2005issue C011-14

Constructing a multi-filter penny stock breakout procedure

A breakout mark is placed on the close only when volume, trend, momentum and volatility tests all pass on the same bar, rather than when any single chart condition fires.

The procedure marks a close only when every listed Boolean test is true on that bar.
MACD, slope bands, volatility-adjusted ADX, compressed variability, signed volume and two price-structure checks are independent gates, not standalone trades.
Slope and variability tests are written to keep the setup in a quiet, compressed base rather than a steep prior trend.
An independent reconstruction can be checked against the same filter family rather than against a single chart pattern.

Entries in this reading3 entries

Breakout system48 readings Current trackAverage Directional Index71 readings MACD115 readings

Several filters, one close mark

The breakout procedure generates a mark on the close only when every listed Boolean test is true on the same bar, rather than when any single indicator fires. The listed tests cover momentum, prior trend, volatility-adjusted trend intensity, compressed variability, signed volume and two price-structure checks.

In platform terms this is a show-me overlay. The bar is marked only when every Boolean condition in the procedure is simultaneously true.

Momentum as a histogram gate

The MACD filter is a price-momentum oscillator built from two exponential averages of close. It is computed from close with a 9-period fast length and a 20-period slow length, then turned into a histogram by subtracting a 9-period exponential average of that MACD.

The MACD histogram must exceed a threshold of -0.003 before that momentum condition is counted as true. A histogram that fails the threshold leaves the whole conjunction false on that bar.

Slope bands for a quiet base

Normalized close slopes over 35, 70 and 170 bars must stay inside separate band tests. Each linear-regression slope is a normalized slope of close over its lookback window.

The three windows are not treated as trend-following triggers. They keep the setup in a quiet base instead of a steep prior trend. If any window leaves its band, the breakout mark is withheld.

A volatility-adjusted trend block

A volatility-adjusted ADX term is formed by multiplying 25-period average directional index by 30-period standard deviation of close divided by its 30-period average. That product must stay below 1.3.

Average directional index is used here as a trend-strength measure scaled by normalized price variability. The product blocks the breakout when volatility-adjusted trend intensity is too high.

Compressed variability

A three-bar stochastic of normalized standard deviation over a 150-bar lookback must stay below 20 so that only compressed variability qualifies.

The ranking starts from price standard deviation divided by a moving average, then scores that ratio as a stochastic. Wide or expanding variability fails the test even if other filters pass.

Signed volume after a cutoff

Finite-volume-element scoring first builds a cutoff from intra-bar and inter-bar log-range variability, then signs volume only when money flow exceeds that cutoff. The resulting percentage must exceed -1.

The accumulator classifies each bar as buying or selling pressure after money flow is compared with that volatility-scaled cutoff. Volume that never clears the cutoff does not contribute a signed reading, and a percentage at or below -1 keeps the conjunction false.

Location and slow K

Two further price-structure tests require close above a 10-period exponential average and 10-period slow K above 30.

Slow K is a bounded oscillator of close versus its recent range. It is used as confirmation that price is not sitting at an exhausted low when the other gates are already true.

The same filter family, reconstructed

An independent platform reconstruction uses the same family of filters, including finite volume element, three regression-slope windows, MACD, a standard-deviation stochastic, an exponential-average location test, an ADX times normalized-deviation product, and a slow stochastic confirmation.

The reconstruction is useful only as a check that the conjunction is fully specified. It does not replace the same-bar Boolean rule that issues the mark.

VION breakout returns with a 10% volume cap

Capping each fill at one-tenth of that day's volume, AmiBroker's reconstruction of the multi-filter penny-stock rules turns a $10,000 VION book into $32,772.65 — 227.73% total and 27.97% a year. The 187.25% and 1,524.77% risk-adjusted lines look larger because the account is only in the market 14.94% of the time. Every figure is copied from the on-screen backtest table, not traced from the price pane.VION · Daily · 2000-01-01T00:00:00.000Z to 2004-11-08T00:00:00.000Z

The original article's 2,728% VION result assumed compounding into tickets as large as the day's tape (about 50,000 shares near $0.30 in April 2004). This run shrinks any ticket that exceeds 10% of daily volume.

Educational research material, not investment advice. Historical source context does not establish present-day performance.

25 of 56 in the Average Directional Index track

20051-1 pp.Next on Average Directional IndexConstruct one playbook that flips with session regimeClassify the session with the average directional index before any long or short rule is allowed to fire.

All readings on this track · 56 readings

By fieldEntry and exit rules206 readings By year2005104 readings

All 71 readings tagged Average Directional Index

Also on Average Directional Index5 readings

1993Confirming n-bar breakouts with ADX and DX filtersBreakout system 2016How an Average Directional Index filter and a breakout entry form one procedureBreakout system 2018An ADX volatility gate for prior-day breakoutsBreakout system 2018Constructing an ADX filter for intraday breakoutsBreakout system 1991Constructing five-session forecasts from stochastic, ADX, and MACD inputs