Everybody knows that trading more costs more in fees. Fewer people have worked out what it costs in total, because the fee is the smallest and most visible of three costs that scale with frequency, and the other two do not appear anywhere you would look for them.
The first is the explicit fee, charged per execution, which everybody sees and nearly everybody counts. It is published, comparable, and it is the smallest of the three for most traders. Our own schedule is below, and it is the same for everybody.
The second is the spread. Every round trip involves buying at one price and selling at another, and the difference between those two prices is a cost paid on every single trade regardless of outcome. It appears on no statement and is charged by nobody, which is exactly why it goes uncounted.
The third is market impact: the amount by which your own order moved the price against you. It is small on small orders and it is not zero, and it also scales with the number of trades rather than with the amount of capital deployed. Together these three form a drag proportional to activity rather than to size.
| Market | Maker | Taker | What it means |
|---|---|---|---|
| Buying and selling assets outright | 0.100 % | 0.100 % | 0.0100 % |
| Leveraged positions on contracts | 0.020 % | 0.060 % | 0.0060 % |
The arithmetic is easy and it is rarely done. Take the total cost of one complete round trip, expressed as a percentage of the position: the fee on entry, the fee on exit, the spread crossed, and whatever impact you caused. Multiply by the number of round trips in a period, and that is the return your strategy has to produce before you keep anything.
For an infrequent trader the result is negligible and can honestly be ignored. For somebody trading many times a day, the same arithmetic produces a hurdle that is a large multiple of what a fee schedule alone suggests, and it applies whether the trades worked or not.
This is the number that changes decisions, and almost nobody computes it. It is not an argument against frequent trading, which can absolutely work; it is the threshold that frequent trading has to clear, and knowing it is the difference between choosing a frequency and drifting into one.
The reason frequency costs more than proportionally is that two things happen at once when you trade more. The obvious one is that you pay the round-trip cost more times. The less obvious one is that the average quality of your trades declines, because the additional trades are the ones you were less sure about.
That follows from how trades get selected. A trader taking their best few opportunities is applying a high threshold; the same trader taking many more is applying a lower one, and the marginal trades are marginal by construction. So the additional cost falls on the trades least likely to cover it.
The two effects compound. Doubling frequency roughly doubles the cost while reducing the average expected value per trade, which means the total drag rises faster than the count of trades. That is the mechanism behind the general observation that activity and outcome are frequently inversely related, and it requires no psychology to explain.
The extra trades are the ones you were less sure about. Cost rises with count while quality falls with it, and the two compound.
The spread is the most under-counted of the three because there is no moment at which it is charged. You buy at the offer and sell at the bid, and the difference has already been paid by the time you look at the position. Nothing debits it, so nothing records it.
Its size relative to the fee depends on the asset. On the deepest markets the spread is a fraction of what the fee costs and can reasonably be treated as secondary. On less liquid assets it can exceed the fee several times over, at which point somebody optimising their fee tier while trading a wide-spread asset is optimising the smaller number.
Measuring it is straightforward and worth doing per asset rather than in general. Record the offer and the bid at the moment you trade, take the difference as a percentage, and that is what a round trip costs you before any fee at all. The result frequently reorders which assets are cheap to trade actively.
Market impact is the least relevant of the three for most individuals and it is worth knowing when it stops being irrelevant. An order small relative to the depth at the touch is absorbed without moving anything and its impact is genuinely negligible.
It stops being negligible when the order consumes multiple levels, which depends on the asset, the size and the hour rather than on the size alone. The same order can be invisible during the deep hours and consume several levels in the thin ones, which means impact is partly a scheduling question rather than purely a size question.
The test is direct: place your usual size and watch whether the top of the book is still where it was. If it moves, you are paying impact and it belongs in your round-trip cost. If it does not, you can honestly exclude it and count two costs rather than three.
Every strategy has a break-even win rate determined by its cost per round trip and its ratio of average win to average loss. Costs raise that threshold, and the extent to which they raise it depends entirely on how large the average move captured is relative to the cost of capturing it.
The consequence is uneven and it is the most useful thing in this article. A strategy targeting large moves is barely affected by cost, because the cost is a small fraction of what it is trying to capture. A strategy targeting small moves can be entirely determined by cost, because the cost is a large fraction of the target.
So the same cost structure is trivial for one approach and decisive for another, and the question is never whether costs are high in absolute terms but how they compare to the size of the move being pursued. That ratio, rather than the fee percentage, is the number worth knowing.
First, order type. On most schedules an order that adds liquidity costs materially less than one that takes it, and a strategy that can wait for a fill rather than crossing immediately changes its cost per round trip substantially. That is a structural choice about how the strategy executes, and it is generally the largest available lever.
Second, frequency itself. Fewer, more selective trades reduce the total drag both by paying it less often and by raising the average quality of what remains. This is not a recommendation to trade less, which would be advice about strategy rather than about cost; it is an observation that frequency is a lever with a knowable price.
Third, timing. Executing when the book is deep rather than thin reduces both the spread paid and any impact caused, at no cost other than patience. For anybody with discretion about when to execute, it is the cheapest improvement available, and for anybody without it, it is a reason to model costs by hour rather than by average.
None of this says that trading less produces better results, and it would be dishonest to suggest it does. Plenty of frequent approaches work, and they work precisely because their edge per trade exceeds their cost per trade, which is a comparison rather than a rule about frequency.
What it says is narrower: the cost of frequency is knowable in advance, most people count only a third of it, and the uncounted two thirds are what turns a strategy that looked viable into one that is not. That is an arithmetic problem, not a behavioural one.
The honest position is that costs are a hurdle, not a verdict. A strategy that clears its hurdle is fine at any frequency; a strategy that does not is not made viable by trading it more, which is the direction people tend to move when results disappoint.
One session with your own records. Take a representative sample of your round trips and, for each, record the fee paid on both sides, the spread at the moment of each execution, and whether the top of the book moved when you traded. Add those into a single percentage per round trip.
Multiply by the number of round trips you make in a typical month, and compare that figure to what your account actually produced in the same month. If the two numbers are of similar magnitude, cost is not a detail of your strategy, it is the dominant term in it.
Almost nobody has computed this figure for themselves, and everybody who does finds it larger than they expected. It is one of the very few quantities in trading that is knowable in advance, controllable in three specific ways, and unrelated to predicting anything at all.
Compare a month of round-trip cost against what the account produced that month. If they are of similar magnitude, cost is not a detail of the strategy, it is the dominant term.
The explicit fee on each execution, which everybody counts; the spread between the price you buy at and the price you sell at, which is charged by nobody and therefore recorded by nobody; and market impact, the amount your own order moved the price against you.
Two effects compound. You pay the round-trip cost more times, and the average quality of your trades falls because the additional ones are the ones you were less sure about. Cost rises with the count while expected value per trade falls with it.
Because there is no moment at which it is charged. You buy at the offer and sell at the bid, and the difference has been paid before you look at the position. Nothing debits it, so nothing records it, and it appears on no statement.
It depends entirely on the asset. On the deepest markets it is a fraction of the fee. On less liquid assets it can exceed the fee several times over, at which point optimising a fee tier while trading a wide-spread asset means optimising the smaller number.
Most individuals should not, and the test is direct: place your usual size and see whether the top of the book is still where it was. If it moves you are paying impact and it belongs in your cost. Note that the same order can be invisible in deep hours and costly in thin ones.
Because what matters is the cost relative to the size of the move being pursued. A strategy targeting large moves is barely affected; one targeting small moves can be entirely determined by cost. That ratio, not the fee percentage, is the number worth knowing.
Order type first, since an order that adds liquidity usually costs materially less than one that takes it. Frequency second, which reduces drag both by paying it less often and by raising average trade quality. Timing third: executing when the book is deep costs nothing but patience.
No, and suggesting it would be dishonest. Frequent approaches work when their edge per trade exceeds their cost per trade, which is a comparison rather than a rule. The point is that most people count a third of the cost, and the uncounted two thirds decide whether the comparison holds.