Exponential breakout detector
Marks a one-minute close that jumps well beyond the recent typical one-minute move.
Despite its name, this indicator does not look for consolidation or exponential growth. On each bar it takes the last volatility_window closes, measures the one-minute percent changes between them, and finds how large those changes usually are (their standard deviation). If the newest percent change is more than breakout_multiplier times that typical size, it returns the change divided by it, a z-score. Otherwise the bar stays empty.
How to read it
- Markers sit on the candle whose close jumped up sharply versus the close before it.
- The marker value is how many typical moves that jump was: 4 means four times the recent norm.
- Only up-moves mark. A sharp drop never shows here, so read markers alongside the trend, not on their own.
- Markers on consecutive candles read as one sustained run of large up-moves.
What it outputs
- Breakout z-score
- The newest one-minute close-to-close change, in typical-move units. Drawn only where the up-move clears breakout_multiplier, so every value is above that multiplier.
When it fires
On a bar whose close rose by more than breakout_multiplier times the typical one-minute move over the window, that bar included. The marker says one up-jump stood out from recent noise, nothing about what follows. On the SPY sample month it marked about 1.6 percent of bars, roughly 14 a day.
Settings
- volatility_window
- Baseline length in closes (default 30, about half an hour). Longer smooths the threshold and steadies marks, but reacts slowly to new regimes. Shorter tracks recent minutes, so the threshold falls in quiet stretches and ordinary moves start marking.
- breakout_multiplier
- How many typical moves the newest up-move must exceed (default 2.5). Raising it to 3 or more leaves only outsized jumps. Lowering it toward 1 marks ordinary up-moves too. At 0 it marks every up-bar, and below 0 markers stop meaning anything.
Pitfalls
- The name overstates it. Nothing checks for a consolidation before the jump or fits an exponential curve, so an ordinary sharp up-bar in noisy trade marks just as readily.
- The newest move is part of its own baseline. A large jump lifts the typical move it is measured against, so its z-score comes out lower than it would against the prior window alone.
- The sample covers one month of SPY one-minute bars, so the marking rate on other symbols, or in quieter or busier months, is unknown.