Skip to main content
Track Hull moving average
4 / 6
Search the archiveLibrary
Library

2017issue C0830-33

A Hull moving-average forecast in a credit-spread trend procedure

A 2017 case study uses a Hull moving-average convolution as a next-bar directional forecast, then lets momentum, a derived probability, and a volatility-rank flag decide when a bull-market credit-spread procedure may enter, exit, or sit out. Concurrent index spreads share that one rule set.

  • The hull-moving-average baseline is written as a convolution-forecast: a one-bar linear-regression extrapolation of price, then a second one-bar extrapolation of that projection.
  • A momentum-strategy check allows a new long-bias entry only when that forecast is higher than on the prior bar.
  • Volatility-rank-abstention keeps the book flat from an implied-volatility-rank cross above 65 until rank later falls below 50.
  • The same trend-following procedure also exits on a weakening forecast, a convolution-probability rise through 60 percent, a partial-credit-exit at one third of the credit, or a stated loss cap.
Entries in this reading3 entries
Current trackHull moving average6 readingsMomentum strategy113 readingsTrend following168 readings

A forecast inside a trend procedure

A 2017 case study models a convolution-forecast on the Hull moving average and uses that forecast as the directional input to a bull-market trend-following procedure on index credit spreads. Trend-following is a rule set that stays with a confirmed rising market state and stands aside when that state is not present.

The hull-moving-average is a fast, smooth moving-average baseline used here as the model for a next-bar directional forecast.

How the convolution forecast is specified

The convolution-forecast is a one-bar linear-regression extrapolation of price, then a second one-bar linear-regression extrapolation of that projection. In the reported test the lookbacks are 9 and 3.

A convolution-probability series is derived from that forecast and is used with it to time entries, exits, and standing aside.

Momentum confirmation and the long-bias entry

A momentum-strategy layer requires the forecast itself to be improving versus the prior bar before a new position is allowed.

A new long-bias entry is allowed only when the forecast is higher than on the prior bar, its derived probability is below 60 percent or decreasing, a historical implied-volatility-rank flag is false, and a stop confirms that price is rising through the close.

Exits and standing aside

The same procedure exits or stands aside when the forecast is lower than on the prior bar, the derived probability rises through 60 percent, the volatility-rank flag turns true, profit reaches one third of the credit, or loss exceeds half the sum of the short strike and the entry price.

The one-third close is a partial-credit-exit: closing a credit position once a fixed fraction of the collected premium has been earned.

The volatility-rank sit-out band

Volatility-rank-abstention is a market-state flag that keeps the system flat while implied-volatility rank sits in a designated unstable band. The flag becomes true when rank crosses above 65 and stays true until rank later falls below 50, marking a period the system treats as unpredictable.

Credit-spread book and fill rules

Positions are concurrent bull-put-credit-spread structures on SPX and QQQ. A bull-put-credit-spread is a defined-risk options structure that collects premium when a short put strike is expected to finish out of the money.

Short strikes are limited to those with at least a 55 percent chance of expiring out of the money, and that probability is not rounded. The reported test covers January 2013 through December 2015, charges 12.50 dollars per round trip, and prices fills from option-chain data rather than from end-of-day marks only.

Holding time in the reported test

Holding periods were short. The longest SPX trade lasted 8 calendar days and the longest QQQ trade lasted 6. More than half of each series closed within two days, and both underlyings were idle about two-thirds of the sample.

Most of the brief holds were attributed to two-bar local reversals in the convolution-forecast rather than to waiting for expiration.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
4 of 6 in the Hull moving average track
20198-8 pp.Next on Hull moving averageConstructing nested regression projections as smoothersA linear projection fits a chosen price series to a running bar index over n bars and reports the fitted value at the current index.
All readings on this track · 6 readings
  1. 2010Constructing a Hull smoother and a Hull-based oscillator gate
  2. 2015Evaluating moving-average lag in cycle filters
  3. 2017Rebuild a Hull moving average with linear-regression intercepts
  4. 2017A Hull moving-average forecast in a credit-spread trend procedure
  5. 2019Constructing nested regression projections as smoothers
  6. 2020Evaluating countertrend ETF rules against a buy-and-hold control
By fieldTrend and mean reversion140 readingsBy year2017182 readings
All 6 readings tagged Hull moving average
Also on Hull moving average5 readings
2010Constructing a Hull smoother and a Hull-based oscillator gate2015Evaluating moving-average lag in cycle filters2017Rebuild a Hull moving average with linear-regression intercepts2019Constructing nested regression projections as smoothers2020Evaluating countertrend ETF rules against a buy-and-hold control