Minimum Pnce (MP)

The conventional wisdom is that the professionals ignore stocks that trade for less than $20 per share. The problem is that a minimum price screen filters out many volathe stocks, wMle less volatile high-priced stocks pass the minimum price screen. For example, if Stock A is trading at $60 and has an ATR of 1.5, and StockB is trading at $10 with an ATR of 12, a screen based on aminimum ATR of 1.5 eliminates the more volatile Stock B. To compare the volatilit} of the two stocks, we divide the ATR by the stock price to calculate the Volalilitj Percentage. For Stock A, the VP is 1.5 / 60 = 2.5%. In contrast, the VP for Stock is 1.2 / 10 = 12.0° o. A volatihtj measure is a better trade filter than Minimum Price, alttiough using both is an even better trade filter.

For low-priced stocks, screen for both volatiht} and hquidity. At the rime of the chart in Figure 1.9, Ariba had a 20-d Volatihty of ov-er 1.5 and traded under $10 per share. Further, the stock traded an av-erage of sev-eral million shares per d. As a result, the stock passed the filtering process and had two trades dimiie this period with gains of ajroxmiately 20° o. For the trader starting out with a smaher stake, these volatile, low-priced stocks are alcgical choice.

ftRBALAST4Jaily Mo>f Avs 1 line 9.767 Acme Rectanflte Acme AH StnAepes

1Z.O00 -11.000 10.000

7.000 0.000 15.000

Rgure 1.9. Ariba Low-Priced Stock Examp I e

III! I I llllSl........ I..t,l,/,1, 11

ihi. .ihhi vi......iii\ ill iii

- ill. V.-i.ip Inn lt.lll,-. . ,.lll.

Example 1.4. Function AcmeVolalHity

AcmeVolatility: Calculate the annualized histori

: volatility

Inputs:

Length(Nunieric};

Variables:

DaysInYear(365), DaysInMonth(30), DaysInWeek(7L TimeFactor(O.o);

AcmeVolatility = 0;

If Close > 0 and Close[lJ > 0 Then Begin

If DataCompression >= 2 and DataCompression < 5 Then Begin If DataCompression = 2 Then (Daily}

TimeFactor = DaysInYear Else If DataCompression = 3 Then {Weekly}

TimeFactor = DaysInYear / DaysInWeek Else If DataCompression = 4 Then (Monthly} TimeFactor = DaysInYear / DaysInMonth;

AcmeVolatility = StdDevCLogCClose / Close[l]), Length) SquareRoot(TimeFactor); End; End;

Vlfl. jii.<llllllm

dull.-.....Ilu

.,i./..ili. I.,..l,ii.ii,ii

vii i i ll.tt)il i. I •"ll

A trader wants the three Vs: volatihty, volume, and a smah vig [21]. Volatiht} creates the opportunity to go long or go short, volume provides the hquiditj to get in and out ofthe position, and the smah vig limits the amount of money that lands in the pocket ofthe market maker or specialist.

Historical Volatility (HV)

Each stock has Historical Volatility (HV). It is an annualized percentage that measures the standard deviation ofa stocks price changes over a period oftime. e.g., the percentage change of todays close compared to yesterdays close for the last thirt} d-s. The historical volatili calculation assumes that stock prices fall in a lognormal distribution and is derived according to the Black-Scholes options model [5].

The EasjLanguage code for calculating HV is shown in Example 1.4 The HV can be calculated for daily, weeldy, or monthly charts. Depending on the charts time frame, the function calculates a multiplier to determine the annualized HV. The HV calculation uses a sample bar range based on the input parameter Lenglhto extrapolate the annualizedvolatility from the closing price changes for the sample period. The steps for calculating the HV are as follows:

a Calculate the TimeFactorbased on the chart periodicity.

a Compute the standard deviation ofthe sample based on the natural

logerifhm ofthe closing price percentages using the last Length bars, a Multiply the standard deviation by the TimeFactor to determine the

annualized HV.

For a daily chart, the TimeFactor is sinqily the mmiber of ds in the year. For a weeldy chart, rt is the mmiber of dajs in the year divided by the mmiber of dajs in a week. Historical volatilities are measured over various periods of time, but the 30-d HV ( ) is common in many options models. TheHV30 gives the trader an estimate of a stocks travel range. For example, a stock trading at $20 with anHA30of20Oo win have traded 20 X 0.2 = 4 points above and below the current price approximataly 68°o" ofthe time during this period, based on a normal distribution [29]. The hv3d ofthe stock in Figure 1.10 is 1.36, or 136° which is extremely high.

TWOLASr-Daly D8/3in001

4cmeHV(3B,B.5) T.3B

1G.O00

-ie.0D0

14.000 12.000 1d.O00 8.000 -6.000 1.50

IiKiri-I.IO. llisi.M»4l v..I.1I1I1IV

"ll.i-<-l.....I>. 1...11 11>1..>.1>.1.1 Mim. I -.Uli.hml.l.vi

In general we require a minimum HV reading of 0.5 to filter out non-volatile stocks; however, higher readings are desirable The trader should experiment with various HV values to scope his or her imiverse of stocks. The IVolatilitj Web site at http:/www.ivolatility.com has the 30-day HV readings as well as Iniplied Vohtility (RO readings.

NarrowRange (NR)

Crabel pioneered the use of Narrow Range (NR) bars by assigning them to categories such as NR4 and NR7 [6]. For example, an NR4 bar is the bar with the narrowest range ofthe last four bars. Other variations of narrow range bars have since been developed, combining them with inside ds to produce other patterns such as the 1D/NR4 day [3].

A narrow range bar can also be de fined by framing its range in the context of the ATR. By definition, a narrow range bars range must be less than the ATR, but tlie NR bar is generally defined by a smaller percentage ofthe ATR For example, if a stocks ATR is two, and the NR percentage is 60° o, then abarwith a range of 2 X 0.6 = 1.2 or less would qualtfj as an NR bar. An NR percentage that ranges between one-half and two-thirds of the ATR is recommended as the maximimi value, as shown inFigure 1.11.

�999999999

If a trading sjstem places a stop at or around the previous bars high or low. then the range ofthe bar diaaies the size ofthe loss. Thus, an NR bar impro\es the lisk/rewaid ratio ofthe trade because the loss is inherently hmited to the narrow range. Further, a narrow range dayimphes a greater-fhen-even probabihty that a wide range (WR) day with a range greater than the ATR wiU occur the next day. A cluster of NR dsys means that the market is anticipating a major news event, such as a Fed meeting on interest rates or a key economic number.

Average Directional Index (ADX)

The ADX is simply a measure ofthe strength of a trend and has been cov-ered in depth by other authors [3, 4]. As a general rule, ifthe ADX is rising, then a stock is trending strongly-either up or down. The ADX is used in combination with the DMI for momentum trading sjstems. Although most sjstems use an absolute value of ADX to assess a strong trend (e.g., aminimum of 25 or 30), the ADX for a strong stock in a pullback wiU fall as low as 15 Thus, when screening for trading candidates, consider the ADX five or ten ds ago along with the current reading.

A characteristic ofthe ADX is that a rising value indicates a strengthening trend. This is tme, but a stock develops a strong trend weh before the ADX reflects the movement ofthe stock. Geometric breakouts from along or short base trigger signals much earher. The stock in Figure 1.12 had a 30°o move before the 14-daj ADX even reached 30 in late September.

IVDALAST-Daly 10 1 01

t*t*nn

I 30.000 26 00

14) .2

lim.- 1.12. V,.Imlrv

Directional Movement Index (DMI)

The DMI has 2 components: +DMI and -DMI. If-DMI is greater than -DMI, then the trend is up, and ifthe -DMI is greater than +DMI, then the txend is down. Figure 1.13 shows a crossover ofthe two lines under the 50-day moving average. Combined with a weakening ADX trend (the thick line), this crossover is tjpicahy a good shorting opportunity, and the same principle apphes to long positions initiated above the 50-day moving average.

In Figure 1.13, the DMI lines widen near the end ofthe chart (beginning of March). When the spread between the two values is wide, a position should be covered. In this case, the down tick in -DMI corresponding with the up tick in +DMI is an opportunity to either cover a short position or go long.

ABGXLAST4}a>ly D3flM/2002 Mov Avg 1 lineClase,SO,D) 25.095

) 20.83 34.43 27.70

32-000 -30.000

18.000

ie.ooo

40.0D -30.OD 20.00

? WA

>«<-......mI M..<• -nl lii.lix

Each technical indicator has its niche, however. As shown in Figure 1.12, high ADX readings are useful forpullbacks (denoted by P) in very strong trends. A retracement of two or more bars is usuaUy interrupted by a resumption ofthe prev-ailnig long-term trend.

Returning to the example, a strong reversal begins in eafly October, and the ADX does not resume rising until weh imo the rev-ersal. Thus, treat the ADX as a lagging indicator-the trader wiU benefit from shortening the studj length from 14 to 7, especiahy for short sales.