Technical analysis is full of folklore. For our strategy lab we wanted a small, defensible set of price-and-volume structure measures — ones with a real literature behind them — to time entries inside an already-chosen universe. Here's what we built and why.
The measures
- Linear-regression channel. We fit an ordinary least-squares line to recent closes; the slope is a continuous, sign-aware trend measure (no moving-average crossover lag), and a band of ±2 residual standard deviations forms a dynamic channel. This is the classic regression / Raff channel.
- Volume profile & Point of Control. We bin the price range and sum volume into each bin. The highest-volume bin is the Point of Control; the nearest high-volume bins below and above the current price are our support and resistance. The idea — prices the market traded heavily get "accepted" and tend to be revisited and defended — comes from J. Peter Steidlmayer's Market Profile.
- ATR (Average True Range). Wilder's volatility measure turns "near support" and "stop-loss" into volatility-relative tests instead of arbitrary percentages, so the same rules behave sensibly across calm and jumpy stocks.
- We keep the familiar Donchian channel (Turtle Traders) and Bollinger Bands too — they power our breakout and band-reversion strategies.
The new strategy: trend-pullback
Pure breakout systems buy strength. Our new channel-pullback strategy does the opposite within a trend: it waits for a confirmed uptrend (positive regression slope) and then buys pullbacks toward the lower channel or a high-volume support node — better entry price, better risk:reward. It exits at the upper channel / resistance, when the trend rolls over, or on an ATR stop below support.
Why we don't trust it blindly
Most single-indicator TA rules do not survive transaction costs and multiple-testing correction. So every strategy — this one included — must clear a walk-forward + Deflated Sharpe gate (López de Prado) before it goes live, and the on-chart overlays on each company page are labelled educational, not advice.
Further reading
Wilder, New Concepts in Technical Trading Systems (1978) · Bollinger, Bollinger on Bollinger Bands (2001) · Steidlmayer & Koy, Markets and Market Logic (1986) · Faith, Way of the Turtle (2007) · López de Prado, Advances in Financial Machine Learning (2018).