Estimating Spread Option Greeks with Finite Differences

Summary

The document raises a practical question about estimating spread option Greeks by applying finite differences to Kirk’s spread option approximation. The author reports that, for options with deltas between 0.45 and 0.6, the resulting gamma estimates fall between 4 and 8. They interpret these values as implying very large changes in delta and exposure when multiplied by position quantity.

The post does not give a derivation, comparison against another pricing method, or an answer explaining whether the estimates are correct. It offers no details about the finite difference step size, contract units, model inputs, or how gamma is scaled. Those omissions make it impossible to assess the reported values from the document alone. The useful takeaway is that numerical Greek estimates should be examined in the context of their units, calculation settings, and practical exposure, especially when the result appears unusually large.

Key ideas

• The author estimates spread option Greeks by applying finite differences to Kirk’s approximation.
• They report gamma estimates between 4 and 8 for options with deltas between 0.45 and 0.6.
• The post questions whether these estimates imply implausibly large delta changes and position exposure.
• It provides no calculation details or evidence that would establish whether the gamma values are accurate.

Tags

# Kirk Spread Approximation, Greeks by Finite Difference


# Kirk Spread Approximation, Greeks by Finite Difference












I am using finite difference on Kirk's Approximation for Spread Options to estimate greeks of the Spread Option. Now this is creating an problem in the estimation of gamma. For at the money options (with delta between 0.45 to 0.6) I am getting Gamma values in the range of 4-8 (implying a 400%-800% change in delta).

I know gamma can go to infinity but in practicality I am getting huge gamma exposures (gamma * quantity). And this to me and the traders look incorrect.

I am desperately seeking some suggestion.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.