Just the highest high and lowest low over N periods — no averaging, no smoothing. The engine behind the Turtle Traders' legendary breakout system, and still the cleanest definition of "a new high."
Richard Donchian, later called the "father of trend following," built his channel from nothing but the highest high and lowest low of the past N days.
An early version bought on a new four-week high and sold on a new four-week low — about as simple as trend-following gets.
Richard Dennis and William Eckhardt built their trainee-trader system around Donchian-style 20-day and 55-day breakouts, proving the rule could be taught to complete beginners.
Still the reference definition of "a new high" or "a new low" for countless breakout systems, prized for its total simplicity.
The upper band is the highest high of the last N bars; the lower band is the lowest low — the midline just averages the two.
Because the bands are raw highs and lows, "price closes above the channel" has one single, unambiguous meaning — a genuine N-period high.
In sideways markets, price can tag both the upper and lower band repeatedly without any trend ever developing, since a short lookback makes "new highs/lows" cheap and frequent.
A well-documented trade entered when silver reached a new 40-day high, exited on a close below the 20-day low — a highly profitable, rules-only breakout trade.
Through that long, grinding range, a short-lookback channel tagged both bands repeatedly with no lasting trend ever developing — a genuine whipsaw stretch.
A trader uses a 10-day Donchian channel and treats every single new 10-day high as a major trend signal worth a large position. Sound?
Silver breaks a genuine, documented 40-day high. A trader takes the breakout seriously, sizing normally and trailing an exit on a shorter, opposite-direction channel. Sound?
A trader assumes Donchian Channels smooth price the same way Keltner or Bollinger Bands do. Accurate?
Price and the channel, watched tick by tick on the left — and the mark it leaves in the ledger on the right. A genuine long-lookback breakout, a choppy short-lookback whipsaw — and the Turtle-style trailing exit.
Read the base leading into the break, then call it: a rare, major breakout, or minor, frequent noise?
The classic error is treating every "new high" the same regardless of lookback. The discipline is mechanical: pick N to match your actual horizon, and exit on a shorter, opposite-direction channel rather than a fixed target.
Nothing but the raw highest high and lowest low gives traders the cleanest, least ambiguous definition of a breakout in this course — but that same rawness makes a short lookback genuinely cheap to whipsaw.
The trend is your friend.