Skip to content

Explain calibrated Heston parameters in plain English#13

Merged
SecondOrderEdge merged 2 commits into
mainfrom
claude/tab-population-KygrS
May 24, 2026
Merged

Explain calibrated Heston parameters in plain English#13
SecondOrderEdge merged 2 commits into
mainfrom
claude/tab-population-KygrS

Conversation

@SecondOrderEdge
Copy link
Copy Markdown
Owner

No description provided.

claude added 2 commits May 24, 2026 19:39
The single fixed initial guess, price-error objective, and absence of
parameter bounds let Levenberg-Marquardt settle in a degenerate corner
(theta=0, Feller violated, ~11 vol-pt RMSE) on steep equity skews.

Replace with a multi-start, bounded fit that minimises implied-vol error
and seeds v0/theta from the at-the-money and median variance of the
quotes. On steep market-like surfaces this drops vol RMSE from ~11 to
~0.5 with no theta collapse, while synthetic recovery stays at 0.008.
Surface the Feller condition on the page so stressed fits are visible.

https://claude.ai/code/session_01AXbuChtmt8Xo4cxEtKrady
Replace the raw parameter dump with headline derived readouts
(instantaneous vol √v0, long-run vol √θ, variance shock half-life) and a
dynamic per-parameter breakdown that interprets each fitted value, plus a
one-line gloss of the Merton jump parameters.

https://claude.ai/code/session_01AXbuChtmt8Xo4cxEtKrady
@SecondOrderEdge SecondOrderEdge merged commit 6e85c7c into main May 24, 2026
2 checks passed
@SecondOrderEdge SecondOrderEdge deleted the claude/tab-population-KygrS branch May 24, 2026 20:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants