No edge without the gap
Before trusting one pattern it is worth checking the obvious alternatives on the same data with the same cost model. Two were checked: the opening range breakout and the opening retracement.
Opening range breakout: range from 09:30 to 09:45, 10:00 or 10:30; entry on close beyond the range or on a stop order at its edge; stop at the other side or at the midpoint; target as a multiple of risk or of the range; with and without break-even; with and without time exit; long only, short only, both. One hundred and twenty variants. None of them has a profit factor above 1.1 on MNQ over the year, and most are below 1.
Opening retracement in the ICT style: direction from the 09:30 to 10:00 candle, limit entry at a retracement of the opening move, stop beyond the opening range extreme, target as a multiple of risk or at the prior day's high or low, optional requirement of a sweep first. Seventy-eight variants. Same picture.
The sweep-then-gap pattern passes where these fail because it waits for two things: a liquidity level has to be taken, and the market has to move away from it with a gap in the candles. The first is a location, the second is a confirmation that someone with size was on the right side. A breakout has only the first; a retracement has only a price.
Related negative results from the same study: the pattern does not pay on the S&P 500 future (MES, profit factor 0.89 to 0.99 in every variant, with a larger drawdown than on MNQ), and it does not pay in the London session on NQ (profit factor 0.73 to 1.00 in twelve variants). The Nasdaq in the New York morning is the instrument and the hour.