GUIDES

What is scalping in trading, and what does it cost?

Contents
Scalping is usually described by how long you hold. That is the least interesting thing about it. What actually defines a scalp is that the target is small enough for the cost of trading to be the dominant term, which means the arithmetic decides the outcome before the chart does. So this guide is about the denominator. Every number below is measured on US equities rather than assumed, and the question throughout is the same one: what does a scalp have to clear before it is even allowed to be profitable?

1. The only ratio that matters

Take your intended move and divide it by the round trip cost of getting in and out. If that ratio is below one, no amount of accuracy saves the trade. If it is barely above one, you are trading for the broker. The mistake is treating the denominator as a constant. It is not. It changes with the price of the stock, the time of day, the size you send, and which side of the spread you are working. Every section below is one of those terms.

2. What the spread costs, by hour and price

Median spread across 264 recorded sessions and 35,207,932 quotes.
Time$1-5$5-20$20-200
09:30-10:0079bp66bp60bp
10:00-12:0068bp48bp38bp
12:00-14:0069bp44bp33bp
14:00-16:0070bp41bp33bp
Two things in that table matter to a scalper specifically. The opening half hour costs roughly 1.6 to 1.8 times what the afternoon costs in the two upper bands. The window with the most movement is also the window where you hand back the most of it. The bottom band never improves. It starts at 79bp and finishes at 70bp, and nothing in the middle of the day helps.

3. The tick is a floor, and on cheap stocks it is a wall

The reason the cheap band never gets cheaper is the minimum price increment. 32.2% of those 35 million quotes are already one cent wide or less, which is as tight as the market is permitted to go. One cent on a $1.50 stock is 67 basis points. On a $3 stock it is 33. On a $150 stock it is under one. The same tick is a rounding error at the top of the range and a wall at the bottom. The consequence is blunt, and it is the single most useful sentence in this guide. A strategy targeting a 0.5% move is sensible on a $50 stock and structurally impossible on a $1.50 one, because a third of the time that one cent alone is a 0.67% round trip. Parameter tuning does not fix arithmetic about the minimum increment. The full table is in what day trading a US stock really costs.

4. Size is a multiplier, not a constant

The spread is what the top of the book costs. It is not what your order costs, because your order is usually larger than the top of the book.
Market buy$1-5$5-20$20-200
$1,00045bp27bp19bp
$5,00088bp38bp21bp
$25,000132bp45bp28bp
Read the columns. In the $20 to $200 band, going from $1,000 to $25,000 costs an extra 9 basis points, because the book is deep enough that size barely registers. In the $1 to $5 band the same increase triples the cost, from 45bp to 132bp. Double any of these for a round trip if both sides are market orders. A $5,000 market buy in a $5 to $20 stock pays about 38bp one way, roughly 0.75% round trip, before commission and before being right about direction. There is also a fill problem hiding under the cost problem.
Order sizeCould not be filled within ten levels
$1,0000.3%
$5,0002.6%
$25,00028.7%
More than a quarter of the time a $25,000 market order exhausts an entire ten-level ladder. For a strategy whose exit has to happen now, that is not a cost, it is a failure mode.

5. The per-order minimum is a scalper's hidden tax

This is the term most often left out, and it is the one that punishes high trade counts with small orders, which is exactly what scalping is. On 36,947 real round trips priced against published schedules:
Order sizeCommission-freeA fixed plan with a $1.00 minimumRatio
$1,0000.35bp24.04bp68.7x
$2,0000.32bp15.13bp47.3x
$5,0000.29bp10.97bp37.8x
At $1,000 per order that middle column is 24 basis points in commission alone. The stock has to move 0.24% in your favour before you break even on commission, on top of the spread. The arithmetic is simple and worth doing yourself. On a $1,000 order the median trade is 38 shares. At $0.005 per share that is 19 cents, but the minimum is $1.00, so you pay $1.00, and you pay it on both sides, so $2.00. At $5,000 the median is 193 shares and the per-share rate finally exceeds the minimum, which is why the ratio falls to 37.8x. So a per-order minimum is a fixed cost, and fixed costs are brutal on small frequent orders. Take your typical order size, divide by your typical share price, multiply by the per-share rate, and check whether the minimum is larger. If it is, that minimum is your real commission rate. Detail is in day trading commissions measured.

6. The names scalpers watch trade in the smallest clips

There is a structural reason the size term bites. The stocks that appear on a percentage gainers screen are not the stocks carrying the money.
GroupMedian priceDollars per tradeDay range
Top 30 by percent$7.37$88728.3%
Top 30 by dollar volume$319.41$17,4423.2%
Everything that qualified$34.56$3,4582.5%
A median trade on the percentage list is $887. That is the size the market is transacting in on those names, and it sits in the price band where size is punished hardest. The movement that draws you there and the cost that penalises you are the same property of the same stock, measured in top gainers trade in $887 clips. For scale on the other side, during regular hours the quoted price supports about $1,155 to $5,150 depending on band, per order book depth measured.

7. The exit is where the arithmetic stops being an estimate

Everything above assumes the market stands still while you leave. It does not, and the failure is one-sided. On falls of 1.00% or more in a single second, the quoted spread averaged 2.11% against 0.50% on falls up to 0.25%, and the money resting at the best bid fell to $720 from $2,720. The spread widens and the bid thins at the same moment, so the cost of an urgent exit compounds rather than adds. That is the number to put against any scalp that relies on cutting quickly. See stop-loss slippage.

8. Putting the denominator together

For a single scalp, the cost you have to clear is the sum of four things, none of which is optional.
TermWhere it comes from
Spread, both sides41bp to 79bp per side depending on band and hour
Impact beyond the top levelGrows with size, from 19bp to 132bp on a market order
Commission0.35bp, or over 24bp if a per-order minimum binds
Exit conditionsUp to 2.11% spread on a fast move against you
Add them before you look at any chart. If your intended move does not clear that sum with room left over, the strategy is not underperforming, it is arithmetically impossible, and no entry rule changes that. Two things reduce the denominator and both are choices rather than opinions. Larger orders in deeper names cut the impact term and defuse the per-order minimum at the same time. And order type decides whether you pay the spread or offer it, which is the subject of market order vs limit order. The broader version of this arithmetic, applied to day trading generally, is in is day trading profitable. What this guide does not tell you It says nothing about whether scalping works, and nothing about what to trade or when. Every figure here is a cost, and costs only establish the bar. Clearing it is a separate problem and this guide does not claim to solve it. The spread and depth figures are measured on US equities that appeared in a day trading scan, which skews toward smaller and more volatile names. They should not be read as describing large caps. The round trip doubling is an assumption rather than a measurement. The ask side was measured directly; treating the bid side as its mirror is reasonable but unverified here. Queue position is not in any public data, so nothing above tells you whether a resting order would have been reached. That is a real gap and it matters most to exactly the kind of order a scalper leans on.

Related reading

← All guides

Originally published by TraderWe on September 24, 2026. You may quote and link to this page. Republishing the full text without a link back to the original is not permitted.

2 replies

BacktestBetty· Sep 2026 ago
Denominator changes by hour and price band, which means any scalp backtest run on a blended cost assumption is quietly overfit to an average that never existed. I'm re-running mine with cost split by time bucket before I trust a single equity curve.
R
RiskFirstRita· Sep 2026 ago
The tick-as-a-wall point goes straight onto my pre-trade checklist: if the target can't clear a round trip at the minimum increment for that price, the setup gets skipped no matter how clean it looks. Cheap names are where I've been sloppiest.
Sign in to reply →