Home › Learn › Lesson 7

Market, limit and stop orders — choosing the one that fits the trade

By TradingPrimer · Updated 18 Sep 2026 · Published 28 Aug 2026

Every order type on every exchange is built from the same single trade-off: you can have certainty about the price you get, or certainty that you trade at all, but not both in the same order. A market order buys execution and pays for it in price. A limit order buys price and pays for it in execution. A stop is not a third kind of order at all — it is a trigger that fires one of the first two on your behalf. Once you see it that way, choosing correctly stops being memorisation and becomes one question: which certainty does this particular trade actually need?

What is the one trade-off every order type is built from?

An exchange matches buyers to sellers through an order book: a list of everyone's resting offers to buy below the current price and sell above it. Every order you can place is simply an answer to one question — do you come to the book's prices, or do you make the book come to yours?

Come to the book, and you trade immediately at whatever prices are sitting there. That is a market order. You are a taker: you remove liquidity that someone else provided. Post your own price and wait, and you trade only if someone chooses to come to you. That is a limit order. You are a maker: you add liquidity to the book.

Everything else on the order ticket — stop-market, stop-limit, trailing stop, post-only, reduce-only, take-profit — is one of those two with a condition bolted on. There is no third mechanism to learn. If the order book itself is still hazy, how the crypto market actually works takes it apart level by level.

What does a market order actually cost you?

A market order says "fill me now, whatever it costs." The exchange walks up the book, consuming resting offers from the best price outward until your quantity is complete. If your size is small relative to what's resting at the top of the book, that walk is one step and the cost is invisible. If your size is large relative to the book, the walk is expensive — and the difference between the price you saw and the average price you got is slippage. How large that walk gets depends entirely on the depth resting beneath you — Lesson 9 on liquidity and spread shows how to measure it in sixty seconds and size the position to fit.

Here is the mechanic with real arithmetic. Suppose the sell side of a book looks like this, and you send a market buy for 3,000 units:

Your average fill is $60,584 ÷ 3,000 = $20.195, not the $20.05 you saw quoted. Had the whole order filled at the best ask it would have cost $60,150, so the walk cost you $434 — about 0.72% of the trade, paid instantly and invisibly, before the position has done anything at all. On a strategy targeting 1.5% per trade, you have just handed over close to half the target on entry, and you will pay something similar on the way out.

Two things follow that are worth internalising now. First, slippage is a function of your size against this book at this moment, not a fixed property of the asset — the same order costs almost nothing at 3pm on a major pair and a great deal at 4am on a thin one. Second, the quoted price on your screen is the price for the next small trade, not for yours. Beginners treat the ticker as a price they are entitled to. It is a headline about the top of the book.

Should you always use limit orders to save on fees?

No. The maker rebate is real but small — $6 on a $10,000 round trip in the example below — and a limit order fails to fill precisely when price runs away from you, which is when you were right. Use limit orders when patience is the strategy, not to save the fee.

A limit order posts your own price into the book and waits. You will never pay worse than your limit — and you may never trade at all. That is the whole bargain, and the second half of it is where the money is actually lost.

The standard advice is that limit orders are strictly better because they earn the maker fee instead of paying the taker fee. The fee part is true. Take illustrative rates of 0.05% taker and 0.02% maker — not any venue's schedule, check your own tier — and a $10,000 round trip costs $10 in fees as a pure taker and $4 as a pure maker. Six dollars a trade sounds trivial; across eight round trips a month it is $576 a year, which on a $5,000 account is 11.5% of capital. Fees deserve to be taken seriously.

But the conclusion — always use limit orders — does not follow, and here is the reason most articles skip.

A resting limit order is a free option you have written to the market

When your buy order sits in the book at a fixed price, you have given every other participant the right, but not the obligation, to sell to you at that price for as long as it rests there. They will exercise that right precisely when it suits them, which is when they know something you don't or when the market is already moving down through your level. When it doesn't suit them, they simply don't trade and your order expires unfilled.

The consequence is that your unfilled orders are not a random sample. Your limit order fails to fill exactly when the market moves away from you — which is to say, when you were right. The fills you do get are weighted toward the trades where price came back to you, and price often came back because the move wasn't there. Professionals call this adverse selection. Retail traders experience it as "I keep getting filled on the bad ones and missing the good ones," and assume it's bad luck.

Put numbers on it. Take twenty setups over a month, a trader whose record is 40% winners at +$100 and 60% losers at −$50, trading $10,000 notional at 0.05% taker / 0.02% maker:

The maker route saved $136 in fees and forfeited $400 of winners, turning breakeven into a $264 loss. The assumption doing the work is that the four unfilled setups are winners rather than average trades. That deserves a test rather than a shrug, so we ran one.

Is "the misses are winners" an assumption, or arithmetic?

We simulated 20,000 setups per row with these conditions, so anyone can rerun it: price moves in Gaussian steps of 2 points per tick; a setup's edge is a drift of μ points per tick that lasts 50 ticks after the signal and then stops (a signal goes stale); the target is +100 and the stop −50, both measured from the signal price, not from the fill; the limit order sits 25 points below the signal and fills only if price touches it before reaching the target; fees are $10 per market trade and $4 per filled limit trade, as above. One point is one dollar on the $10,000 position. Seed 7; three other seeds move the per-20 figures by about ±$20.

Two things in that table are worth more than the table above it. First, the 100% column is not a pessimistic assumption at all — it is geometry. If your limit sits between the signal price and your stop, a setup that never reached your limit could never have reached your stop either, so every unfilled limit in that layout was a winner you did not get. There is no such thing as a resting limit that "kept you out of a loser" unless it sits below your stop, which would be absurd. Second, look at how little the limit column moves — −$65 to −$78 whatever the edge — while the market column swings from −$214 to +$178. The limit route enters $25 better on every fill and pays $6 less, and it still cannot keep the edge, because the edge lives in the first 50 ticks and is spent waiting for the dip. That is adverse selection with numbers on it: the fills you get are the setups where the drift had already faded.

None of which means never use limit orders. It means the choice depends on what your entry is worth. Two practical rules fall out of the arithmetic:

• If missing the trade is worse than paying the spread, take. Momentum entries, breakouts, anything where you are trying to get in because price is moving — use a market order and accept the cost. You are buying the trade, not the tick.
• If your edge comes from patience, make. Mean-reversion entries, scaling into a level you expect price to revisit, exits at a planned target — the whole point is that price comes to you. A limit order here isn't just cheaper, it is the strategy.

There is a second reason to be careful with resting limits, and it comes from behaviour rather than statistics. In the free course, Part 8 lists four reasons a polarity trade — old resistance turning into new support — fails, and the first is entering in a rush, with limit orders set in advance. A limit parked at a level is a bet that the level holds, placed before the market has shown you anything: no reversal structure on the lower timeframe, no confirmation candle, nothing. The fee saving is small; what you have really given up is the right to look first.

And one hybrid worth knowing: a marketable limit — a limit order placed slightly through the current price (a buy limit just above the best ask). It fills immediately like a market order but caps your worst-case fill, so a sudden thin book can't walk you up thirty levels. It is the default for anyone entering size in an illiquid pair.

When is "take the market order" the wrong advice?

When you have no edge at the moment of the signal, when the setup is the revisit, or when the book is so thin that the walk costs more than the edge. In each case the limit order wins, and the first case is the one most beginners are actually in.

Go back to the first row of the simulation. With no edge, the market route loses $214 per twenty setups — fees, plus nothing to show for paying them — while the limit route loses $75: it pays $6 less per trade and enters $25 better on the 80% of setups that fill, and there is nothing for adverse selection to take from you, because there is nothing there. If your journal cannot show positive expectancy from the signal bar, use limit orders. The adverse-selection argument only bites when you have something to be selected out of; until then, fees are the only variable you control, and the maker side of the book is the right place to be.

The advice flips a second time when the price coming to you is the idea, not a delay of it: a mean-reversion entry at a level you expect to be tested, a scale-in below a breakout, an exit at a planned target. In the simulation's terms, the drift there begins when the level is touched, not when you first noticed it, so waiting costs nothing. And it flips a third time on a book where the market order's walk — the 0.72% above — is larger than the edge you are chasing. That is the marketable limit's job, and the reason Lesson 9 teaches you to read the depth chart before you size.

Why did your stop-loss not fill at the price you set?

Because a stop is not a resting order — it is a trigger. When your trigger price prints, the exchange submits a market or limit order on your behalf, and that order meets whatever the book looks like at that moment: a stop-market fills at the best prices available, which in a fast move can be well past your level, and a stop-limit may not fill at all.

This is the single most consequential misunderstanding on the list, because it is what people get wrong at the worst possible moment.

A stop-loss does not rest in the order book. Nobody can see it. It sits with the exchange as a conditional instruction: when the trigger price prints, submit an order on my behalf. Which kind of order it submits is your choice, and that choice determines exactly how it can fail.

Stop-market: when triggered, sends a market order. You are guaranteed to exit. You are not guaranteed a price — and stops trigger during exactly the fast, one-sided moves when the book is thinnest and slippage is largest. Your stop can fill well below the level you set.

Stop-limit: when triggered, sends a limit order at a price you specify. You control the worst fill. But if the market gaps straight through your limit price, there is nothing to fill against — the order rests, unfilled, and you are still in the position while it keeps going.

Work the second case through. You are long 0.05 BTC entered at $60,000, with a stop trigger at $58,000 and a stop-limit price of $57,900 — a planned loss of $105, which is 2.1% of a $5,000 account. News hits. The tape prints $58,050, then the next print is $56,800. Your stop triggers correctly and a sell limit at $57,900 goes into the book — where the best bid is now $56,800. Nobody buys at $57,900. Your order sits there, and you are still long. You notice and exit manually at $56,500.

The planned risk became 1.67× larger than intended, and the mechanism that was supposed to cap it is the mechanism that failed. This is why the honest default for a protective stop on a leveraged position is stop-market: bad slippage on an exit you definitely got is a smaller problem than a clean price on an exit you never got. Reserve stop-limit for situations where a specific bad fill is genuinely unacceptable and you are watching the screen. The default flips in exactly one case: a book so thin that a stop-market would fill several percent below the trigger — a small altcoin at 4am, or any pair whose depth chart shows almost nothing beneath you. There, a stop-limit with a wide gap between trigger and limit (1% or more, against the 0.17% gap between $58,000 and $57,900 in the example) caps a fill that a market order would have made catastrophic — and you accept the obligation to watch the screen that comes with it. If leverage is in play, your stop may never get the chance to fire at all — Lesson 8 on leverage and margin works out where that crossover sits. It is the same failure geometry described in the anatomy of a liquidation cascade — and the reason your liquidation price should never be close to your stop.

Last price or mark price — which one should trigger your stop?

Mark price is the safer default for a protective stop on a leveraged position; last price is defensible only if you want your stop to respect exactly what that one venue printed. Whichever you choose, choose it on purpose — the dropdown is already set to something, and most beginners have never opened it.

Here is the detail that separates people who have read about stops from people who have used them on a derivatives venue.

Your stop needs a price to watch, and on most futures and perpetual venues you can choose which one:

• Last price — the most recent trade that occurred on that specific exchange. Literal, but it is one venue's tape, and a thin book can be pushed a long way for a couple of seconds by one aggressive order.
• Mark price — a smoothed reference derived from an index of several venues, which is also what the exchange uses to calculate your unrealised PnL and your liquidation. Far harder to move with a single order on one book.
• Index price — the underlying spot index itself, without the derivative's basis.

The same stop, at the same level, behaves differently depending on which one is selected. A brief wick that prints on one venue's last price but never registers in the index will take you out under a last-price trigger and leave you untouched under a mark-price trigger. Traders who get "wicked out and then watch it go straight to target" are usually describing a last-price trigger on a thin book — and the candle it leaves is read properly in Lesson 10 on candlesticks — and are usually blaming a conspiracy rather than a dropdown they never opened.

The one thing to do after reading this lesson: open your exchange's order settings, find the stop trigger source, and confirm which one is the default on your account. Set it deliberately. Mark price is the safer general default for protective stops on leveraged positions; last price is defensible if you genuinely want your stop to respect exactly what that venue printed. What is not defensible is not knowing.

Which order type should you use for which trade?

Take when missing the trade is the real cost, make when your edge is patience, protect with a stop-market, and flag every exit reduce-only. The table is the same rule set laid out per order type.

Print this, or rebuild it in your own words — the second is better.

The last row is a small habit with a large payoff. Flagging every exit reduce-only makes it structurally impossible for a mistyped exit to open a new position in the opposite direction — a mistake that turns a closed trade into an accidental short at the worst moment. It costs one checkbox. On the Binance Futures app, as captured in the course screenshots, that checkbox sits beside the quantity box labelled Reduce only, next to Post only; the open-position row shows Entry price, Mark price and Liq. price side by side with a Stop TP/SL control. Exchanges rename controls, so treat those labels as what to look for, not a promise. Build the habit before you build the account, which is the argument running through knowledge capital before trading capital.

What do people get wrong at this stage?

Treating the ticker as your price. The quote is for the next small trade at the top of the book, not for your size. Using market orders on illiquid pairs. This is where the 0.72% walk in the example above comes from; use a marketable limit instead. Using stop-limit for protective stops. The one time it fails is the one time you needed it. Setting stops at obvious round numbers. Everyone's stop sits at $60,000, which makes it a pool of resting sell pressure that price is drawn toward; put yours where your idea is actually invalidated, not where the number is tidy. Never checking the trigger source. A dropdown you have not opened is deciding whether wicks take you out. Chasing the maker rebate on time-sensitive entries. You are saving $6 and risking the trade. Forgetting reduce-only on exits. One typo and your closed long is an open short.

FAQ

What's the difference between a market order and a limit order? A market order guarantees you trade but not at what price; a limit order guarantees the price but not that you trade. Execution certainty and price certainty are the two ends of one lever — you pick a position on it, you don't get both.

Should I always use limit orders to save on fees? No. The saving is real but small — $6 per $10,000 round trip at 0.05% taker versus 0.02% maker. Limit orders fail to fill precisely when price runs away from you, so the trades you miss skew toward winners. In the worked example above, $136 of fee savings cost $400 of forfeited winners.

Why did my stop-loss not fill at the price I set? Because a stop isn't a resting order. It is a trigger that submits a market or limit order once your price prints. A stop-market takes whatever is there in a fast move; a stop-limit doesn't fill at all if price gaps past your limit.

What is a marketable limit order? A limit order placed slightly through the current price — a buy limit just above the best ask. It fills immediately like a market order but caps your worst fill, so a thin book cannot walk you up thirty levels. It is the default for entering size in an illiquid pair.

Should my stop trigger on last price or mark price? Mark price is the safer default for protective stops on leveraged positions, because it is drawn from a multi-venue index and is much harder to move with one order. Last price is defensible if you want your stop to respect exactly what that venue printed. The real answer is to open the setting and choose on purpose.

Where do I actually click all this? The theory above is venue-agnostic on purpose, because interfaces change. For the button-by-button version on a live exchange — product tab, pair, side, unit labels, and reconciling the fill afterwards — follow how to place your first spot order on Binance.