2 / 6 · 9 min
The arithmetic of style: is the travel enough to cover costs
One calculation answers whether your chosen horizon is possible at all: how far the instrument travels in that time versus what it costs to get in and out. If the first is smaller than the second, skill is not the issue.
Where profit comes from
Every trade starts in the red: you gave up the spread on entry and will pay a fee twice. Profit appears only once price has travelled further than that round trip cost. So a style has a simple feasibility check: take the instrument's typical travel over your holding horizon and divide it by the cost of a round trip. The result shows how many times over there is something to take.
Average travel comes from NATR — how far the instrument moves over a window, in percent of price. Round-trip cost is the spread plus two fees; on major venues at market that is about 0.1%. Headroom below one means the average trade loses money BEFORE any markup: you pay more than the instrument manages to travel.
Bitcoin and an average coin: two different worlds
Measured 09.08.2026. BTC's average travel: 0.04% over 5 minutes, 0.11% over an hour, 0.53% over 4 hours, 1.96% over a day. Median across 166 liquid coins: 0.54% over 5 minutes, 1.84% over an hour, 3.18% over 4 hours, 7.46% over a day. At a round-trip cost of 0.1%, bitcoin's headroom on a five-minute horizon is 0.4 — the average travel is THREE TIMES SMALLER than costs. On an average coin over the same window headroom is 5.4. The same entry scheme on these two instruments is work in one case and guaranteed loss in the other.
Headroom by horizon, at 0.1% round-trip cost
| Horizon | BTC | Average coin | Suits |
|---|---|---|---|
| 5 minutes | 0.4× | 5.4× | volatile coins only |
| 1 hour | 1.1× | 18× | day trading almost anywhere |
| 4 hours | 5.3× | 32× | day trading and swing |
| a day | 20× | 75× | swing and position |
Read the table not as bitcoin is bad but as every instrument has its own minimum horizon. On BTC meaningful work starts around four hours; on a volatile coin even five minutes is possible. Headroom of 5 is not a promise of profit, only the condition under which profit is possible at all: everything you get wrong still comes out of that same headroom.
Counting only the fee as cost
The fee shows up in the statement, so people remember it. The spread shows up nowhere: you simply get filled slightly worse than the price you saw. On a liquid coin it is negligible; on a coin turning over a hundred thousand dollars it is half a percent, ten times the fee. On top comes slippage: a thousand-dollar order on a thin market moves price by itself. When computing round-trip cost, take all three parts, or the calculation will show headroom where there is none.
The second cost — paid for time
Short horizons are dominated by the number of trades, long ones by funding. The market median rate is 0.0051% per eight hours, that is 5.6% a year: on a week-long trade it comes to 0.1%, exactly one more round trip. But the rate is never average where things are interesting: on a coin in the middle of a move it reaches tenths of a percent per settlement, and then a week of holding costs several percent. Before opening a multi-day position, look at the rate — unlike direction, it is known in advance.
Settlement happens three times a day, so the daily cost is three times the rate. At 0.01% that is 0.03% a day and 0.2% a week; at 0.1% it is already 0.3% a day and 2.1% a week. The second number is comparable to the target of the trade, and that changes the decision: the same idea on the same chart can work on spot and lose on a perpetual.
How often you must be right just to break even
Headroom says whether there is something to take, not how often you must guess right. If you target one unit of travel and stop out at one unit, then with costs at a tenth of that travel you need slightly more than half your trades to win — around 52–53% to break even. Every extra round trip of cost adds roughly a percentage point to that bar. For a scalper with forty trades a day the bar climbs to 60% and above, and that is why the same markup method is profitable on hourly charts and lossy on minutes.
The NATR column in the screener shows average travel over different windows, from a minute to a day. Sort the table by the window that matches how long you usually hold, then look at the bottom of the list: those are the instruments where your style will not work, however good the chart looks.
About volatility and NATRCompute headroom for your own instruments
Take three instruments you trade and for each write down the NATR of the window equal to your usual holding time. Then estimate the round-trip cost: your venue's fee times two, plus the spread visible in the book right now. Divide the first by the second. Cross out instruments with headroom below three — your horizon is too short for them. If all three get crossed out, what needs changing is not the instruments but the horizon.
A coin's average 5-minute travel is 0.2% and the round trip costs 0.1%. Worth scalping?
Headroom is two, and that is thin. Half of the typical travel goes to costs, so you have to guess direction far more often than half the time — and on five-minute noise that is hard. A practical guide: below headroom of three, a style survives only on very high entry accuracy, and the first losing streak eats the result. It is wiser to take either a more volatile instrument or a longer horizon — the second is almost always cheaper.
Why is the same strategy profitable for one person and lossy for another?
Usually not because of execution but because of conditions the strategy never wrote down. Different fees — 0.02% as a maker for one, 0.05% at market for the other. Different instruments — on BTC the five-minute travel is three times smaller than costs, on a volatile coin five times larger. Different account size — on a small one slippage is invisible, on a large one the order moves price itself. A strategy describes entry and exit, while profit is also decided by those three numbers; copying somebody's system without recomputing it for your conditions is pointless.