Good-Til-Cancelled vs Day Orders: What Happens When the Order Type Expires
You want 100 shares of a stock trading near $50, but only at $48. So on Friday, March 6, at 3:55 p.m. Eastern, you enter a limit order and close the laptop. Monday morning the order is gone, and the stock never traded below $49.
That’s a day order behaving exactly as designed. It expires at the close of the regular session it was entered in — 4:00 p.m. Eastern for US equities — filled or not, and nothing carries it into the next session. A good-til-cancelled order would have survived the weekend. But GTC does not mean forever either: most US brokers purge resting GTC orders after 60 or 90 calendar days, which is roughly 42 or 63 actual trading sessions once weekends and holidays come out.
The clock starts when you click, not when the day starts
The thing that expires is the instruction, not the position and not the account. A resting order is a standing message sitting on an exchange book or in your broker’s system, and expiry deletes the message.

Here is the part most people have backwards. A day order entered at 9:31 a.m. and a day order entered at 3:59 p.m. expire at the same instant. Entering later does not buy you more time — it buys you less. Your 3:55 p.m. limit on March 6 had five minutes of life, not a day, which is 1.3% of the 390-minute regular session.
GTC differs in one way and matches in another. It rolls from session to session instead of dying at the bell. It is still only live during the sessions your broker routes it into, so a GTC limit will not execute at 2 a.m. because a headline moved index futures overnight. FINRA’s investor overview of how stocks trade and what order instructions do is the regulator’s framing of this, and it’s worth reading before your broker’s marketing page.
Why a broker kills your GTC at all
Because “cancelled” has to mean something. A limit entered during one price regime firing two years later into a completely different one is a housekeeping problem, and brokers solve it with a hard cap. The number varies by firm, so you have to look up your own — but the two figures you will see almost everywhere are 60 and 90 calendar days.

Say you re-enter the $48 limit as a GTC on Monday, March 9, at a broker with a 60-day cap. It dies on May 8. You will not necessarily get an email about it.
Sixty calendar days is not sixty chances
Calendar days and trading days diverge fast. About 69% of calendar days are trading days in a normal US equity year — 252 out of 365, a ratio of 0.6904.

Every figure below assumes the US equity regular session, 9:30 a.m. to 4:00 p.m. Eastern; a 252-day trading year; holidays spread evenly at 9 per 252 sessions; and a window starting on a Monday. Start midweek and the counts move by a day in either direction, because the number of weekend days a window catches depends on which weekday it opens on.
Working the 90-day case line by line:
- 90 calendar days = 12 full weeks plus 6 days
- Weekend days in 12 weeks = 24
- Weekend days in the trailing Monday-to-Saturday stretch = 1
- Weekdays = 90 − 25 = 65
- Holidays at 9 per 252 sessions ≈ 65 × (9/252) = 2.3, call it 2
- Trading days ≈ 63
Across the range of caps brokers actually use:
| Broker GTC cap (calendar days) | Weekend days | Holidays | Trading days the order lives | Usable share of the window |
|---|---|---|---|---|
| 30 | 8 | 1 | 21 | 70.0% |
| 45 | 12 | 1 | 32 | 71.1% |
| 60 | 16 | 2 | 42 | 70.0% |
| 90 | 25 | 2 | 63 | 70.0% |
| 180 | 50 | 5 | 125 | 69.4% |
Find your broker’s cap in column one. Column four is how many chances your order actually gets at the market. Your March 9 order at a 60-day broker gets 42 sessions, not 60, and the difference is 18 days the market wasn’t open at all.
The administrative cost of holding one price
Here’s where the choice between order types stops being stylistic. If you want a single price level held continuously across some span of trading days, the number of separate order entries you need is that span divided by the life of one order, rounded up.
To keep the $48 bid resting for a full trading year:
- Day orders: 252 ÷ 1 = 252 entries, one every morning
- 60-day GTC: 252 ÷ 42 = 6 entries
- 90-day GTC: 252 ÷ 63 = 4 entries
Add one to each GTC figure if your start date doesn’t line up cleanly with the cap boundary, which it usually won’t. Stretch the horizon and the gap widens: two years of coverage is 504 day-order entries against 12 entries on a 60-day GTC. Every one of those 492 extra clicks is a chance to fat-finger a price, a quantity, or a side.
A sceptical reader will point out that nobody re-enters a day order 504 times. They place it when they think of it. Correct — and that is the real cost, not the clicking. Discretionary re-entry means your order isn’t there on the days you’re distracted, and price gaps have a way of landing on exactly those days.
What the expired order actually costs
Nothing directly. No broker charges a fee for an order that dies unfilled. The cost is the trade you weren’t in.
Follow the March 9 order forward. On April 14 the stock prints $47.50 for twenty minutes on a bad open, then recovers and closes the month back near $50. If the GTC was live, you filled 100 shares at $48.00 — a $4,800 basis against a $5,000 mark, $200 better off. If you’d been running day orders and hadn’t re-entered that morning, your outcome is a blank line. No fill, no basis, no loss.
That blank line is why people underrate expiry. You didn’t lose money on the expired order. You simply weren’t in the position, and “flat” registers emotionally as neutral in a way that a $200 gap on a statement does not. The specific price path here is invented for illustration; the asymmetry it shows is not.
Partial fills split into two fates at the bell
Expiry only touches the unfilled remainder. If 300 of your 1,000 shares filled during the session, you own 300 shares, permanently, and the other 700 cease to exist as an instruction at 4:00 p.m.
The filled portion settles normally. Since May 2024, US equity trades settle one business day after the trade under the SEC’s move to a T+1 settlement cycle, so a fill on Tuesday settles Wednesday regardless of what happened to the rest of the order.
A GTC behaves the same way except the remainder survives. That matters for anyone tracking cost basis carefully: a GTC that fills in three tranches on three different days produces three trade dates, three settlement dates, and possibly three separate commission or fee events depending on your broker’s schedule. One order, three lines on the statement.
Two things people consistently get wrong
“GTC means good until it fills.” It means good until cancelled, and your broker is one of the parties who can do the cancelling. Your March 9 order is dead on May 8 whether or not the stock ever came near $48. Worse, some platforms keep recently cancelled orders visible in the same list as live ones, distinguished only by a status field that is easy to skim past. Check the status column, not whether the row is still there.
“A day order gives me all day.” Only if you enter it at the open. The regular session runs 390 minutes, and your exposure is whatever is left when you click:
| Entry time (ET) | Minutes live | Share of session |
|---|---|---|
| 9:30 a.m. | 390 | 100% |
| 12:00 p.m. | 240 | 61.5% |
| 2:30 p.m. | 90 | 23.1% |
| 3:30 p.m. | 30 | 7.7% |
| 3:55 p.m. | 5 | 1.3% |
An order placed over lunch gets 61.5% of the day. Placed after the last coffee break, 7.7%. If you routinely enter day orders in the afternoon and wonder why nothing fills, that is most of the answer.
Running it backwards: what cap do you need?
Flip the question. You know how long you want an order to sit untouched; the cap you need is trading days ÷ 0.6904, rounded up.
Want a full quarter of coverage — 63 trading days? That needs 92 calendar days. A 90-day cap falls two days short. This is the trap in the whole exercise: “one quarter” sounds like it fits comfortably inside a 90-day GTC, and it fits only if the window is holiday-light and opens on a favourable weekday. Plan on one re-entry per quarter, not zero.
Where this arithmetic breaks
The trading-day math above fails in specific, nameable situations, and each one costs you sessions in the same direction.
Holiday clustering. Thanksgiving, Christmas, and New Year’s land inside about six weeks. A 60-day GTC spanning November into early January loses three holidays plus two shortened sessions, not the two that a flat rate assumes.
Half days. The Friday after Thanksgiving and, in some years, Christmas Eve close at 1:00 p.m. The tables here count those as full sessions. They are 3.5 hours, not 6.5.
Corporate actions. A split, reverse split, or large special dividend can void a resting order or leave it at a nonsensical price. A 2-for-1 split takes a $100 stock to $50, and an unadjusted $100 buy limit is suddenly sitting 100% above the market — it would fill instantly at the open. Most brokers cancel rather than adjust. “Most” is not “all,” and this is a question worth asking your firm directly rather than discovering the answer.
Halts. An order can be live and unable to trade. Fifteen minutes of a halt is fifteen minutes of your 390 gone, and the reopening auction can print through your limit price without filling you the way you expected.
Extended hours. At most brokers, standard day and GTC orders are regular-session-only unless you flag them. A day order does not become an extended-hours order by sitting there.
Modification resetting the clock. Some brokers restart the cap when you change an order. Adjust your price by a penny and a 47-day-old GTC may become a brand-new one, or may not. Both behaviours exist in the market, and the disclosure is where you find out which you have.
What this does not tell you
All of the above is calendar arithmetic. It tells you how many sessions an order can be live. It says nothing whatsoever about whether it fills.
Fill probability depends on where your limit sits relative to the market, how much size is resting ahead of you in the queue at that price, how deep the book is, and how your broker routes the order. None of that appears in these numbers. A GTC sitting 20% below the market for 63 trading days has 63 sessions of opportunity and close to zero chance of anything happening in any of them.
The 252-day trading year is also a convention. Real years run 250 to 253, which moves every count here by about 1% and changes none of the re-entry conclusions.
Nothing above covers GTD, IOC, FOK, AON, or market-on-close instructions, each of which has its own expiry logic. And the tax consequence of a fill landing on December 31 versus January 2 is a real issue that order-type mechanics simply don’t address.
FAQ
Does a GTC order expire if I never cancel it?
At most US brokers, yes — after 60 or 90 calendar days, which is roughly 42 or 63 trading sessions. A few brokers offer longer or effectively open-ended GTC. Futures brokers frequently use different rules from equity brokers on the same platform, so check both if you trade both.
What time does a day order expire?
At the close of the regular session it was entered in, normally 4:00 p.m. Eastern for US equities. Not midnight, and not 24 hours after you placed it. Enter at 3:50 p.m. and you have ten minutes of exposure, or 2.6% of the 390-minute session.
Do day orders carry into pre-market or after-hours?
No, unless you explicitly select an extended-hours order type. A standard day order covers the 390-minute regular session. Full extended hours at many brokers runs 4:00 a.m. to 8:00 p.m., a 960-minute window, so a regular-session day order is live for about 40.6% of the total tradeable day.
How many times would I re-enter a day order to cover a year?
252 — one per trading session. A 90-day GTC does the same job in four entries, a ratio of 63 to 1. A 60-day GTC needs six.
What happens to the part of a day order that already filled?
It stands. Fills are final and settle T+1 under the current US cycle, independent of what happens to the rest of the order. If 300 of 1,000 shares filled, you own 300 shares and the remaining 700 vanish as an instruction at the close.
Will my GTC limit adjust automatically after a stock split?
Sometimes, and it depends entirely on your broker. Treat it as probably not adjusted and possibly cancelled outright. A $60 buy limit on a stock that does a 3-for-1 split should logically become $20; if it stays at $60 against a $20 stock, it is sitting 200% above the market.
Where to look next
Open your own broker’s order-handling disclosure and find three numbers: the GTC cap in calendar days, whether modifying an order resets that clock, and how corporate actions are handled on resting orders. Those three answers rewrite every figure on this page for your account.
After that, the mechanic worth understanding is queue priority — it explains why a limit order can sit at the exact price a stock is trading at and still not fill.
This article is general information, not financial advice. See our disclaimer.
Also worth reading: Ex-Dividend Date: Why the Stock Price Drops by (Roughly) the Dividend
Also worth reading: Stock Splits: Why 4x the Shares Is Worth the Same
Also worth reading: Rule 10b5-1 Trading Plans: What Insiders Are Actually Allowed to Do
Also worth reading: Treasury Auctions: 2 Bid Types and What You Get
Sources
Related articles
- Extended-Hours Trading: Why the Spread Triples After 4pm Work out what a wider after-hours spread costs on your own order size, from 100 to 5,000 shares, with the arithmetic shown line by line.
- T+1 Settlement: What Happens Between Your Trade and Your Money See exactly how a stock trade moves from execution to settled cash in one business day, with a worked timeline and a real slip-up scenario.
- Ex-Dividend Date: Why the Stock Price Drops by (Roughly) the Dividend Work out why a $0.90 dividend knocks about $0.90 off the open, why your account value barely moves, and what taxes do to the gap.