TtmSqueeze

John Carter's volatility squeeze: Bollinger Bands sitting inside a Keltner Channel signals a coiled market about to expand. Paired with a detrended-close momentum line that gives the breakout direction.

Quick reference

Field	Value
Family	Bands & Channels
Input type	Candle (uses high, low, close)
Output type	TtmSqueezeOutput { squeeze, momentum }
Output range	squeeze ∈ {0.0, 1.0}; momentum unbounded
Default parameters	period = 20, bb_mult = 2.0, kc_mult = 1.5
Warmup period	period
Interpretation	Squeeze releases (squeeze flips from 1.0 to 0.0) signal a breakout; trade in the direction of sign(momentum).

Formula

squeeze  = 1.0 if BollingerBands(period, bb_mult)
               ⊂ KeltnerChannels-like(SMA(period), ATR(period), kc_mult)
           else 0.0

hl_mid   = (max(high, period) + min(low, period)) / 2
detrend  = close − (hl_mid + SMA(close, period)) / 2
momentum = LinearRegression(detrend, period)        // endpoint

The "Keltner-like" envelope here uses an SMA centerline (not the EMA of typical price that Keltner uses) plus an ATR offset, exactly as Carter's original publication and every chart-vendor implementation define it. The squeeze is on (1.0) when both Bollinger rails sit inside the corresponding Keltner-like rails.

Parameters

Name	Type	Default	Constraint	Source
period	usize	20	>= 2	TtmSqueeze::new (ttm_squeeze.rs:85)
bb_mult	f64	2.0	finite, > 0	ttm_squeeze.rs:91
kc_mult	f64	1.5	finite, > 0	ttm_squeeze.rs:91

period < 2 returns [Error::InvalidPeriod] (the momentum regression needs at least 2 points); a non-finite or non-positive multiplier returns [Error::NonPositiveMultiplier]. TtmSqueeze::classic() returns (20, 2.0, 1.5). Python defaults come from #[pyo3(signature = (period=20, bb_mult=2.0, kc_mult=1.5))]; the Node constructor takes all three arguments explicitly.

Inputs / Outputs

use wickra::{Indicator, TtmSqueeze, Candle, TtmSqueezeOutput};
// TtmSqueeze: Input = Candle, Output = TtmSqueezeOutput
const _: fn(&mut TtmSqueeze, Candle) -> Option<TtmSqueezeOutput> = <TtmSqueeze as Indicator>::update;

• Python streaming. update(candle) returns (squeeze, momentum) or None.
• Python batch. TtmSqueeze.batch(high, low, close) returns an (n, 2)np.ndarray with columns [squeeze, momentum]; warmup rows are NaN.
• Node streaming. update(high, low, close) returns a { squeeze, momentum } object or null.
• Node batch. batch(high, low, close) returns a flat Array<number> of length n * 2, interleaved [squeeze0, momentum0, …].

Warmup

warmup_period() returns period. The three sub-indicators (Bollinger, SMA of close, ATR) are fed unconditionally so they warm up in lock-step, and update emits once all three are ready — candle period (index period − 1). Pinned by warmup_returns_none (period 20: candles 1–19 return None, candle 20 emits).

Edge cases

• Flat market. Both envelopes collapse to a point, so the squeeze is trivially on (squeeze == 1.0) and the detrended regression gives momentum == 0 (test flat_market_has_zero_momentum).
• Binary flag. squeeze is always exactly 0.0 or 1.0 — never a fraction (test squeeze_is_binary).
• Reset. reset() resets all three sub-indicators and clears the rolling high/low/close windows.

Examples

Rust

use wickra::{BatchExt, Candle, Indicator, TtmSqueeze};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let candles: Vec<Candle> = (0..40)
        .map(|i| {
            let m = 100.0 + (f64::from(i) * 0.4).sin() * 2.0;
            Candle::new(m, m + 1.0, m - 1.0, m, 10.0, i64::from(i)).unwrap()
        })
        .collect();
    let mut s = TtmSqueeze::classic(); // (20, 2.0, 1.5)
    for o in s.batch(&candles).into_iter().flatten() {
        println!("squeeze={} momentum={:.4}", o.squeeze, o.momentum);
    }
    Ok(())
}

Python

import numpy as np
import wickra as ta

s = ta.TtmSqueeze(20, 2.0, 1.5)
out = s.batch(high, low, close)  # shape (n, 2): [squeeze, momentum]
squeeze, momentum = out[:, 0], out[:, 1]

Node

const ta = require('wickra');
const s = new ta.TtmSqueeze(20, 2.0, 1.5);
const o = s.update(101, 99, 100); // null during warmup, else { squeeze, momentum }

Interpretation

Carter's setup reads in two steps — when and which way:

1. The coil (squeeze == 1.0). Bollinger inside Keltner means realised volatility has dropped below the ATR baseline. The longer the squeeze persists, the more energy is stored.
2. The release (1.0 → 0.0). When the squeeze turns off, volatility is expanding — enter in the direction of sign(momentum) at the release bar and hold while momentum keeps expanding in your favour.

momentum is a histogram-like reading: positive in a breakout up, negative in a breakout down. Most charting renders it as colored bars; here it is the raw regression endpoint so callers can color/scale it themselves.

Common pitfalls

• Trading the squeeze-on bar. The signal is the release (1.0 → 0.0), not the squeeze itself — entering while still squeezed front-runs a breakout that may not come.
• Expecting Wickra's Keltner numbers in the squeeze test. TTM's internal Keltner-like envelope uses an SMA-of-close centerline, deliberately different from Keltner's EMA-of-typical-price.

References

• John Carter, Mastering the Trade, McGraw-Hill, 2005. The original "TTM Squeeze" was distributed as a proprietary TradeStation indicator through Carter's Trade The Markets / Simpler Trading service.

See also

• BollingerBands — one of the two envelopes.
• Keltner — the other envelope (note Carter's Keltner-like definition uses an SMA centerline, not Keltner's EMA).
• BollingerBandwidth — a continuous measure of "how squeezed" the bands are.