Scalping strategy for US stocks with a spread filter

A scalping strategy for US stocks where the spread filter, not the entry, is the strategy. Momentum filter, spread guard, time stop. Copy it, then tighten each condition against your own recordings.

1. The rule

# BUY if ChangePct > 5 and Strength > 120 and (Ask1 - Bid1) / Price < 0.003: Buy()# SELL if ProfitPct >= 0.8 or ProfitPct <= -0.5 or HoldTime >= 300: Sell()

2. Why the spread line is the whole strategy

ChangePct > 5 is a selection filter rather than a timing one — it picks names that are already moving today, which is where scalping has something to work with. Strength > 120 asks that buyers have been paying up. But the line that decides whether any of this can work is the third one. A scalp targeting 0.8% pays the spread twice. At a 0.3% spread the round trip has eaten three quarters of the target before the trade does anything at all.

3. What the round trip asks of it

Read that second bar carefully, because it is the most important number on this page. Against our measured cost for a $5,000 order in a $5–20 stock — about 0.75% round trip — a 0.8% target needs a 96% win rate to break even. That is not a strategy, it is a rounding error. This is not a flaw in the figure. It is the reason the spread condition exists. That measured 0.75% is a *median across the population*, and the spread filter's job is to keep you in the part of the distribution well below it. If your fills are averaging anywhere near the median, scalping a 0.8% target does not work, and no entry condition will rescue it. So the honest reading of this strategy is: it is a bet that you can select names where your real round trip is a quarter of the median. Measure your own fills before you believe you can — see what day trading a US stock really costs for how, and the quoted price is only good for about $1,200 to $5,200 for why your size matters as much as the quote.

4. When it fails

It fails at the open, when spreads on the exact names that pass ChangePct > 5 are at their widest. The condition rejects most of those entries, which is correct behaviour, but it leaves the strategy with very few trades in the first half hour — the window where the moves are.

5. What to change first

The target and the spread cap move together. If you widen the cap to 0.005 you must raise ProfitPct >= 0.8 to at least 1.5, or you are keeping the same target while paying nearly twice as much to enter. Change them as a pair and the comparison stays honest. Then compare the whole thing against the baseline your trading strategy has to beat. Scalping has to clear a much higher bar than a slower strategy, and it is worth knowing by how much before you commit to it. Test on your own data: how to backtest a day trading strategy on US stocks.