Price impact: how your own order moves the price against you

Price impact is the amount the price moves because of your own order arriving. You see one price on the screen, you send size that is larger than the quantity available at that price, and the rest of your order fills at worse prices behind it.

Nothing went wrong when this happens. It is the normal consequence of a market where every fill needs a counterparty who has already committed to a price and a quantity.

Why there is never just one price

The price you see quoted is not a price the market is offering you in unlimited quantity. It is the best resting order on one side of the book, and it comes attached to a size.

An order book is a queue of commitments. Participants post offers to sell at specific prices and bids to buy at specific prices, each for a specific quantity. The best offer and the best bid sit at the top. Behind them sit worse prices with their own quantities, stacked outward in both directions.

A market order does not ask for a price. It asks for a quantity, and the matching engine fills it against the resting orders in price order: the best one first, then the next best, until the quantity is satisfied. If your order is larger than the top level, it keeps going. This is called walking the book, and your fill price is the weighted average of every level you touched.

So the quoted price is better understood as the price of the smallest possible trade. Any larger trade gets a worse one, and the question is only how much worse.

Walking the book, with made-up numbers

Suppose, and every figure here is invented for the example, that the offers on a book look like this:

You want 45 units, bought at market.

1. The first 20 units fill at 100.00, costing 2,000.
2. The next 15 fill at 100.10, costing 1,501.50.
3. The final 10 fill at 100.40, costing 1,004.

Total cost is 4,505.50 for 45 units, an average of about 100.12. The quoted price when you pressed the button was 100.00, so your price impact was about 0.12 percent of notional, or roughly 5.50 in money.

Now change one thing: you wanted 60 units instead. The first 45 cost the same 4,505.50, and the remaining 15 come from the 101.00 level at 1,515. Total 6,020.50 for 60 units, an average of about 100.34. Order size went up by a third. Impact per unit went up nearly threefold.

That is the shape of the thing. Impact is not proportional to size, because the levels you reach last are both worse in price and, typically, further apart from each other. Liquidity concentrates near the touch and thins as you move out. The first part of your order is cheap. The last part is expensive, and it drags the average for everything.

The part that comes back and the part that does not

After a large order lands, the price does not usually stay where the order left it. Some of the move decays. Market makers who were filled see an opportunity to re-post, and the book refills behind your order over the following seconds or minutes. That portion is temporary impact, and it is the cost of demanding immediacy.

Another portion persists. Your order was information. A participant who buys 60 units when the book only held 45 near the touch has told everyone watching something about their conviction, or their urgency, or their need. Resting liquidity reprices around that, and the new price level holds. That portion is permanent impact, and no amount of patience recovers it.

The practical difference matters. Temporary impact can be reduced by trading more slowly. Permanent impact cannot, because it is a response to the position you took, not the speed at which you took it.

Why the exit costs more than the entry

Entering a position is usually voluntary in both size and timing. You can choose a quieter hour. You can take a smaller position. You can wait for the book to look better. If conditions are poor, you can simply not trade today.

Exits are rarely that comfortable. The reasons you want out, a stop being hit, a margin call, a risk limit, a piece of news, tend to be the same reasons everyone else wants out at that moment. Resting liquidity on the side you need has thinned exactly when you need to consume it, because the participants who would have posted it are reassessing too.

So the arithmetic that looked fine on the way in gets applied to a worse book on the way out. Worse, if you are leveraged, the exit is the leg that cannot be postponed. A position you cannot afford to hold is a position you will sell into whatever is there.

The round trip, not the entry, is the real cost of a size decision. If you size a position by what you can get in at, you have priced half the trade. Our position size calculator and liquidation price calculator are useful here mainly because they force the exit into the same arithmetic as the entry.

The mistake: treating the quote as the price of any size

The common error is not ignorance of impact. It is a scaling error. Someone trades a small size for months, sees fills land essentially at the quote, and concludes that the quote is the price. Then they increase size, and the model silently breaks, because it was never a model of the book. It was a model of being small enough not to notice the book.

The specific failure modes follow from that:

Sizing against account equity instead of against available liquidity. A position is large or small relative to what is resting, not relative to what you can afford. The same quantity is nothing in one market and dominant in another, and account size tells you nothing about which.

Measuring execution quality against the quote. If you compare your average fill to the top-of-book price at the moment you sent the order, every large order looks like a failure. The right comparison is to a reasonable average over the period you were trading.

Trusting a stop to fill at its trigger price. A stop that becomes a market order is an order that will walk whatever book exists at the moment it fires, and the moment it fires is correlated with a thin one. The trigger is where the order is born, not where it fills. If filling at a known price matters more than filling at all, that is a stop-limit decision, with the unfilled-order risk that comes with it.

Assuming a deep book stays deep. Depth is a snapshot of commitments that can be withdrawn. The quantity you measured a minute ago is not a quantity anyone owes you.

The useful habit is small: before sizing, look at how much is actually resting within a price range you would accept, and treat that as your real capacity. If your intended size is a meaningful fraction of it, you already know the fill will be an average rather than a price, and you can decide whether to work the order, shrink it, or accept the cost with your eyes open. More on the general case in what is slippage, and on the order-type trade-off in market order vs limit order.

Education only, not financial advice, and trading carries a risk of loss.