An options liquidity screen answers one question: can this position be exited at a fair price when exiting matters? The practical screen has three parts: minimum open interest at the traded strikes, minimum daily volume judged against how much of the session has elapsed, and a ceiling on bid-ask width as a fraction of the mid price.
Entries are optional; exits are not. An entry into a thin market can simply be skipped, but a position already open in one must eventually negotiate with whoever is quoting, at whatever width they choose, on a day that may be exactly when everyone else wants out too. Liquidity screening at entry is therefore a promise about the exit: strikes with real open interest and tight quotes today are the ones most likely to have them under stress.
Raw daily-volume floors fail honestly at 9:35 AM: minutes into the session, even the most liquid strikes have printed a fraction of their daily volume, so a static floor rejects everything each morning and then waves the same strikes through by noon. Intraday volume follows a U-shaped curve, heavy at the open and close, thin midday. A session-aware floor requires only the fraction of the daily floor that the elapsed session could plausibly have produced, reaching the full requirement by the close. Open interest and quote width need no such scaling, since neither depends on the clock; quote quality polices every hour of the session, which is why Options Scanner enforces those two in full at all times.
Multi-leg structures exit as packages. An iron condor with one liquid wing and one dead one trades at the dead wing's width, because the package quote inherits the worst leg's uncertainty. Screening per leg and judging the structure by its weakest passes only positions whose entire exit path is real. The same principle demotes a beautiful expiry score on an illiquid weekly: a top-scoring expiration whose strikes fail the floors is a trap, and the honest selection takes the top-scoring expiration that clears them.