Limit Order: The Fills You Miss Are the Cost
A limit order specifies the worst price you will accept, and will not execute beyond it. You gain control over the price and lose the certainty of being filled, and the orders that never fill are a cost that appears in no statement or trading journal.
How it works
A limit order names a price and says: fill me here or better, or not at all. A buy limit at 100 will fill at 100 or lower and never above it.
That is the whole trade-off against a market order, stated in one line. One of them guarantees the price; the other guarantees the fill. There is no order type that does both, and every execution decision is a position on which you would rather have.
A resting limit order joins a queue at its price, usually filled in the order it arrived. If a hundred contracts are ahead of you and only fifty trade there, you do not fill even though price “reached your level.” That is the most common reason a limit order missed a trade the chart says it should have caught.
The cost nobody records
Every unfilled limit order is a decision with a consequence, and nothing in your account records it. The broker statement shows fills. The journal shows trades taken. The move that ran without you because your bid was a tick too low is absent from both.
Which systematically flatters limit-order execution in review. The filled sample looks better than the full set of intentions, and the difference is exactly the trades that got away.
The fix is to log intended entries as well as fills. A line in a spreadsheet for every order placed, including the ones that never filled, turns an invisible cost into a countable one. It is tedious for a week and settles the question permanently.
There is a second, subtler asymmetry. A resting buy limit fills when sellers come to it — that is, when price is falling toward your level rather than moving away from it. So the fills you get are biased toward the cases where price kept going against you afterwards.
That is called adverse selection and it is a real, structural feature, not bad luck. It is the reason market makers charge a spread for the privilege of resting orders on both sides.
In practice
Most venues run a maker-taker fee structure: a small rebate for adding a resting order and a slightly larger fee for removing one. That difference is small per trade and it compounds over a lot of trades, which is why it appears in every serious cost model and in almost no beginner material.
The round trip on this site’s shared history is 2% of a median bar’s range, and a limit order saves part of that by not paying the spread on entry.
And here is the figure that decides whether it matters. On this history, bar ranges run from 0.17 at the tenth percentile to 1.10 at the ninetieth — a 6.5-fold spread. The same fixed cost is 2% of a median bar and 45% of the smallest bar in the series.
So execution quality matters most in exactly the conditions where it is hardest to achieve: quiet, thin, small-range bars where the spread is a large share of the move and there is nobody to trade with.
Volume is what moves the queue. Heavy participation means orders ahead of you clear quickly; thin participation means your order can sit at the touch and never fill while price trades around it.
A range is where limit orders work best. Price oscillates, comes back repeatedly, and resting orders fill at good prices — which is exactly what market makers rely on.
On a longer timeframe an order rests for hours or days, and an order resting through news is a position you have committed to at a price chosen before the news existed.
A gap past your limit fills you at the opening price if the gap goes through it, which is better than your limit — or skips you entirely if the market opens the other way.
What a limit order is not
It is not a guarantee of execution. It is a guarantee about price only.
It is not visible as a stop. Resting limit orders appear in the order book; stop orders are held at the broker and do not.
It is not free. The unfilled orders cost you the trades, and the filled ones carry adverse selection.
And it is not always the better choice. On a fast move in a liquid instrument, the certainty of a market order is worth more than the fraction of a spread a limit might save.
When it fails
In a fast market it does not fill. Price moves through the level faster than the queue clears, or trades a tick away and leaves. The trades a limit order misses are concentrated in the moves that went furthest.
The second failure is chasing. The order does not fill, price moves, the limit gets moved up, and the process repeats until the entry is worse than a market order would have been at the start.
A third is the review bias. Judging execution on fills alone measures a filtered sample.
A fourth is resting orders through events. An order left in the book across a scheduled release is a commitment made before the information existed.
And a fifth is using one where certainty matters. Exiting a losing position with a limit order means the exit happens only if price comes back to you, which is precisely what it may not do.
The original data
Bar ranges on this site’s shared 576-bar history run from 0.17 at the tenth percentile to 1.10 at the
ninetieth — a ratio of 6.5 — with a median of 0.493 and a smallest bar of 0.022. The round-trip cost of
2% of a median bar is 0.0098 in price units, which is 45% of that smallest bar, and it exceeds 10% of the
bar’s range on 15 of the 576 bars. All of it is in research/series-measurements.json, produced by
site/measure_series.py.
The 6.5-fold spread in bar ranges is the number that should decide your order type, and almost nobody computes it. A fixed execution cost is trivial on a large bar and most of the opportunity on a small one, so the same order type is a good decision in one regime and a bad one in another. Compute the range distribution for your instrument and timeframe, express your round-trip cost as a percentage of the tenth-percentile bar, and you will know whether you can afford to trade quiet conditions at all — which is a more useful answer than any general rule about limit versus market orders.
Related
Order types is the parent page covering the full set. Market order is the other side of the trade-off. And bid-ask spread is the cost a limit order is trying to avoid.
Limit orders fixed my costs and hid my mistakes for about a year. Every fill looked better, and the trades that ran away without me were not in my journal at all - so I was reviewing a filtered sample and concluding my entries had improved.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.