Aggressive trend detector
Flags a bar where a short rise is backed by heavier volume and is speeding up; it almost never fires at defaults.
Over the last `window` one-minute bars it fires only when three tests pass on the same bar: the net change from the first to the last close is above `threshold`, the last five bars' volume totals more than 1.5 times the first five, and the second half of the window gained more, in percent, than the first half. The output is a true or false flag per bar, so it draws markers rather than a line.
How to read it
- Markers sit on the candles where all three tests passed on that bar. No line is drawn.
- A marker says a short rise is building on heavier volume and speeding up into the close. It does not confirm a breakout.
- The rise it describes runs over the last `window` bars ending at the marker.
- Markers are rare at defaults. On the sample SPY month there were none, so a chart with no markers is the normal state.
What it outputs
- Aggressive trend
- True on bars where all three tests pass together: the rise clears the threshold, volume builds and the move accelerates. No marker on the other bars.
When it fires
Fires on the bar closing a window where the net rise clears the threshold, the last five bars' volume exceeds 1.5 times the first five, and the second half gained more than the first half. It marks a rise backed by volume and speeding up at that close.
Settings
- window
- Bars the rise, volume and acceleration span, ending at this bar. Longer windows make the rise threshold easier to clear. Keep it at 10 or more, since below that the first-five and last-five volume sums overlap.
- threshold
- Minimum net rise over the window as a fraction (0.005 is 0.5%). Raising it fires less. Lowering it admits smaller rises, and below zero even shallow falls pass. The volume and acceleration tests ignore it.
Pitfalls
- Zero signals on the sample: SPY one-minute bars for September 2026, about 20,160 bars, never fired at default settings. All three tests must pass at once, and the output does not show which one blocks it.
- The price test compares only the first and last close of the window, so a spike that reverses before the close does not count.
- The try presets were not run on the sample; only the defaults were measured, so check the markers yourself after changing a preset.