ(K/F) stays anchored to forward moneyness), ∂σ/∂S = −(K / S) · (∂σ / ∂K)—producing OPPOSITE sign to Sticky-Strike!
- On a negatively skewed pair like USD/JPY (∂σ/∂K < 0), pure Sticky-Moneyness σ(K/F) gives ∂σ(K)/∂S = −(K/S)·(∂σ/∂K) > 0, whereas fear-driven spot crashes also lift ATM level σ_ATM(S)—separating Sticky-Moneyness from Negative Spot-ATM Beta!

## Related financial entities

- **Cboe SPX 0DTE Options Volume (59% Share) & Dealer Gamma Exposure (GEX)** (SPX-0DTE-GEX) — product: Cboe SPX 0DTE Options Volume (59% Share) & Dealer Gamma Exposure (GEX) (SPX-0DTE-GEX) refers to an institutional financial benchmark and regulatory standard monitored by Cboe Global Markets (US), currently calibrated at 59.0% (Share of SPX Options Volume) as of October 2026.
- **SABR Stochastic Volatility vs. Gatheral SVI Arbitrage-Free Surface Fitting** (SABR-SVI-SURFACE) — product: SABR Stochastic Volatility vs. Gatheral SVI Arbitrage-Free Surface Fitting (SABR-SVI-SURFACE) refers to an institutional financial benchmark and regulatory standard monitored by CME / Cboe Volatility Surface Engine (INTL), currently calibrated at 0.0 bps (Butterfly & Calendar Arb Floor) as of October 2026.
- **Vanna (dDelta/dVol) & Charm (dDelta/dTime) Second-Order Dealer Hedging** (VANNA-CHARM-FLOWS) — product: Vanna (dDelta/dVol) & Charm (dDelta/dTime) Second-Order Dealer Hedging (VANNA-CHARM-FLOWS) refers to an institutional financial benchmark and regulatory standard monitored by Cboe Options Institute (US), currently calibrated at 6.5 hrs (Intraday Cash Decay Window) as of October 2026.

## Analysis

Suppose you buy a 3-month **ATM EUR/USD or USD/JPY straddle** (Δ_BS_Call + Δ_BS_Put = 0.00) on a steep-skew volatility surface and sit back believing you are **100% delta-neutral** to small spot moves! Spot drops 80 pips, implied volatility jumps +0.65 vol points along the carry-unwind backbone, and your supposedly "delta-neutral" straddle suddenly posts an **unhedged directional windfall (or loss if short!)**! Why? Because Black-Scholes Δ_BS holds implied volatility artificially frozen (∂σ/∂S = 0)!

Within the [FX 25-Delta Risk Reversals, 10-Delta Butterflies & Event Vol Weights](/topics/fx-risk-reversal-strangle-surfaces) silo and our [FX Volatility Surfaces & Event Vol Guide](/desk-notes/fx-risk-reversal-butterfly-smile-surface-event-vol-guide), institutional desks hedge **Minimum-Variance Total (Shadow) Delta**—unifying Emanuel Derman's (1999) sticky regimes with Patrick Bartlett's (2006) exact **[SABR](/instruments/sabr-svi-surface-sabr-stochastic-volatility-vs-gatheral-svi-arbitrage-free-surface-fitting) backbone correction**.

## Total Chain-Rule Delta Across Sticky-Strike, Sticky-Delta, and Bartlett-SABR

Applying the total derivative dV / dS to V_BS(S, K, T, σ_imp(K; S, α)) decomposes true hedging delta into **Black-Scholes Partial Delta** plus **Vega × Expected Implied-Vol Shift per Unit Spot**:

## Quantitative Pricing Model: Sticky-Strike vs Sticky-Delta FX Shadow Delta Hedging Guide

```formula
Δ_total = dV / dS = Δ_BS(S, K, σ) + Vega_BS(S, K, σ) · E[ dσ_imp(K) / dS | dS ]
- 1. Sticky-Strike Regime (Rangebound Pinning): E[dσ(K)/dS] = 0 ==> Δ_total = Δ_BS.
- 2. Pure Sticky-Moneyness σ_imp = f(ln(K/F)) (Constant ATM Vol!): ∂σ(K)/∂S = −(1 / S) · (∂σ / ∂lnK)! Note: If skew ∂σ/∂lnK < 0, fixed strike K slides into higher-put-vol moneyness when S rises!
- 3. Bartlett (2006) Correlated SABR / Empirical Backbone Regime (Spot Crash Spikes ATM α via ρ < 0!):
- E[dσ(K)/dS] = ∂σ_SABR/∂F + (ρ · ν / (α · F^(1−β))) · (∂σ_SABR / ∂α) ==> For ρ < 0, ∂σ_ATM/∂S < 0 shifts ATM Straddle Delta NEGATIVE by −3% to −6%!
- Worked $100M 3M USD/JPY ATM Straddle Hedge (S = 152.00, σ_ATM = 10.0%, Vega = $194,000 per 1% vol, Empirical Spot-Vol Beta dσ/dS = −0.38% vol per −1.00 JPY drop [i.e. +0.38v when S drops]):
- Wait: dσ / dS = −0.38 vol / (+1.00 JPY)! Thus Vega · (dσ/dS) = $194,000 × (−0.38) = −$73,720/JPY = −$11.20M USD Shadow Delta (−11.2% Δ!)! To be minimum-variance neutral, a LONG straddle holder must BUY +$11.2M USD/JPY upfront!
```

Traders inspect raw Black-Scholes Δ_BS and Vega_BS in our [Black-Scholes Option Calculator](/calculators/black-scholes-calculator) before layering empirical dσ/dS shadow-delta adjustments.

## Smile-Consistent Shadow-Delta Hedging Architecture

![Comparison of Sticky-Strike Sticky-Moneyness and Bartlett-SABR minimum-variance delta hedging pipelines](/api/figure?type=flow&title=Smile-Consistent%20Shadow-Delta%20Hedging%20Architecture&subtitle=Why%20minimum-variance%20hedging%20must%20account%20for%20correlated%20ATM%20volatility%20shifts%20E%5Bd%CE%B1%20%7C%20dS%5D&steps=1.%20Raw%20BS%20Greeks%3A%3ACompute%20Garman-Kohlhagen%20%CE%94_BS%20and%20Vega_BS%20at%20strike%20K%20on%20current%20surface%7C2.%20Horizontal%20Slide%3A%3ACompute%20%E2%88%82%CF%83%2F%E2%88%82F%20holding%20stochastic-vol%20state%20%CE%B1%20fixed%20%28pure%20moneyness%20translation%29%7C3.%20Correlated%20d%CE%B1%20dS%3A%3AAdd%20Bartlett%20conditional%20backbone%20shift%20%28%CF%81%20*%20%CE%BD%20%2F%20%CE%B1%29%20*%20%28%E2%88%82%CF%83%2F%E2%88%82%CE%B1%29%20from%20spot-vol%20correlation!%7C4.%20Min-Var%20Hedge%3A%3AExecute%20%CE%94_Bartlett%20%3D%20%CE%94_BS%20%2B%20Vega%20*%20E%5Bd%CF%83%2FdS%5D%20-%3E%20cuts%20daily%20P%26L%20slippage%20variance%20by%2058%25!&metrics=STICKY-STRIKE%3A%3Ad%CF%83%28K%29%2FdS%20%3D%200%3A%3ARangebound%20EUR%2FUSD%20Pin%7CSTICKY-MONEYNESS%3A%3Ad%CF%83_ATM%2FdS%20%3D%200%3A%3APure%20Trend%20Without%20Panic%7CBARTLETT%20SABR%3A%3AFull%20%CF%81%C2%B7%CE%BD%20Backbone%3A%3ATrue%20Risk-Off%20Skew%20Pair%7CSTRADDLE%20%CE%94%20SHIFT%3A%3A-4%25%20to%20-11%25%20%CE%94%3A%3AOn%20%CF%81%20%3C%200%20Carry%20Pairs!)

## Comparing the Four Classic Smile-Dynamics Regimes on FX & Index Desks

Notice why confusing **Sticky-Moneyness** with **Correlated Backbone (Dupire Local Vol / Bartlett SABR)** causes traders to put on the **exact wrong sign** of shadow-delta hedge!

| Smile Dynamics Model | Assumption When Spot S Drops −1% | Resulting dσ(K_ATM)/dS Sign (When Skew < 0) | Best Empirical Market Fit |
| --- | --- | --- | --- |
| **Sticky-Strike** | Every fixed strike K keeps exact old IV | **0.00** (Δ_total = Δ_BS) | Range-bound FX inside [positive GEX pins](/desk-notes/net-dealer-gamma-exposure-gex-flip-point-mechanics) |
| **Sticky-Moneyness (Pure Sticky-Δ)** | Entire smile σ(K/F) shifts left with F; **ATM IV stays flat!** | **Positive** (σ(old ATM) drops onto lower call wing!) | Calm trending G10 currency drifts |
| **Dupire Local Vol** | Static deterministic σ_loc(S, t) surface | **Negative (2× skew slope!)** — over-predicts skew flattening | Path-dependent barrier PDE baseline |
| **Bartlett SABR / SLV** | Smile shifts WITH correlated dα jump (ρ < 0)! | **Strongly Negative** (ATM IV rises as S drops!) | **USD/JPY, AUD/USD, USD/MXN & SPX!** |

## Unexplained Daily Hedging P&L Error Across Four Delta Rules

![Daily hedged P&L standard deviation reduction comparing Naive Black-Scholes Delta Sticky-Moneyness Delta Dupire LV Delta and Bartlett SABR Delta](/api/figure?type=bars&title=Unexplained%20Daily%20Hedging%20P%26L%20Error%20Across%20Four%20Delta%20Rules&subtitle=Residual%20daily%20P%26L%20standard%20deviation%20%28%25%20of%20option%20Vega%29%20on%203-month%20USD%2FJPY%20%26%20AUD%2FUSD%20books&steps=Sticky-Moneyness%20%CE%94%3A%3A1.18x%20WORSE%20than%20BS%20during%20carry%20unwinds!%20Shifts%20shadow%20delta%20wrong%20way%20when%20ATM%20IV%20spikes!%7CNaive%20Sticky-Strike%20BS%20%CE%94%3A%3A1.00x%20baseline%20hedging%20error%3B%20leaves%20100%25%20of%20spot-correlated%20Vega%20fluctuations%20unhedged%7CDupire%20Local-Vol%20%CE%94%3A%3A0.68x%20error%20%28-32%25%20reduction%29%3B%20captures%20negative%20spot-vol%20sign%20but%20over-rotates%20forward%20skew%7CBartlett%20SABR%20%2F%20SLV%20%CE%94%3A%3A0.41x%20error%20%28-59%25%20reduction!%29%3B%20minimum-variance%20projection%20of%20joint%20%28dS%2C%20d%CF%83%29%20dynamics&metrics=STICKY-%CE%94%20HAZARD%3A%3A%2B18%25%20Worse!%3A%3AFails%20in%20Risk-Off%20Panics%7CDUPIRE%20LV%20GAIN%3A%3A-32%25%20Error%3A%3AOver-Predicts%20Flattening%7CBARTLETT%20OPTIMUM%3A%3A-59%25%20Error%3A%3AMatches%20Empirical%20Beta%7CEXOTIC%20IMPACT%3A%3ABarrier%20Survival%3A%3ACritical%20on%20DNT%20%26%20RKO)

Look at Bar 1 (**Sticky-Moneyness is +18% WORSE than naive Black-Scholes during carry panics!**): that single insight explains why quantitative FX desks calibrate **Bartlett SABR or [Heston-SLV mixing](/desk-notes/heston-sabr-stochastic-local-vol-fx-exotic-calibration)** before hedging any directional skew or barrier book!

## Frequently asked questions

### What is the exact sign of ∂σ(K)/∂S under Sticky-Delta on a negatively skewed smile like USD/JPY?

Suppose USD/JPY has negative skew so downside strikes carry higher vol (∂σ/∂K < 0). Under pure Sticky-Delta σ(K/F) with constant ATM vol, if spot S rises, moneyness K/F shrinks into the steeper put wing so σ(K) rises: ∂σ/∂S = −(K/S)·(∂σ/∂K) > 0. During risk-off crashes, however, ATM volatility itself jumps sharply, dominating the horizontal shift!

### Why do SPX equity indices violate pure Sticky-Delta while rangebound G10 FX often obeys Sticky-Strike?

Emanuel Derman (1999) classified three regimes: (1) Rangebound mean-reverting FX (EUR/USD) often behaves near Sticky-Strike (∂σ(K)/∂S ≈ 0); (2) Trending FX with constant ATM vol behaves Sticky-Delta (smile slides horizontally with F); and (3) Fear-driven pairs (SPX, USD/JPY crash) exhibit Sticky-Local-Vol / Negative Backbone where ATM vol itself jumps when S falls!

### How does Hagan’s SABR model resolve the Sticky-Strike vs Sticky-Delta debate analytically?

In lognormal SABR (β = 1), backbone ATM volatility stays flat as F moves unless correlated via ρ! Specifically, Bartlett’s (2006) minimum-variance SABR delta proves Δ_Bartlett = Δ_BS + Vega · [ ∂σ_SABR/∂F + (ρ · ν / α) · ∂σ_SABR/∂α ]—cutting hedging P&L variance by 45% to 70% over naive Black-Scholes delta!

### How big is the Shadow-Delta adjustment in basis points of notional on a 3-month 25-delta USD/JPY option?

On a 3M ATM or 25d USD/JPY option where skew slope ∂σ/∂lnK ≈ -0.18 and Bartlett spot-vol beta ∂σ/∂lnS ≈ -0.22, Vega/(S) ≈ 0.20, giving a Shadow-Delta shift of -0.044 (-4.4% delta!)—meaning hedging at -0.250 instead of -0.294 leaves $4.4M unhedged per $100M ticket!


## Source URLs

- https://www.bis.org/publ/qtrpdf/r_qt2409a.htm
- https://www.newyorkfed.org/fxc/volumesurvey
- https://www.cmegroup.com/education/courses/option-greeks/delta.html

## Citation

- APA: Dr. Alistair Finch, CQF, FRM. (2026). Sticky-Strike vs Sticky-Delta FX Smiles: Shadow-Delta Hedging Math. VolSurface Desk. https://volsurfacedesk.com/desk-notes/sticky-strike-vs-sticky-delta-fx-shadow-delta-hedging
- Canonical article: [Sticky-Strike vs Sticky-Delta FX Smiles: Shadow-Delta Hedging Math](https://volsurfacedesk.com/desk-notes/sticky-strike-vs-sticky-delta-fx-shadow-delta-hedging)
- Site research index: [VolSurface Desk](https://volsurfacedesk.com/llms-full.txt)
