back start next


[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [ 38 ] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54]


38

Hgure 10.13. Acambis News Continuation

Figure 10.14. Rambus Breakout Continuation

The breekout continuation is a two-day pattem. On Day One. the stock must consolidate the entire day untU later m the aftemoon. Then, one or two hours before the closing beh, the stock trends strongly in either direction. This is an example of a pent-up mo-e that will probably continue on Day Two, unhke a stock that trends ah day on Day One and continues into Day Two. The strategy is a simple application ofthe alternation principle that a trending day fohows a non-trending day, but the key is to reccgnize the beghming ofthe trend phase late in the aftemoon of a predominantly flat day.

The breakout continuation is a pattem for real-time scanning prcgrams such as FirstAlert. The breakout follows the pattem principle of a rectangle, where the rectangle height of the breakout period is compared to a smaller range height for a longer time frame preceding the rectangle. In this case, we divide the trading day into half-hour interi-als, so the trading day has a total of thirteen interi-als. Since we scan for abreakout during the last two hours ofthe trading day, we want to calculate the ratio of the range height of the last four thirty-minute intervals divided by the range he ight ofthe previous nine thirty-minute intervals, e.g., a ratio of21 or 3:1. Further, the height ofthe reference range must be narrow, as little as -20° of the stocks ATR. Figure 10.14 shows abreakout continuation for Rambus (RMBS:Nasdaq).

10.3.3 Block Trading

Previously, we presented some examples ofhow block trade analj-sis can detect imusual activit} in a stock. Before we explore the mechanics ofthis technique further, lets compare the a-ere blocktrade with the a-erage trade size for the Nasdaq stock market. As shown in Table 10.4, the a-erage block size is o-er twent} thousand shares, but remember that a trade quahfies as a block only in the context of a stocks ai-erage daily volume (ADV).

For example, if the ADV of one stockis 100,000 shares, then a trade of5000 shares could be considered a block trade. For a stock with an ADV of one mil-hon shares, then trades of 25K and 50K would be considered block trades. For stocks such as Cisco Sjstems with an ADV ofmiHions ofshares, blocks become more difficult to interpret.

Table 10.4. Nasdaq Block Size

Ycur

Awragf Trade Sizr

Awnig,- Bloct Size

.>l)l)l)

.>(in

: 1,"



One accepted truism of day trading is that large blocks signal a trend reversal. This concept has appeal because a large block to be sold wih temporarily force down the price. A trader who is alerted to a large block on the tape maybe able to ride the momentum back up; however, this tjpe of trade has two problems. First, by the time the "print" occurs in Time & Sales, the stock may have aheadj rewrsed because Nasdaq members have up to ninety seconds after execution to report transactions; thus, the print can be delayed. Second, the trader has no idea whether or not another large block is coming down the pipe

Without any insight as to order flow, fading a block trade may not be worth the risk. Only the participant with access to order flow can buj or sah ahead. A trader ofNYSE stocks subscribing to the OpenBook ser\ice (raleasedJanuarj 24, 2002) can wew tiie spedahsts Ihnit order book from 7:30 am to 4:30 pm. Here, the trader can devalop a sense ofthe technical le-els that maybe breached and anticipate any movement towards those areas.

The key to block trading is to measure the frequency of certain block sizes across the spectrum ofmarket capitalization. Ifalarge block needs to be bought or sold for a smah-cap stock, then the probability is greater that the block is a one-off, and the trader may be able to participate in a reversal. We recommend that the trader set up a group of separate tickers, segregated by ADV to display only those trades that meet the minimum block size. Table 10.5 shows sample block sizes sorted by ADV,

Table 10.5. Volume-Based Block Size

: Size

0 -

j 0

j,(

10,000

E 0

7,

l,viO,0vi -

10,VK)

j,o---- -

10,0----0

25,0-3

We caution the trader not to place too much emphasis on a single block trade. People seem to get excited about seeing a large print above the offer if they are long or a large print below the bid if they are short. The isolated print serves only as apsjchological boost to the nervous trader, who should probably not be in that position ifhe or she is dwelling on every tick and consulting the oracle of the Yshoo board. More importantly, examine a string of block trades to see how many were executed on dov\nticks and how many on up ticks. The trailer is simply trying to assess trend and possibly impending news. When a low cap stock suddenly shows up on the ticker, then that is a sign to get involved.

Ni>i, \vi4-\.iiiimi ihi iuii.i(l.i\ I

ik.-i,

10.3.4 Spread Trading

When spreads were wider in the fractional dajs, spread trading was an actiitj best reserved forthe market maker. Still, a trader could "play market maker" in a liquid stock by simultaneously placing abuy order at the best bid price and a seh order at the best offer price using an ECN such as the Island. For example, if a stockwere trading at 40 4 x 41, then a trader could bid the stock at 40 5/16 and offer it at 40 1516-there may have been a seher and buyer who were willing to take ad\antage ofthe betterprices given by the spread trader.

Unfortunataly, for stocks with wide spreads, the trader with no knowledge of order flow is a sitting duck. If a traders bid is hit, then it probably happened for one reason-the stock is gcdng down, and the trader still has an offer to sell his or her shares. Now, other market participants see that the bid was hit, and start going low offer. In an attempt to sell the shares, the tradergoes low offer as weh, cutting the spread and sowing the seeds ofthe stocks demise. The trader wih be lucky to get out ofthe trade without a loss.

Currently, the spreads are as narrow as possible, so unless the practice can be automated, spread trading is intense and is not the best use ofthe traders time. A large-cap stock such as Cisco trades with a penny spread. To make money on the spread, the commission costs must be factored into both sides ofthe spread trade. For example, ifthe commission is $10. then the total cost ofthe trade is $20, one trade for the bid and one trade for the offer. At least 2100 shares must be spread (2100 X $0.01 = $21 - $20 = $1) to make any profit at ah. Even with a stock such as Cisco, a price j ump could move the stock twenty or thirtj cents, and ah ofa sudden, the trader has risked se-eral hundred dollars to make a buck, converting a spread trade into a position trade.

Spread trading has been subsumed almost enthely within the domain ofthe computer. For the laige-cap stocks, many ofthe ECNs are lined up on either side of the Level 11 window with thousands of shares displayed on the screen. Traders that used to watch the volatilitj on the Level II screen are now forced to watch as automated programs swap hundreds ofthousands ofshares before any appreciable price movement. The free-flowing volatilitj of the past has erolved into apattemoftight consohdations alternating with sudden price shocks.

So far, we have covered the fohowing day trading techniques:

- Gap Tradmg Continuation Trading Block Trading

- Spread Trading



Period

Before the Bell 08:00 pm-09:30 am EST

The Open 09:30 am - 11 0 am

Lunch Hour 11:00 am - 02:00 pm

The Close 02:00 pm - 04:00 pm

After the Bell (M:00 pm - 08:00 pm

The fime period in the first row is not a tjpographical error. The new trading day starts just after the close of after-hours trading at 08:00 pm, putting us on the 24-hour cycle. One may question our designation of the Open and Close segments with their expanded time fi-ames, but they serve to delineate the time periods whan trades are entered. New trading positions between 11:00 am and 02:00 pm are rare.

10.4.1 Before the Bell

The pariod before the beh is divided into two \.

a Research phase (08:00 pm - 08:00 am)

a Trading phase (08:00 am - 09:30 am) The research phase-the process of downloading price data, scanning charts, and salecting stocks is discussed in Chapter 9. The trader should have ah ofthis work done before trading begins at 08:00 am, although some traders prefer not to trade either before or after the beh, in which case the trader can hit the snooze button. Use the time before the opening bell to set up charts, enter alerts, and scan for gaps. The trader may also hai-e other research services and pubhcations to re\iew beforehand. Finahy. any remaining time can be spent laughing at the gm-s on Squawk Box.

> 1 .11 .-. I......-.............I.,...mil...

S.l..vvls lt..>, l: iii.............W-l-W-........-1

It is a daily moming ritual for some traders-the business sertion, coffee, and the futures check. One of the first things to do is flip on the television and get the latest S&P futures quote displayed in the lower-right hand comer ofthe screen. This quote, also known as the S&P bug. shows the positive or negative change from yesterdays S&P 500 futures contract close.

The purpose ofwatching the S&P futures in the moming is to assess the general direction ofstocks because futures are aleading indicator ofstockprices. Unless the futures are \er\ strong (e.g., greater than +5.00 or less than -5.00\ then the market open whi be difficult to predict. As a trader ewntuahy teams, a positive futures change does not imply a strong opening, and a negative futures diange does not imply a weal: oijening. This price dscrqaancj- is e>q3lained by the trading concept movmssfair value

Fair v-alue is an estimate ofwhat an S&P 500 futures contract is worth; it is a formula that fectors in borrowing costs and dividends. Fair value is computed at the and of each trading day to compare with the actual futures price. Before the market opens, the S&P futures serv-e as a market prox\, digesting any news to trade abov-e or below fair value. For example, a bullish economic release at 8:30 am whi send the futures soaring beyond their fair value. The key point is to know where futures are trading relative to fair value. Some business channels such as CNBC display this value before the opening beh.

Market commentators alwajs give their perception of a strong open or weak open for the market. For now, hit the mute button to make your own determination. Firsl, calculate the net change for the S&P futures fiom yesterdays close. This value is displayed with a "-i-" or "-" point value on the tale\Tsron screen. For example, ifthe S&P futures are +2.50. then the futures are trading two and a half points higher in the moming trading session. Then, get the fair value displayed on the screen; this value is also displayed as a "+" or "-" point value. For example, if fair value is "-6.00", then futures closed sis pornts abo-e fair value yesterday. Ifthe fair value rs "+6.00", then the futures closed sis pornts below fair value yesterday.

Now, compare the current S&P futures quote against iarrvalue to determine how the market is going to open this moming. Simply subtract the fair value number (F) fiom the S&P futures quote S). Apositive number indicates abras to the upside; anegativenumberriidicatesadownsrdebras; andzero means that the market will open flat. Some sample combinations of fair value and flitures are shown in Table 10.7.

III. N....l.i,(-......,l\ll,

..........I ill! (......illi.ltim I, !\Liif;l l4,ii,l4.

ir 1 ..<1.(....... imiii.........1.......1..............ill (............ till i,

IMI I ImIIMI i.t.tliiiLtiiJ lim iIh- N,v.iI.u 1(1(11 l> !< Dl .

10.4TheTradingDay

The stock market is expanding on either end ofthe d, a natural extension into round-the-clock trading. Assuming a 24-hour trading day, we divide the day into five natural segments as shown in Table 10.6.

Table 10.6. Trading Day Segments



[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [ 38 ] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54]