Measuring Slippage: Planned Entry Versus Actual Fill
Slippage is the difference between the price you intended to trade at and the price you actually got. Recorded per trade and normalised against your risk, it turns "the market always moves against me" into a number you can analyse and change. And it answers the most important question of all: do you have a strategy problem or an execution problem?
What exactly is slippage?
The gap between your intended price and your realised price, carrying a sign. It can fall against you when you are filled worse than planned, and in your favour when the fill comes in better. That is precisely why it belongs in your journal with a sign rather than as an absolute figure.
Fees are not the same thing. They are known, plannable and printed on your statement. Slippage only comes into existence at the moment of execution and can be estimated beforehand at best. Both belong in separate columns, because the levers available to you differ completely between them.
Nor is the bid ask spread the same thing. The spread is a property of the market and exists before your order, while slippage measures what happened beyond it. In practice the two overlap, which is exactly why choosing your reference price is the decisive preliminary question.
Which reference price is the right one?
The price that applied at the moment of your decision, meaning the price that triggered your signal. Not the price you saw when submitting the order, and certainly not a level that felt fair in hindsight.
The reason is simple: only the decision price makes the figure comparable across different situations. Using the price at submission excludes exactly the part that costs most in hectic phases, namely the time between deciding and clicking.
What matters above all is that you settle on one definition and leave it unchanged for months. A metric whose reference point moves is not a time series. Write the definition down, as you should for your other metrics, covered under the metrics that matter.
How do you normalise slippage?
In R, meaning relative to your initial risk on that trade. An example with freely chosen numbers: if the distance between your entry and your stop is two euros and you were filled twenty cents worse, that is 0.1 R.
This normalisation has three advantages. It makes trades of different sizes comparable, it makes different instruments comparable, and it nets directly against your result, since your result should be measured in R anyway.
Percent of price is the second best choice. It is easier to calculate but says nothing about how heavily the deviation weighs against your risk. Twenty cents is a lot on a tight stop and very little on a wide one, and only R captures that difference.
Where does slippage come from?
From four sources, each of which can and should be observed separately.
- Order type. An order meant to fill immediately takes whatever price it gets. An order with a price limit does not, but may go unfilled instead.
- Liquidity. Thin books and edge hours cost more. That applies to smaller listed names as much as to smaller trading pairs.
- Events. Scheduled moments with expected movement produce jumps where nothing trades between two prices at all.
- Stops. The most expensive case, because a triggered stop order becomes an immediately executing order under many order types, which means it absorbs slippage exactly when the market is moving fast.
A practical recommendation follows from that last point: record slippage separately for entries and exits. The two have different causes and different remedies, and in many journals the larger figure sits on the exit side.
How do you separate a strategy problem from an execution problem?
By calculating your metrics twice, once on intended prices and once on actual ones. The distance between the two figures is your answer.
If the strategy is already weak on intended prices, you have a strategy problem, and working on execution will not help. If it holds up on intended prices and fails on actual ones, the idea is fine and the path from idea to position is not. That is good news, because execution is far quicker to change than a setup.
The best metric for this is one that is dimensionless and reflects both sides of your trades. How to calculate it is covered under calculating your profit factor. Run it once on intended prices and once on actual ones, and your execution quality becomes a single number.
What should you analyse it by?
By the three dimensions where you can actually change something. Everything else is interesting but leads nowhere.
- By instrument or trading pair. The most common finding is that one single instrument carries a disproportionate share of all your slippage.
- By time of day. Especially informative in continuously traded markets, where liquidity shifts markedly across the day.
- By order type. A direct comparison between immediately executing and limited orders, including the trades that never happened because of the limit.
All of that requires clean categories per trade. How to set them up without over engineering them is covered under tagging your trades. And analyse one dimension at a time, or your sample falls apart.
What can you do about high slippage?
The obvious answer, use more limit orders, is only half an answer. A limit trades slippage for unfilled trades, and whether that trade is worth making depends on which trades you lose.
So anyone switching to limits has to count the trades they miss. Record every order that went unfilled because of the limit and what became of the signal. Without that counter calculation you see only the slippage you saved and not what it cost.
Three other levers are usually more effective because they demand no such trade off: avoid edge hours with thin liquidity, reduce position size in instruments with conspicuously high slippage, and treat scheduled events with expected movement as their own category rather than as an ordinary trading day.
Does slippage depend on your position size?
Yes, and this is the most commonly overlooked analysis of the lot. The larger your order relative to what the market is currently offering, the further you have to reach up or down the order book to be filled in full.
In practice: plot your order size against measured slippage and look for a point beyond which the deviation rises disproportionately. That point is your practical capacity limit in that instrument, and it has nothing to do with your risk budget and everything to do with liquidity.
Knowing where it sits answers a question that arrives inevitably as an account grows: whether the strategy you have been running is still the same strategy at larger sizes. In thin instruments it often is not, and this column is where that shows up first.
How many trades do you need?
Enough within each individual value, not in total. Drawing a conclusion about an instrument's execution quality after ten trades in it measures individual events rather than a relationship.
Look at the distribution too, not only the average. Slippage is skewed in many journals: many trades near zero and a few with large deviations. A mean hides exactly those few cases even though they account for most of the cost. How to approach analyses like this in a structured way is covered under reviewing your trades.
In practice: sort by slippage once a quarter and look individually at your ten most expensive fills. You will usually find two or three recurring situations there, and those can be addressed directly.
What has to be in your journal for this to work?
Four fields per trade, three of which an exchange connection supplies on its own: intended price, actual fill price, timestamp and order type. Only the intended price has to come from you, and it has to come before execution.
That is the real effort behind this metric and the reason it is so rarely tracked. An intended price cannot be reconstructed after the fact, only remembered, and memory adjusts itself to the outcome. A trading journal holding the intended price before the order goes out delivers the analysis afterwards with no extra work.
Anyone who keeps that one column consistently has, after a quarter, an answer to a question that otherwise stays open indefinitely: whether the result comes from the idea or from the route to it.