Fat-Finger Limits And Price Collars
A trader who means to sell 100 shares but types 100,000, or means $50.10 but types $5.010, needs a mechanical check to catch the error before the market does — fat-finger limits and price collars are that check, and they exist precisely because humans and algorithms both make this exact mistake.
A "fat-finger" error is any order that's wrong by an order of magnitude or more, typically because of a typo, a unit mix-up, or a bug in an algorithm's sizing logic — an order for 1,000,000 shares instead of 1,000, or a limit price of $5.00 instead of $50.00. These aren't rare edge cases; they happen regularly enough at every trading firm and exchange that catching them mechanically, rather than relying on someone noticing before hitting send, is standard practice everywhere.
Two separate checks, two separate mistakes
A fat-finger limit (also called a maximum order size check) simply rejects any order above a pre-set quantity or notional value threshold — if the desk's normal order size tops out around 5,000 shares, an order for 500,000 is almost certainly an error and gets bounced before it reaches the exchange. This catches size mistakes.
A price collar catches price mistakes instead: it rejects (or the exchange itself rejects) an order whose limit price is too far from the current market price — say, more than 5% away from the last trade or the current best bid/offer. This stops an order priced at $5 in a stock trading at $50 from ever executing, even if the order's size is perfectly normal. Collars are often set both by the client's own risk gateway and, independently, by the exchange itself, so a mistake that slips past one layer still has a second chance to be caught.
Why both checks are necessary and neither replaces the other
A fat-finger size limit does nothing to catch a wrong price on a normally-sized order, and a price collar does nothing to catch a wrong quantity at a correct price — the two errors are independent and need independent checks. This is also why serious market-access risk frameworks (in the US, formalized under SEC Rule 15c3-5, the Market Access Rule) require brokers to have both types of pre-trade check in place, rather than treating one as sufficient coverage for the other.
What breaks without them
Absent these checks, a single mistyped order can execute against a large fraction of the visible order book in an instant, moving the price sharply before anyone can react — exactly the kind of event that price collars and circuit breakers were built to prevent from cascading further once it starts, but that a well-placed fat-finger check would have stopped from ever executing at all.
Fat-finger limits catch orders of abnormal size; price collars catch orders at an abnormal price. They check independent things, and a robust order-entry system needs both, applied both at the client's own risk gateway and, as a second layer, at the exchange itself.
Setting fat-finger and collar thresholds too loosely "to avoid false rejects" defeats the purpose — a limit generous enough to never inconvenience normal trading is often generous enough to let a genuine error straight through, which is exactly the scenario these checks exist to prevent.
Practice in interviews
Further reading
- SEC Rule 15c3-5, Market Access Rule
- Harris, Trading and Exchanges, ch. 26