Skip to main content
Track Volatility stop
1 / 12
Search the archiveLibrary
Library

1989issue C071-4

A close-only volatility reverse bound to average true range

This archive construction keeps a trend-following book always positioned and reverses only on the close. Average true range times a constant factor sets how far that reverse sits from the extreme favorable close, so the allowed distance expands and contracts with market speed.

  • Stay always positioned in a trend-following book and reverse only on a close-only stop that trails the extreme favorable close of the open trade.
  • Measure volatility as average true range. The reverse sits the average-range constant away, which is that average times a constant factor.
  • Raise the constant factor to place the reverse farther from price and flip less often. The trail widens when ranges expand and tightens when they contract.
  • A close-only rule can let a large adverse move print during the session before the stop can act, especially in markets without daily limits.
Entries in this reading3 entries
ATR position sizing4 readingsTrend following7 readingsCurrent trackVolatility stop12 readings

An always-positioned trend-following reverse

The construction is an always-positioned trend-following procedure. It stays long or short and reverses with a close-only stop that trails the extreme favorable close of the open trade.

Entry, reverse, and the decision to remain in the market are one rule set. The reverse is a volatility-stop: it may fire only on the close, and its distance from the most favorable close of the trade expands or contracts with measured volatility.

Average true range as the speed input

Volatility is measured as average true range over n days. True range is the greatest of the session high-low span, the session high versus the prior close, or the session low versus the prior close.

After the first n-day average, later average true range values can be updated as the prior average times n minus one, plus today's true range, divided by n. A 10-day example multiplies the prior average by nine, adds today's true range, and divides by 10.

The same volatility input loosens the trail when ranges expand and tightens it when ranges contract. Stop distance therefore changes in proportion to average true range.

Constant factor and the average-range constant

Stop distance equals average true range multiplied by a constant factor. A larger factor places the reverse farther from price and produces fewer flips.

Exit and reverse are defined at the average-range constant, which is average true range times the constant factor, measured from the extreme favorable close.

In the supplied construction example, a 200-point average range and a 3.5 factor put a long close-only stop 700 points below the highest close of that trade.

What the construction search varied

The construction search examined volatility lookbacks from 6 to 20 days and constant factors from 1.0 to 4.0.

Editorial note: the archive records that range as the historical workflow. The facts allowed here do not say which pair was chosen or how any pair behaved.

Closed profit and peak drawdown after the volatility-system search

Each market stayed always in the book and reversed only on the close, with the reverse sitting average true range times a chosen constant from the best close. Coffee, T-bonds and the yen posted the largest closed profits; gold barely paid, and the S&P 500’s 26,775 of profit sat on a 61,550 peak-to-valley hit. Every bar is copied from Aan’s published test table, not read off a plot.
Each market stayed always in the book and reversed only on the close, with the reverse sitting average true range times a chosen constant from the best close. Coffee, T-bonds and the yen posted the largest closed profits; gold barely paid, and the S&P 500’s 26,775 of profit sat on a 61,550 peak-to-valley hit. Every bar is copied from Aan’s published test table, not read off a plot.Daily, about 5.5 years · 1983-04-01T00:00:00.000Z to 1988-06-30T00:00:00.000Z

Aan subtracted $100 from every trade for slippage and commission. The day count and constant factor he lists for each market sit inside the search he ran: 6–20 days for average true range and 1.0–4.0 for the multiplier. Sample windows differ slightly by contract, mostly mid-1983 through mid-1988.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
1 of 12 in the Volatility stop track
19921-5 pp.Next on Volatility stopEquity-curve average as a live-capital gateDaily account equity is plotted as its own series so it can be judged as rising, falling, or flat, independent of the market-price chart.
All readings on this track · 12 readings
  1. 1989A close-only volatility reverse bound to average true range
  2. 1992Equity-curve average as a live-capital gate
  3. 1992Constructing volatility-adaptive trailing stops
  4. 1993Constructing skew-adjusted volatility stops and pyramid size
  5. 1999Evaluating a long-only breakout system with a volatility stop
  6. 1999When markets burst, not trend
  7. 2005Building entry rules with ratchet volatility stops
  8. 2005Pricing entries, stops and exits in range units
  9. 2013Constructing asymmetric volatility bands for reversal, trend, and stops
  10. 2015Mark the stop, the target, and the invalidation line before entry
  11. 2019Bounding capital risk with phase-aware stops
  12. 2019Measure the Bollinger Bands touch before adding engulfing and a volatility stop
By fieldTrend and mean reversion6 readingsBy year19895 readings
All 12 readings tagged Volatility stop
Also on Volatility stop5 readings
1993Constructing skew-adjusted volatility stops and pyramid sizeATR position sizing1999Evaluating a long-only breakout system with a volatility stopATR position sizing2002Constructing volatility stops from average true range and parabolic SARATR position sizing1999When markets burst, not trendTrend following1989Evaluating always-in parabolic SAR trailing stopsTrend following