Order types

Definition

An order type is the instruction set attached to a trade: what price is acceptable, how long the instruction lives, and what triggers it. The taxonomy is small — market, limit, stop, stop-limit, plus time-in-force and venue modifiers — but the differences carry the KB’s most literal risk content: the flash-crash and August-2015 records are largely stories about order types executing exactly as specified into conditions their users never imagined.

How it works / structure

• Market order: execute now at whatever the book offers — certainty of fill, no bound on price; its cost is the spread in calm and the vacuum in stress (ms-liquidity).
• Limit order: execute at the stated price or better — bounded price, no guaranteed fill; a marketable limit (priced through the touch) buys near-certainty of fill WITH a disaster cap, the platform’s default execution primitive.
• Stop (stop-market): dormant until the trigger trades/quotes, then becomes a MARKET order — the documented dislocation amplifier (mgmt-stop-loss, episode-2015-etf-dislocation).
• Stop-limit: triggers into a limit instead — caps the fill price but can miss entirely in a gap (the protection/certainty trade is explicit, not escapable).
• Time-in-force & modifiers (engine-relevant): day vs GTC (resting GTC stops age badly as context changes); IOC/FOK (fill-or-vanish primitives for algos); market-on-open/close (auction participation — ms-sessions-auctions, the benchmark-flow order types); extended-hours limits (thin-book sessions force limit usage).
• Options specifics: market orders in wide-spread contracts donate the spread; multi-leg orders execute as native spreads at net prices (the structural alternative to legging risk).

When it applies

Every execution — the order type IS the last risk decision before the market takes over. The platform’s parameterized defaults: limit or marketable-limit for entries; stop-limit or alert-plus-decision over resting stop-markets; auction orders for size at benchmarks; never market orders in options, extended hours, or dislocations.

Risk profile & failure modes

• Market-order vacuum risk: the flash-crash penny and $100,000 prints were market orders meeting empty books — certainty of fill is certainty of SOME price, not a fair one.
• Stop-market cascades: clustered stops trigger each other into gaps — documented in every dislocation episode this KB carries.
• Stop-limit stranding: the gap through both trigger and limit leaves the position unprotected exactly when protection was the point — pairing with position-size discipline (risk-fixed-fractional) is the honest fix.
• GTC staleness: orders resting weeks execute against news their placer never saw — expiry hygiene is risk hygiene.

Evidence & limits

Order-type mechanics are exchange- and SEC-documented; the failure modes are documented in the episode record (2010, 2015). Venue-specific behaviors (auction eligibility, odd-lot handling) vary and change — the entry carries the stable taxonomy, not per-venue minutiae.

Falsifiable-thesis examples

Illustrations only, not signals:

• “Replacing resting stop-markets with stop-limits plus alerts changes this book’s dislocation-day realized slippage by less than 10bp in normal months while capping worst-day fills (order-policy audit)” — falsified by the replay comparison.
• “Marketable-limit entries fill within one tick of arrival mid over 95% of the time in liquid names (default-primitive check)” — falsified by the fill log.

Cross-references

• The liquidity reality: ms-liquidity; the sessions: ms-sessions-auctions
• The protection debate: mgmt-stop-loss
• The scaled-up machinery: ms-execution-algos
• The case law: episode-flash-crash-2010, episode-2015-etf-dislocation