Explosive parabolic detector
A 0 to 100 surge score that sits near 25 on ordinary bars and mostly ignores whether price is rising or falling.
The score adds four capped parts: acceleration of the last four one-minute changes (up to 35), how often a bar's percent move beats the one before it across the window (up to 30), current volume against the earlier bars' average (up to 25), and a two-bar gain that needs about 0.5% to score a single point (up to 10). The acceleration part needs each change to be larger than the one before, in signed dollars, so a fall that is slowing down scores the same as a rally that is speeding up. The percent-move part is close to a coin flip on noisy bars, which is why an ordinary chart reads about 25.
How to read it
- The line sits in its own pane below the price, scaled 0 to 100. It is a score, not a price level.
- On SPY one-minute bars for September 2026 the median bar read 25, the top 5% read 50 or more, and the highest was 90.
- Readings above 50 were rare there, so check the price beside a high reading to see whether the surge is up or down.
- The first bars of a chart read 0 while the window fills (11 bars at the default), which is not a flat reading.
What it outputs
- Parabolic score
- A 0 to 100 score of how much recent one-minute bars look like an accelerating surge. Higher means more of its four parts fired at once.
Settings
- window
- Bars looked back over. Below 5 almost nothing scores, as three parts need 5 bars. From 5 to 8 the percent-move part has few comparisons, so it swings more. Above about 20 that part pins at its 30-point cap, so typical readings sit near 30 and flatten.
Pitfalls
- The score is direction-blind: the acceleration and percent-move parts reward a slowing fall as much as a speeding rally, so check the price beside it before reading it as upside.
- About half of percent moves beat the one before, so that part adds about 25 points to ordinary bars at the default window. The score is not a probability and has no built-in signal threshold.
- The acceleration part is in absolute dollars (scaled by 10,000), so the same shape scores far less on a cheaper stock and cannot be compared across symbols.