Sweep + Volume — What 500 Sessions Say
Every cell below is measured, not modelled: SPY 1-minute bars, 500 regular sessions, entries at the sweep bar's close, stop beyond its wick, path-ordered with the stop checked first, 60-bar horizon. Build a setup and see what it actually returned.
2 · Volume condition on the sweep bar
How to read a confidence interval here. The interval is where the true expectancy plausibly sits. If it straddles zero, the honest reading is "no measurable edge" — not "a small edge". With a per-trade standard deviation near 1R, a sample of 350 can only exclude effects larger than about 0.1R, so a genuinely small edge could hide inside these intervals. What the data does rule out is the large one the setup is usually sold as. Costs are not included: add the spread and every cell moves further left.