Scalping Strategy Explained: How Scalping EAs Work and Which Timeframe Suits Them

What scalping is, and what an EA scalper actually does

A scalper aims for a few pips, or a few points on gold or an index, per trade, and makes up for the small size with many trades. A manual scalper sits at the screen. An Expert Advisor does the same with rules: when conditions line up, open a position with a tight stop and a nearby target, close it when either is reached or when a short timer ends. The attraction is that each trade risks a small amount and the exposure to overnight events is close to zero.

Most EA scalpers fall into a few families. Breakout scalpers place pending orders around a recent range or a session open and ride the first impulse. Mean-reversion scalpers fade a stretched move back to an average. Momentum scalpers enter in the direction of a short burst confirmed by volume or candle strength. News scalpers trade the seconds around a release, which is the riskiest version because spreads blow out. Each has its own failure mode, and all of them share one dependency: execution.

Why costs decide everything: a worked example

Take EURUSD on a raw-spread account. Suppose the average spread is 0.2 pips and the broker charges a commission that works out to 0.7 pips per round trip. The round-trip cost is about 0.9 pips, paid on every trade whether it wins or loses. Now compare what a typical target looks like on three charts. The numbers are illustrative, but the proportions are what matter.

Read the first row slowly. A one-minute scalper with a 3-pip target and stop has to win 65 percent of its trades just to break even once costs are counted, compared with 50 percent in a world with no costs. Few strategies hold that edge reliably, and a spread that widens from 0.2 to 0.6 pips for a minute, which happens routinely at rollover and around news, wipes it out. On M15 the same cost is a rounding error. This is the entire argument for slower scalping timeframes, and you can test it on your own numbers.

Which timeframe is best for scalping?

There is no universal best, but the trade-offs are clear enough to give a ranking for Expert Advisors, which cannot compensate for poor fills the way a discretionary trader sometimes can.

• M1. The most trades and the least room. It needs the tightest spreads, the fastest execution and the most testing on real ticks. Backtests on a one-minute chart without real ticks are especially unreliable. Best left to experienced traders with a proven broker.
• M5. The workable middle. Targets large enough to absorb costs, enough signals to build a record in weeks, and fewer false starts than M1. Most scalping EAs that run on retail accounts sit here or on M15.
• M15. Strictly a short-term trader's chart rather than a scalper's, with a handful of trades a day. Costs are small relative to targets, and the EA is far less sensitive to the broker. The trade-off is fewer trades and a longer wait to see whether the edge is real.
• M30 and above. Not scalping. They belong to our trend following guide.

The honest summary: for an EA you will run unattended on a retail broker, M5 to M15 is where a scalping edge is most likely to survive real costs. The shorter the chart, the more the result depends on conditions you do not control. A gold scalper compounds the issue, because gold's spread changes sharply across sessions.

Execution: the part the backtest does not show

A backtest fills every order at the price it assumes. A live broker fills at what the market gives, and for a scalper the difference is the strategy. Four things matter.

1. Spread and commission. Raw or ECN accounts with a commission usually beat standard accounts with a wide spread for scalping. See ECN, standard and cent accounts.
2. Slippage. A stop order filled several points worse than its price turns a planned small loss into a larger one. It is far more common in fast markets, which is exactly when scalpers are active.
3. Latency. The time from signal to fill. A VPS close to the broker's server cuts it; see our VPS setup guide.
4. Fill policy. Some brokers re-quote, reject or delay orders that look like scalping. MT5 has different order filling modes that suit different brokers; see order filling modes explained.

One more habit separates scalpers who last from those who do not: they measure their own costs instead of trusting the broker's advertised figures. Record the spread at entry and the difference between the requested and the filled price on every live trade for a month. Divide the total by the number of trades and compare it with the cost your backtest assumed. If the live figure is more than a third higher, the backtest is flattering you, and the right response is to rerun it with the real cost, not to hope the gap closes. A scalping EA that stays profitable at the real, measured cost, and at a cost half a pip worse, has a margin worth trusting.

Broker rules that can end a scalping strategy

Some brokers and prop firms restrict very short trades. Common rules are a minimum holding time (a trade must stay open for, say, 60 seconds), limits on orders per minute or per day, and bans on trading during high-impact news. A strategy that violates them may have its profits cancelled or its account closed, whatever the backtest said. We researched published broker agreements and prop-firm rulebooks in our long guide on scalping brokers and minimum holding times, and the point carries over: read the rules before you pay for an EA, not after.

How to evaluate a scalping EA before you trust it

We apply the same questions to scalpers we review. In our ThunderGold Scalper verdict the main finding was a missing live record and a strong dependence on broker conditions. In our Gold Snap verdict the live signals held up but the trades lasted four minutes on average, which made slippage the key risk. For a wider view of how scalpers compare with grids and trend systems, see our strategy comparison and the short guide on scalping Expert Advisors.