When the trader looks at a cell, they have to be able to trust it. A blank cell is honest — it says the system doesn't know. A cell that shows data that has quietly gone stale is a trade error waiting to happen. So the engine is built to fail loud: a contract repeated throughout the code is that stale data is worse than no data. On a dropped connection, realtime prices are designed to read #N/A rather than holding a last good value — so a stale price surfaces as a visible gap rather than a silently frozen number.

Holding to that design principle — across broken connections and the undocumented quirks on both sides of the bridge — is the part of production engineering that real money makes non-negotiable.

That principle, and the war stories behind it, are the subject of the writing →

Interactive Brokers ships a sample Excel RTD server with the TWS API. Per IBKR's documentation, it is "not intended to be used as production level trading tools." It is a teaching artifact that streams top-of-book quotes, with callbacks for account values, positions, and orders left as empty stubs.

Closing the gap means implementing the API to subscribe and receive all market-data fields, account values, positions, and orders; supporting multiple accounts; implementing COM-compliant multithreading so the stream survives even when Excel's UI is busy or blocked; hardening against TWS exceptions — detecting silent disconnects and rate limits, reconnecting and resubscribing automatically; late-seeding and deduplicating between RTD topics and TWS subscriptions; and ensuring multiple Excel instances can run RTD streams against the same TWS. That's the work StreamXLS does — at first for my own desk, now for anyone with the same needs.

The full case for why a spreadsheet is the right host is its own essay. Read it →

— StreamXLS