Scalping is the day trading strategy with the smallest margin for error: dozens of trades a day, a few cents each, with fees and the spread taxing every single one. It can work, but only if execution and discipline are your actual edge; for everyone else it’s the fastest way to bleed an account ten dollars at a time.
This page covers what scalping is, the logic behind it, the conditions where it falls apart, specific setup criteria, and a complete worked trade with the costs included. The cost math is the part most write-ups skip, and it’s the part that decides whether you have a strategy or a donation schedule.
What scalping is and why it can work
A scalper takes many small trades over a session, holding each for seconds to a few minutes, targeting moves of a few cents to a few dimes per share. No swinging overnight, no waiting for the big trend. You take the small move that’s in front of you, lock in, and look for the next one.
The logic rests on two ideas. First, small moves happen far more often than large ones, so a trader hunting ten-cent moves gets more shots per day than one hunting two-dollar moves. Second, time in a trade is exposure: a position held for ninety seconds can’t get caught in a midday news flush the way an all-day hold can. Each trade risks little. The strategy’s return comes from repetition, base hit after base hit, not from any single winner.
That repetition is also the catch. Every trade pays commissions, regulatory fees, and usually some of the bid-ask spread. A strategy built on $50 winners is far more sensitive to a $6 round-trip cost than one built on $500 winners. Scalping is the strategy where transaction costs stop being a footnote and become a core variable, which is why the worked example below carries a full cost ledger.
When it fails
Scalping breaks in predictable ways. Know them before you risk the first dollar.
The asymmetry problem comes first. When your average winner is $75, a single undisciplined loss of $150 erases two of them, and the temptation to “give it room” is constant. One blown stop can wipe a morning of clean trades. Every scalper’s account dies the same way: not from the small losses the plan allows, but from the big one it didn’t.
Costs compound against you. Twenty round trips a day at $6.50 each is $130 in fees before you’ve made a dime. On thin names the spread is worse than the commission: pay the ask in, hit the bid out, and a five-cent spread costs $0.10 per share before the stock moves at all. The spread will eat you alive on illiquid stocks, which is why the liquidity criteria below aren’t optional.
Slippage scales with urgency. Scalp entries are time-sensitive, so scalpers lean on marketable orders, and marketable orders on a fast tape get filled worse than the screen showed. The same applies on the way out when a trade goes against you and you bail with the crowd.
And the strategy demands your full attention for the whole session. Entries appear and vanish in seconds. If you have a day job, or you check the market on your phone between meetings, scalping is structurally wrong for you; slower setups like momentum trading at least give you minutes instead of seconds to act. This is one of several strategies we break down, and it’s the least forgiving of them.
Skip scalping entirely if any of these describe you: you’re new and haven’t logged months in a simulator, your broker charges per-trade flat fees that dwarf small wins, your platform lags on fills, or you already know you struggle to take a stop. None of those are character flaws. They’re disqualifiers for this specific strategy.
Setup criteria for a momentum scalp
There’s no single scalping setup; traders scalp breakouts, VWAP reclaims, mean reversions, and order-flow imbalances. What follows is a momentum-scalp framework with numbers attached, because vague criteria can’t be tested or journaled.
The stock must be in play. A catalyst (earnings, news, a sector move) plus relative volume above 2, meaning it’s trading at least twice its normal pace. Several million shares already traded by mid-morning is the kind of liquidity that lets 500 shares in and out without moving the price.
The spread must be a penny or two. The cost math below shows why: on a nickel spread, the round trip costs more than most scalp targets pay.
Trade the window that has volume. The first hour, roughly 9:30–10:30 ET, carries the volatility and the participation. After 11 it’s a chop fest on most names, and chop is where scalpers churn their account into commissions. Many of the best scalpers are done by lunch.
The entry needs a defined level: VWAP, high of day, the premarket high, a round number. You’re buying the reclaim or the break of a level everyone can see, with Level 2 and time and sales confirming buyers are actually stepping up, not just printing small lots into a fading bid.
The stop goes where the setup is wrong, not where the loss feels comfortable. If you’re buying a VWAP reclaim, the trade is wrong below VWAP; that’s the stop, and the share size gets calculated from it, never the other way around.
Tooling matters more here than in any other strategy. Real-time data, hotkeys, and fast executions are table stakes, and direct-access routing exists for exactly this job; the trade-offs are covered in direct access versus retail brokers.
A worked example, costs included
The account: $20,000 in a margin account. Risk per trade: 0.25 percent, or $50. Conservative on purpose; at 20+ trades a day, per-trade risk multiplies fast.
The setup, 9:42 a.m.: a stock gapped up on earnings, four million shares traded, RVOL above 3, penny-wide spread. It pulled back from $18.55 to test VWAP at $18.40 and held. Buyers stack the bid at $18.40 on Level 2.
The trade:
| Item | Value |
|---|---|
| Entry | $18.42 on the reclaim |
| Stop | $18.32, below VWAP and the pullback low (10 cents of risk) |
| Share size | $50 risk ÷ $0.10 = 500 shares |
| Target | $18.57, the premarket high (15 cents, 1.5R, $75) |
Now the ledger nobody shows you. At a half-cent per share each way, commissions run $5 round trip on 500 shares; regulatory and ECN fees add roughly $1.50 more. Call it $6.50. If both your entry and exit cross the penny-wide spread, add another $10 (2 cents x 500 shares). So the all-in cost of this trade sits between $6.50 and about $16.50, against a $75 target. That’s 9 to 22 percent of the winner, gone before skill enters the picture.
The breakeven math makes the point sharper. With a $75 winner, a $50 loser, and no costs, you break even at a 40 percent win rate. Add the $6.50 in fees and breakeven climbs to about 45 percent. Run the same trade on a five-cent spread and the spread alone adds $50 per round trip; breakeven jumps past 85 percent, which no one sustains. The spread is the biggest fee on the ticket, and it’s the whole reason scalpers live on penny-wide names.
Scale it to a session: 20 trades at a 50 percent win rate with 1.5R targets grosses $250 (ten winners at $75 minus ten losers at $50). Fees take $130 of it. Net: $120, or 0.6 percent of the account, on a disciplined day where everything went to plan. One revenge trade where you pull the stop and eat a 3R loss hands back $150, and the day’s red. The entire strategy lives or dies on never letting that trade happen. Run your own numbers in the risk-reward calculator before you run them with money.
Risk management for scalpers
Size from the stop, every time. Risk divided by stop distance equals shares; the position size calculator does the arithmetic so you don’t improvise it at 9:42 with the tape moving.
Use live stop orders, not mental stops, until you’ve proven across hundreds of logged trades that you take mental stops without hesitation. Most people don’t, and at scalping speed there’s no time to negotiate with yourself.
Set a daily max loss of 2 to 3R ($100–$150 on this account) and walk away when it’s hit. Scalping generates more trades than any other strategy, which means more chances to overtrade, chase, and average down. A hard daily stop is the only structural defense.
Never average down on a scalp. The trade was wrong at the stop. Adding to it is good money after bad, at speed.
On the rules side: the old pattern day trader rule and its $25,000 minimum were eliminated in 2026, replaced by intraday margin requirements effective June 4, 2026. There’s no trade counting and no $25,000 floor; instead your firm monitors that your equity covers your open positions during the day, $2,000 remains the minimum equity for trading with leverage, and repeatedly failing to cover intraday margin deficits can get an account restricted for up to 90 days, per FINRA’s investor guidance on the new requirements. Firms have until October 2027 to transition, so confirm which regime your broker is on. Full breakdown in our guide to intraday margin requirements. A high trade count no longer triggers a designation, but it doesn’t make the costs or the risk smaller by a cent.
The reality check
Most day traders lose money, and the research is consistent that more frequent trading correlates with worse results, not better; the SEC’s day trading risk publication and FINRA’s guidance on frequent intraday trading both say so plainly. Scalping is the highest-frequency strategy there is, which puts it at the sharp end of that finding. The full numbers on day trading success rates are worth reading before you commit capital to the fastest version of the game.
That $120 disciplined day above annualizes to a decent return on $20,000, and it also assumes a 50 percent win rate with 1.5R winners, flawless stops, and no fat-finger errors, sustained across hundreds of sessions. Build the win-rate evidence in a simulator first. If the sim numbers don’t clear the cost-adjusted breakeven, real money won’t fix that.
Next steps
Scalps start with finding stocks in play before everyone else does; our guide to scan settings that surface these setups covers the gappers, RVOL, and halt alerts that feed a scalper’s watchlist. And because this strategy produces more trades per week than any other, the data piles up fast: log every one in the trading journal template and let your own numbers tell you whether your win rate clears the bar the math sets.