Home › Learn › Lesson 6

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

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?

The one trade-off everything is made of

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.

Market orders: you buy speed and pay in price

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.

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.

Limit orders: you buy price and pay in uncertainty

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. If your venue charges 0.05% taker and 0.02% maker, 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 skew toward winners rather than being average — and that assumption is not pessimism, it is the definition of how limit orders fail. They fail when price leaves without you.

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.

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.

Stop orders are not orders — they are triggers

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. If leverage is in play, this 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.

The setting almost nobody opens: what price triggers your stop

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 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.

The decision table

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. Build the habit before you build the account, which is the argument running through knowledge capital before trading capital.

Common mistakes 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 didn't my stop 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.

Last price or mark price for my stop trigger? 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.

← Curriculum · Previous: Lesson 5 — Choosing a trustworthy exchange