Everything we have written about building and running an automated options strategy, in the order it tends to matter. It is written from running execution software on live options markets rather than from theory, which is why it concentrates on the parts that decide outcomes and most coverage skips.
If you are deciding whether to automate at all, the honest sequence is: understand what a backtest can and cannot prove, learn what paper trading hides, then look at the broker plumbing. Most people do this in reverse, buy tooling first, and discover the constraints afterwards.
Data entitlements, symbology, chain discovery, rate limits and failure behaviour, in the order the work comes in.
Gamma dominates, decay becomes a clock, and liquidity collapses away from the money.
What one loss really costs, and why consecutive losses decide survival.
Historical replay, payoff modellers and execution simulators answer different questions. The fill assumption is where results get invented.
Six diagnostics, every one of them found by building execution software.
What paper fixes that live does not: fills, size and latency.
Why implied volatility on a thin strike is confident nonsense.
Every retail gamma figure rests on an assumption exchanges do not publish.
Market data entitlements, keys and the things that block a first run.
Why a harness can report success a live account never sees.
Eleven minutes of refused quotes mid-position, and the four fixes.
Fills, not signals. Almost every retail discussion is about picking direction. Almost everything that decides whether a system survives is execution: the fill assumption, the spread crossed twice, the gap between where a stop triggers and where it fills. Direction is the part people enjoy arguing about and the part that matters least.
Liquidity at the strike, at the moment of entry. A chain that looks fine before the bell can be unusable thirty seconds after it, and the reverse. Screening once at selection throws away good trades and still admits dead ones.
What happens when the data stops. Every automated system eventually loses its quote feed mid-position. What it does then should be a deliberate design decision. Ours closes a position it can no longer price, because a position you cannot see is one you cannot manage.
Options Sniper is free, runs paper or live, and goes open source on 30 October 2026. It deliberately ships with no strategy — a default strategy inside an open-source program is a published strategy — so you either build your own in its pack builder or license ours. Investors University is the other route: forty students, direct access to a coach, and a month of the licensed pack included.