For the Acme FP short entnes in Figure 4.5, Jumper broke down twice from low bases. In the first instance, the fioat tumed o-er in only six trading sessions. After the fioat tumed o-er, the stock did not go down much further. In the second instance, the trade could have been stopped out on the third day when the stock gapped up. Since the stockwas in a downtrend, we waited out the gap on the open to see ifit held before covering the short position.

JNPR LAST-Daily Acme Float Channetjl .0,7.0.3) 32.850 24.010

52.000 48.000 44.000 40.MO 36.000 32.000 28,000 SSri?>24.M0

Heme Float Percenlfl .0,7.90.0,0.3) 84.1i

4 11 18 25 J S

4.5. Juniper Networks

When a long base is being formed, an FP signal may precede an FB signal. In this example, the price continued downward and broke below the extended base ofthe lower float channel the same d-; however, in both cases the price did not form a table near the lower float channel, therefore an FB signal would not have triered. Essentially, both float sj-stems work in concert. The trader gets an opportunity to enter on a pullback (FP) or waits until the stock consohdates near the base and breaks out (FB). Ifbofh conditions occur, then the traderwiU get the FP signal first, fohowed by the FB signal usuahy one or two days later.

In cases where multiple signals occur within one or two ds,we simply treat each siinal separately, even if it means doubling; a position in the same stock. A second signal serves as confirmation for the first signal; we are more confident with a second entry in the same stock rather than an isolated entry in a different stock; the concept of multiplici applies here. too.

4.5.3 Ariba

The Acme FP short entrj- in Figure 4.6 bounces off the upper float channel. This is a strongly trending stock, emulating an Acme M or N swing entry. Although the FP sjstem rehes stricfly on DMI, the trader is free to add an ADX filter for FP entries only. Using the ADX filter on FB entries only eliminates good trades because when a stock forms a base, the ADX is bound to be low.

For non-trending stocks, the float channels ser-e as support and resistance -be wary of taking breakouts in trading range stocks. In the example, after the short entrj- was co-ered, there was a break below a fourteen-bar base. The FB sjstem did not go short on this breakout because the stock did not form a table near the lower float channel. Further, the stock had aheady declined over 30° in seven days. In these situations, atest usually occuis, fohowed by arebound for a few days before a new trend is estabhshed.

Eigiire4.6. Ariba

With experience, the trader learns that stocks have symmetric consohdations after protracted moves. A stock that advances or declines 10°o in four or five days usually requires the same amount of time to diciest the move. The leneral market tends to follow a similar cycle, alternating an NR day with a WR day, or an NR2 day followed by two WR days, etc. Recently, the market had six NR days followed by six WR days, not the swing traders optimal pattern

4 Float Trading

4.5 Examples

4.5.4 Ciena

Figure 4.7 is an example ofan FB short e tI5 Ciena was m a downtrend and its high touched the upper float charmel on August 2" 2001. Six d-s later, it approached the lower float charmel. After gapping down and closing near the high ofthe day, the stock formed a 180 pattem [4], completing a three-day table in the process. The FB short entrj triered the following day.

CIEN LAST4)ally Acme Float ChanneK1.0,r.(l-3) 19.771 13.390

-38.000

32.000 0.

1 8.

-28.000

24.00

120.000

Figure 4.7. Ciena

This chart is another illustration of pattem conveigence. Although the chart - does not show all ofthe Acme indicators, we highly recommend the practice. When the trader sees all ofthe indicators working in concert, his or her eyes are opened to all ofthe possibilities, leading Uttle to chence. We prefer this technique to interpreting some vagae head-and-shoulder formation.

The ultimate goal is to encode as much technical and fundamental information as possible and annotate the chart with it: the letters and lines approach. Too many times to count, we have chosen between multiple signals only to miss a critical piece ofinformation that distinguished one ofthe signals among the many others. Through the process ofobser\ation, the trader can simply encode the newpiece ofinformation and use the chart as an artists palette, each color representing; a unique characteristic ofthe stock. By automating; the stock selection process as much as possible, the trader is free to focus on new techniques.

4.5.5 Check Point Software

Figure 4.8 shows an example of float turnover for CHKP in late November 2000, as shown by the Float Percentage histtram reaching 1009-O after breaking below a long base. Although this technique is not encoded in the Acme F Sjstem, an alternative way of plaj-ing float tumover is to let a stock break out. then wait for it to reach lOOo tumover wldle it is trending. Once the float turns over, take a counter-trend trade on an exhaustion move, using either the Acme M or Acme V sj-stems as confirmation. In the example below, a long position taken at the 100° tumover mark would have netted o-er 30 points in five daj-s. Because ofthe volatilit} this may be a strategy better suited for options.

10 Acme Float Channel[1.0,7,03) 112.8M .250

Figure 4.8. Check Point Software

Now, examine the second Float Percentage histtram in Figure 4.8-it is not a mirror image ofthe stock price hke the first one. This FP started at zero in late November (the first histogram overlaps it) when CHKP broke below the lower fioat channel. The stock bottomed and reversed, approaching the upper fioat channel in eariy December. At this point, the FP reading was 70° o, with price near resistance at 110 established in eariy November.

Moreover, the stock had formed a V-bottom pattem, encompassing a fifty-point move. In this case, the tiader does not wait for a full tumover to exit the trade because ofthe dispari in volume between the down leg; and up leg; ofthe "V. Over 60°o ofthe flost changed hands durin« tiie down le«, but only 30°o of the float turned over on the up leg. Clearly, the stock had rebounded on some hot air, outpacing M

f last-daily - / ! acme float channel1-0,7.0.3) 4s.550 33.750

50.ddo 48.000 46.000 44.000 42.000 40.000 38.000 36.000 34.000

l5 22 29 it 5 12

j-32.0l

fure 4.9. flir systems

4.6 Float Trading Strategies

1. when the trade entry coincides with a float percentage greater then 90° o, the probabihty ofsuccess increases dramatically. use the alert feature ofthe float percentage indicator to find charts completing a float tumover.

2. for acme fb long and short entries, use the double bottom and double top indicators to assess the trade entry. if the stock price has traversed from the top of the channel to the bottom of the channel and ttien back up again, ttie breakout may be false and ttiere may be a test

pattem first. in this case, buy a puhback near the upper channel and sell short on aretracement up from the lower channel.

3. because tradestation generates its performance report for one instmment at a fame, we have included trade losing with the function AcmeLogTrade. when trade logging is tumed on, each trade is written to a common text file that can be imported into a microsoft excel spreadslieet. by running the tradestation 2000i worl:si3ace assistant on a portfoho, trade wiu be recorded for each sjmbol stored in the globalserver.

4. the current float value is vahd only as far back as the most recent split date. unless the software automaticehy adjusts the float for stock splits, interpreting the float tumover wiu be meaningless for pre-split data. for example, ifa stock has a float of 10 million shares and a 2:1 stock split occurred on january 1, then the pre-split amount was 5 million shares. when analyzing historical data, aust the float value to its pre-spht level.

45.6 fur systems

the cliart in figure 4.9 appears to be the picture-perfect chart. three times in the month of october, the stock attempted to break out aboe the upper float channel. the stock consohdated for five days in early november, and finally it broke out the channel. after one day of going slightly higher, the trade was stopped out for a loss and then feh into awaterfah oer the next five ds. clearly, if the apparently perfect trade does not follow tluough, then take the loss and fade the original trade consider how many people got into this trade waiting for it to break out.