What Is Scalping?

Michael Whitman · Trading since 2017, full-time since 2020 · Published 2026-09-05 · Updated 2026-10-04 · Reviewed 2026-10-04

▶ Watch: Scalping for Beginners (Full Course)

The four style pages on this site describe one market watched at four speeds, and the comparison between them comes down to arithmetic rather than opinion. This is the fastest view of it.

How it works

Hold for seconds or minutes, take a small move, do it again. A session that a position trader sees as one candle is dozens of candles here.

The typical bar is the raw material. It is the move available inside one bar, before the spread and fees are taken out of it, and at this speed it is the smallest move any trading style works with.

The scalper’s cost problem

Take a hypothetical market where range grows exactly with the square root of time, a round trip costs 0.02, and the typical scalping bar spans 0.50. The cost is 4.0% of that bar.

Here is the same cost against the other three views of the identical hypothetical market:

Same market, same cost, eight times the burden.

Why scalping costs grow faster than its moves

Trade 64 times more often and each move is only 8 times smaller — not 64 times.

That is the table above: 4.00 down to 0.50, a factor of 8, across a 64-fold increase in trade count. The table builds the rule in exactly; real markets follow it only roughly, and the measured SPY figures further down show how roughly.

The reason is that range grows roughly with the square root of time, which is the ordinary behavior of a price series that wanders. Costs do not — a round trip costs the same whether you held it for ten seconds or ten months.

So the number of costs scales linearly and the size of what you are trading scales with a square root. That gap is the entire economics of the choice, and no amount of skill closes it.

Five-minute bars inside a 390-minute session

A regular US stock session is six and a half hours, which is 390 one-minute bars or 78 five-minute bars. A single daily candle on a stock chart is built from all of them.

That is worth holding on to, because it is easy to feel that a fast chart contains more information. It contains the same information at higher resolution, and resolution is not the same thing as edge.

Path against progress on a five-minute chart

Most of what a fast chart shows is back-and-forth. Add up every five-minute move in a session, up or down, and the total is many times the distance from the open to the close. That total is the scalper’s raw material, and it is genuinely there.

The catch is that each piece of it is paid for separately. Capturing more of the path means more round trips, and the measured share of five-minute moves too small to cover one is in the data below.

When scalping works, it is usually the costs

The obvious question this page raises is why anyone does it, and the honest answer is not about skill.

The arithmetic above turns on one input you can change: the cost per round trip. Halve it and 4.0% becomes 2.0% — the share day trading pays in the hypothetical table, with nothing about the market having changed.

That is why serious short-term traders spend so much attention on execution. Direct market access, a venue where you can add liquidity rather than take it, and an instrument whose minimum price increment is large relative to its spread all move that one number, and it is the number the whole style rests on.

The implication for someone starting is uncomfortable and worth saying. If your costs are ordinary retail costs, you are attempting the style at its hardest setting, and the fix is a slower chart rather than more screen time.

A worked example

Work out your cost per round trip first. Spread plus commission plus the regulatory fees on the sale. For 200 shares of a $750 fund with an assumed one-cent spread: $2.00 of spread, plus the SEC’s Section 31 fee of $20.60 per million on the $150,000 sale ($3.09), plus FINRA’s $0.000166 a share ($0.0332). That is $5.1232, or about $0.0256 a share, before any commission.

Divide it by the size of the move you intend to trade. A five-minute move of $0.20 a share leaves $0.1744 after that cost, so the cost takes 12.81% of it. A move of $0.05 leaves $0.0244, and the cost takes 51.23%.

Decide whether that number is acceptable before looking at a single chart. If it is not, the answer is a slower chart, not a better strategy.

Then the ordinary work applies — a read, a level, and an invalidation, exactly as on the chart reading page. Scalping does not have its own analysis; it has the same analysis on a faster chart with a worse cost structure.

The original data

The data. Yahoo Finance five-minute bars for SPY, the S&P 500 fund, over the 60 full sessions from 10 July to 2 October 2026: 4,680 regular-hours bars, 78 a session. Closes ranged from $729.54 to $779.22.

The cost. Fee rates read on 4 October 2026: the SEC’s Section 31 rate of $20.60 per million dollars of sales, in force from 4 April 2026 (SEC fee advisory), and FINRA’s Trading Activity Fee of $0.000166 a share sold (FINRA By-Laws, Schedule A). Both apply to the sale; the test assumes both are passed on to the trader and adds an assumed one-cent spread; no commission is included. Per share that is $0.01 + 0.0000206 × price + $0.000166, which came to $0.0252 at the lowest close and $0.0262 at the highest. Each move was compared with the cost at its own closing price.

How many moves cleared the cost? A move is the absolute change from one window’s close to the next, with each session’s first window measured from the open. Windows start at 9:30 AM; the one-hour rows cover 9:30 AM to 3:30 PM, six a session.

306 of the 4,680 five-minute moves were no bigger than the round trip, and 1,444 were no bigger than five round trips, the size at or below which the cost takes at least a fifth of the move. Going from five minutes to one hour cut that share from 30.85% to 8.89%. The window table is published.

By hour of the session, the share of five-minute moves no bigger than five round trips was 18.89% from 9:30 to 10:30 AM, 25.00% from 10:30, 28.47% from 11:30, 32.08% from 12:30, 41.25% from 1:30 to 2:30 PM, 38.33% from 2:30 and 33.06% in the last half hour. The intraday trading page measures bar ranges by half hour; this is the same afternoon lull seen through close-to-close moves.

Path against progress. For each session, every five-minute move was added up regardless of direction and compared with the net distance from the open to the close. The median session’s travel was $23.06 a share and its median net progress $2.115, and the median of the 60 ratios was 11.5 to 1. The range ran from 2.1 to 1 on 3 August 2026, when SPY rose $8.23 from open to close, to 141.2 to 1 on 10 September, when it closed $0.15 from its open after traveling $21.19.

These are 60 sessions of one very liquid fund in one season, a small sample, and the session table is published so it can be redone on another stretch. The momentum trading vs scalping comparison measures the median move against cost from five minutes out to three months.

Most of the teaching is small, and little of it covers failure. Across our study of 24,971 trading videos, 60% of those with scalping in the title never pass 50,000 views, and the median one runs 12.4 minutes. The corpus carries description text for 126 of them, and 3 of those mention invalidation, failure, being wrong, a loss, a mistake or risk.

When it fails

Most moves are not worth trading

A fixed cost against a variable move means most attempts start behind. A style that pays the same round trip every time and gets a different move every time has to be selective, and the temptation of a fast chart is the opposite of selective.

And the thin moves cluster. The measured afternoon stretch had the largest share of small moves, which is exactly when a bored scalper is most likely to force a trade.

The cost is invisible until you total it

A cost of a few cents per share does not feel like anything. Twenty round trips in a session is a different sentence about the same number, and that is the one worth writing down.

You confused activity with progress

A median session traveled more than eleven times its net move. A fast chart shows you a great deal happening, and almost none of it goes anywhere.

You picked the trade afterwards

On a finished fast chart the good trades stand out. In the middle of one there are dozens of bars, most of them small, and no way to tell which is which — which is the same problem as every other page here, arriving many more times per hour.

Related

Day trading is the same market one step slower, and the direct comparison for whether the cost share is worth it.

Timeframes is why the same market looks different at different resolutions.

And risk management matters more here than anywhere, because a high trade count turns a small sizing error into a fast one.

This page is educational, not financial advice. Test every idea on your own charts before risking money.