Scalping — what is it and does it make sense for retail?

Scalping targets small price movements during short holding periods. A quick exit does not make monetary risk small: volume, costs and execution matter. Before choosing indicators or buying hardware, calculate what actual entry and exit prices leave after fees.

What scalping is — positions held for minutes

Scalping is a short-term style, often with positions held for seconds or minutes. It is not a complete strategy and has no universal trade count, pip target or chart interval. Automated high-frequency trading is a separate implementation category rather than a mandatory neighbor on a retail speed hierarchy.

• Holding time — specify your model’s validity window and exit rule; one to five minutes is not a requirement.
• Profit target — calculate the net monetary result at the assumed exit price; a small pip distance is not proof of a fill.
• Stop loss — entry distance, volume and fees determine planned loss; ordinary stops can slip.
• Number of trades — record opportunities and executions. No daily quota establishes an edge.
• Timeframes — M1 and M5 are possible observation intervals; record when signal data become final.
• Currency pairs — compare actual quotes and execution at your volume. A pair name does not guarantee a spread.

Six assumed gross gains of 10 pips and four losses of 5 pips total 40 pips, or 4 per trade. This hypothetical sample is not an estimated edge. With constant cost c per trade the mean becomes 4 − c pips; unequal wins, losses or costs require their own calculation.

Check the instrument and implementation

Reliable connectivity and measured order latency matter, but no hardware list or universal millisecond threshold guarantees execution quality. Test the platform and provider using the intended order type. On an eight-pip gross target, one pip of adverse slippage consumes one eighth of that target; whether it occurs depends on the actual market and execution conditions.

The brutal arithmetic of costs

First calculate one trade. Assume 10,000 EUR on EUR/USD, a USD account and pip 0.0001 worth USD 1. Initial Bid/Ask is 1.1000 / 1.1001; a long fills at Ask 1.1001. Assume USD 1 total commission and no financing or conversion fees. Exit Bid 1.1006 gives gross USD +5 and net +4; exact-stop Bid 1.0996 gives −5 and −6; a worse fill at 1.0994 gives −7 and −8. Spread is already in the executed Ask/Bid difference: do not subtract it twice. Comparing chart Bid alone overstates this entry’s result by USD 1.

In an assumed 100-trade sample, 60 × USD 4 and 40 × USD −6 total zero despite 60% net winners. Gross profit and total commission are both USD 100. If 20 stops instead lose USD 8 net, the total becomes USD −40 with the same winning proportion. The simplified two-result break-even proportion is 6/(4+6) = 60%, not a forecast.

Keep a separate annual-volume illustration: fifty one-lot trades daily with the following assumed prices and fees.

At the assumed USD9 round-trip cost, fifty one-lot trades per day over 220 days incur USD99,000 in transaction expenses. On USD50,000 of starting capital, that is 198 percent of initial equity: gross trading profit must cover USD99,000 just to offset those expenses. Two hundred trades at the same assumed cost incur USD1,800, fifty-five times less. These figures exclude financing, slippage and changes in trade size or pricing; they are hypothetical costs, not provider quotes.

What a typical scalping setup looks like

The following three-timeframe recipe is an unvalidated example specification. A directional filter does not prove that all scalping must follow a trend. Define the candle patterns, swing rule and expiry precisely before testing.

1. For this candidate rule, use the last closed M15 bar: longs only when Close exceeds its EMA50, shorts only when below. Equality gives no direction.
2. On M5, identify a pullback from previously confirmed anchors. A chosen 38.2–61.8% band is a model parameter, not a price law.
3. On M1, define the pin-bar or engulfing condition and wait for required data. Record the later executable quote; these are alternative rules.
4. If the model chooses a stop five pips beyond an extreme, calculate its full distance from entry. A five-pip buffer is not a five-pip trade risk.
5. Calculate both distances. Entry Ask 1.1001, SL Bid 1.0991 and TP Bid 1.1011 give gross reward:risk 10:10, or 1:1. At 10,000 EUR and USD 1 total commission the net ratio is 9/11, about 0.82, despite a ten-pip target.
6. Specify a target or trailing exit before testing. Standard MT5 trailing is terminal-managed and stops updating when the terminal stops; the last accepted server SL is separate. Neither rule guarantees maximum loss.

The setup's win rate must be measured under explicit entry and exit rules; no reliable 50–60 percent estimate is established here. At a fixed 2:1 gross reward-to-risk ratio, the theoretical break-even win rate is one third before costs. Realised average wins, losses and transaction expenses determine the actual threshold.

Workload and changes to your rules

Monitor whether workload interferes with following the procedure. These are process checks, not findings from personal interviews:

1. Attention. Record monitoring time and interruptions. Set a session limit suited to the implementation; no two- or four-hour optimum is established here.
2. Responses to losses. Define pause and session-ending conditions beforehand. Two or three losses can be chosen limits, not demonstrated universal psychological thresholds.
3. Guard against overconfidence. Keep the predefined position size and stop rules after wins as well as losses. A five-win sequence does not mechanically cause a larger next loss; review any rule changes rather than assuming such a law.
4. Late entry. Recalculate risk and reward at the new quote. Delay can worsen or improve a fill; it does not universally produce a purchase at the top.

Small gross targets make execution costs material. Compare the actual round-trip spread, commission and slippage with the planned and realised outcomes.

What to do before you try scalping

1. Calculate actual round-trip costs. Record spread, commission and observed slippage at the intended size, in both money and pips. Compare net average wins and losses with their frequencies. A cost above twenty percent of the target is a warning to investigate, not a mathematical proof that every possible scalping strategy is unprofitable.
2. Check monitoring and recovery. Record when you can supervise the procedure, what happens after a disconnect and where automation runs. No minimum number of hours guarantees viability.
3. Evaluate losing sequences on demo. Record at least the entry rules, net result, costs and consecutive losses. Compare different periods and execution assumptions. A demo sample does not determine the length of a future live losing streak; plan loss limits without assuming an exact multiplier.
4. Evaluate the executed subset. Four fictional immediate entries yield USD +4,+4,−6,−6, totaling −4. Suppose a one-pip-better entry limit fills only the two losing opportunities, with unchanged exits and commission. Each loss improves to −5, yet their total is −10. Compare all opportunities and fill status: better prices on filled orders need not improve the series. A limit touch does not guarantee full execution.
5. Compare another style before buying hardware. A move to swing trading changes holding time, financing and data requirements. Test its own rules rather than assuming the same setup works better on H4 or needs one daily decision.

Costs, setups and session planning are covered in more depth in the trading strategies section in our library.