Options spread ranking that shows its work.

FerroSpread ranks candidate spreads only after construction and quality gates have produced a clean candidate set.

Why this is an engine problem, not a UI filter

!

A single score is not enough when spread quality depends on liquidity, width, Greeks, and pricing assumptions.

!

Ranking becomes fragile when model outputs cannot be traced back to leg-level inputs.

!

Users lose trust when the engine cannot explain why one spread outranks another.

How FerroSpread handles it

01

Start from constructed spreads

Ranking operates on spread candidates that already satisfy construction and leg-quality constraints.

02

Score on explicit objectives

Expected return, risk/reward, probability, capital efficiency, breakeven distance, and liquidity can be surfaced as separate components.

03

Return the explanation

The output keeps score components, dominant sensitivities, comparables, and assumption trails beside the candidate.

What makes this FerroSpread-shaped

Designed around Construct -> Rank -> Explain as one typed path.

Supports products where ranking must be inspectable.

Feeds retail and professional surfaces without changing the analytical substrate.

Common concerns

Can ranking criteria be product-specific?

Yes. FerroSpread is positioned around configurable ranking surfaces rather than one fixed score.

Does ranking hide the risk model?

No. Pricing and Greeks come from FerroRisk, and the ranked candidate carries the relevant analytical context forward.