PDS Trader: Options Strategy Analysis and Portfolio Modeling

Building multi-leg spreads strike by strike in broker UIs is slow and combinatorial. There is no way to see every viable spread across a strike range in one view

No single grid for Entry Cost, Downside Risk, Upside Risk, Downside Potential, and Upside Potential across every spread variant

No way to project a Black-Scholes-shaped P/L curve per spread without exporting to a separate tool

Selected positions could not move cleanly between an analysis tool and the execution platform without re-keying

Without a personal risk filter, the spread matrix is overwhelming and traders cannot narrow to spreads inside their risk budget

Combinatorial spread enumeration

Strategies define each leg as a strike range, not a single strike, so the engine generates every valid (leg1, leg2, ..., legN) combination. One cached option chain fetch feeds all legs, reversed ranges are normalized, empty filtered sets short-circuit, and a List-aggregating Cartesian product returns an empty list rather than crashing on misconfigured inputs.

Hand-rolled options pricing for the risk graph

The projection sweeps the underlying price by ten percent in either direction per spread, including legs that have already expired at the projection point. We built the normal CDF on top of an Abramowitz and Stegun erf approximation, a Black-Scholes-shaped curve for pre-expiry legs, and a piecewise payoff for at-expiry legs, with adaptive sweep step size for low and high priced names.

Atomic strategy save with favourite invalidation

Favourites are stored as comma-separated indexes into a deterministically ordered Cartesian product, so any change to leg count, expiration, strike range, qty, or option type silently invalidates them. The save runs the strategy update, the legs delete-and-reinsert, and the favourite-reset decision inside a single SQL transaction, so partial saves cannot leave dangling favourites pointing at a different surface.

Compressed cross-service option chain transport

Option chains are heavy. The internal cache server returns gzipped Base64 payloads, which the API decompresses, JSON-deserializes, and either Base64 decodes or treats as raw bytes depending on payload shape. The Generate Spread endpoint in turn gzip-compresses the full spread plus risk-graph payload before returning it to the SPA, keeping wire size manageable for large strike grids.

Dual-product SSO and bi-directional TradePartner sync

PDS Trader does not own its user store. Login federates to the SmartTrading identity provider, mirrors a local AppUser row on first sign-in, derives Free vs Paid from the latest subscription, blocks Free users at the door, and issues two JWTs in the response (one scoped to PDS, one to TradePartner) so the same SPA session can talk to both products and both SignalR hubs. TradePartner can POST a trade into PDS via API-key auth and have a portfolio auto-created; PDS can push a selected portfolio back to TradePartner as a watchlist.