9.4. 2003–2005 BULL MARKET
When the swiftness with which U.S. forces penetrated into Iraq and took control of the capital city of Baghdad took the crowd by surprise, the market followed through to the upside on March 17, 2003, launching a new bull market. One of the first stocks out of the gate was Amazon.com (AMZN), which we began to play heavily (Figure 9.35). AMZN did not show immediate power as it broke out to new highs on March 17th, right in synch with the market's follow-through on the same day. The stock pulled back over the next four weeks before finding support at the 10-week (50-day) moving average. Volume had dried up sharply on this pullback, so we bought the stock right there.
For the next six weeks the stock closed right at the peak of the weekly ranges for each of those weeks; five of those six weeks closed at new price highs. O'Neil was quick to point out just how constructive this action was, and that he had calculated an upside price target for AMZN of $60 a share. In an interesting coincidence, we owned the stock as well and had used Point & Figure charting techniques to calculate an upside price target of $61 based on the upside thrust of those six weeks where the stock closed right at the peak of each weekly range. AMZN finally peaked at 61.15, with a little climax top type of move in October 2003, and then began building a big, ugly base (Figure 9.36).
Figure 9.35. Amazon.com (AMZN) weekly chart, 2003–2004.: Chart courtesy of eSignal, Copyright 2010
Figure 9.36. Amazon.com (AMZN) daily chart, 2003.: Chart courtesy of eSignal, Copyright 2010
As AMZN continued to build its base, it put in an initial low of 47.67 on the left side of the base on November 19, 2003. The stock rallied for a few days and then spent the next two to three weeks drifting back downward before undercutting the initial 47.67 low at an even 47.00 on December 17th. This was the only time we ever saw O'Neil implement the old Livermore trick of buying off the bottom on a "shakeout-plus-three" situation in real time. On December 23rd, he called to declare that AMZN was a shakeout-plus-three buy based on adding three points to the 47.67 low of November 19th in order to come up with a new buy point of 50.67. On that day, as Figure 9.36 shows, AMZN pierced the 50.67 buy point with volume up sharply on the day. In this particular case, the "shakeout-plus-three" or "SO+3" buy signal on AMZN resulted in a tepid rally, and the stock eventually rolled over and failed, gapping down hard in late January 2004.
One significant takeaway from O'Neil's implementation of the SO+3 buy rule that Livermore would make use of from time to time is that the shakeout in the pattern should be exactly that—a quick and sharp shakeout that immediately turns and begins rallying to the upside, as AMZN did over the next four days following the 47.00 low on November 17th. The stock should not have undercut the prior low at 47.67 and drifted lower or sideways for a few days. The turn off the lows following the undercut should be swift, as this confirms the stock's ability to recover quickly and with some authority, hence giving the SO+3 a look of extreme buoyancy in the stock as it immediately rises back to the surface—the top of the base.
9.4.1. The Past Doesn't Repeat, but Often Rhymes
O'Neil frequently had interesting stories to tell about stocks he played in the past and how they were similar to what was going on in the present. When Netflix, Inc. (NFLX) emerged as an IPO in May 2002 at $7.50 a share, the market was still in a bear trend so the stock immediately tanked, finally bottoming out in December 2002 at $2.42 a share. The stock then began working its way higher. Along with Amazon.com and other dot-com e-commerce survivors like eBay, Inc. (EBAY), NFLX (Figure 9.37) rode the second wave in the Internet stock move among those companies that had actually figured out how to make money running a business on the Internet. This new wave began in late 2002 and into early 2003, and NFLX's DVD-by-mail movie rental model was the perfect service for a consumer population that was opting to stay home more in the wake of the 9/11 attacks.
After a long price run, Netflix began to shows signs of topping as it failed from a late-stage base in April of 2004. Note how this move occurred on low volume, as Figure 9.37 shows. The two weeks just before the week of the late-stage-base-failure it moved up straight from the bottom of the stock's chart base on below-average volume, with the second week occurring on extremely light volume, a sign that buying demand was beginning to wane. Talk of Blockbuster, Inc., the grandfather of movie rental chains, moving into the DVD-by-mail business in early 2004 was weighing on NFLX stock. The stock had, by this time, had quite a price run, so it was obviously late in its "cycle" when this news hit, and O'Neil made an interesting observation at the time, comparing Netflix and Blockbuster to motor-home maker Winnebago and car maker General Motors back in the 1960s and 1970s: "Years ago when Winnebago was making RVs and selling them like gangbusters, their stock was a huge winner. General Motors came along and announced they were going to build these RVs also, but they never made much money in that line of business. This is similar to Blockbuster announcing they are going to compete with Netflix."
Figure 9.37. Netflix, Inc. (NFLX) weekly chart, 2003–2004.: Chart courtesy of eSignal, Copyright 2010
As it turned out, Blockbuster never made a go of the DVD-by-mail business and never became a competitor to Netflix. At the time of this writing, in early 2010, Netflix is still in business while Blockbuster has declared bankruptcy. It is also interesting to note that in 2010 Winnebago is still in business, and General Motors has been taken over by the U.S. government, garnering the nickname Government Motors—proof that some individuals know how to run a business correctly, and some don't, and it is usually the early innovators that prevail over time.
O'Neil viewed 2003 as the first year of an economic cycle, and in February 2004 he was keeping an eye on where we were in the cycle. As he put it, "There is a certain order to a bull cycle. In the first year or initial phase of the bull market the hot, newer names will have their move. Sometime later, 8–12 months or so, big business starts to participate in the new economic cycle and you will see cyclical growth stocks come on. Later on, the capital spending stocks will come on as businesses are flush with cash and start spending on capital equipment. In the old days, capital equipment stocks were the machinery stocks; today they are technology stocks." The tone of the market in 2003 going into 2004 was also much different from the heady days of the 1990s when markets were in a grand, parabolic upside move that ended in the March 2000 market top and bubble-bursting. O'Neil, however, was familiar with how this could cross us up, and he warned, "When I made a lot of money in Syntex many years ago I had trouble making money right after that big success. I would sell stocks too soon, since I figured if they weren't up six points a day there was something wrong—I had gotten acclimated to SYN's strong uptrend and couldn't adjust to slower, yet still profitable stocks." In 2004 we had to adjust from an environment that was simply slower and less parabolic than that of the late 1990s.
In early 2004 there were several textbook examples of topping action that O'Neil called to our attention. Often the phone would ring, and O'Neil would be on the line wanting to discuss a particular stock, making note of its characteristics in real time. In March of 2004, O'Neil pointed out the two-weeks-down-and-two-weeks-up sell signal in Boston Scientific (BSX), a leading stock at the time (Figure 9.38). This occurs after an extended price run when a stock makes a new price high and then immediately pulls back two weeks before shooting back to new highs over the next two weeks, creating a "two-up-and-two-down" formation on the weekly chart. This was an old sell rule of O'Neil's, which is often misinterpreted, but in this case the two-down-and-two-up rule was a textbook sell signal, and the action in BSX here illustrates it well.
Figure 9.38. Boston Scientific (BSX) weekly chart, 2003–2005.: Chart courtesy of eSignal, Copyright 2010
A number of biotech stocks had big moves in 2003, including Gilead Sciences (GILD) and Celgene (CELG). Joining in on the group move was Martek Biosciences (MATK), which had found a way to turn microalgae into two fatty acids, DHA and ARA, essential to brain and eye development. These in turn were added to baby formula, producing a "super formula" for infants. Martek Biosciences formed a late-stage double-bottom base in the first four months of 2004. The major flaw in the pattern, as shown in Figure 9.39, is the mid-point of the "W" formation. Normally, in a proper double-bottom base this should close below the left sides of the "W," but in this case the mid-point exceeds that high. As Martek Biosciences attempted to break out on the right side, O'Neil pointed out this flaw. As it turned out, the breakout failed, and the stock broke down altogether over the ensuing months.
Figure 9.39. Martek Biosciences (MATK) weekly chart, 2003–2004.: Chart courtesy of eSignal, Copyright 2010
9.4.2. The IPO "U-Turn" Base
In the Institutional Services department of William O'Neil + Company, Inc., there was never any shortage of banter and debate about the market and leading stocks. As well, there was always somebody coming up with some new "concept" to which they had given a clever name. During 1999 the moves in stocks were so crazy that we would give some of their strange, short consolidations as they dashed to the upside descriptive names like Double-Flying-Eagle or IPO U-Turn. The only one that survived was the one that Michael Lowrey, an institutional salesman at O'Neil, first dubbed the IPO U-Turn. Two of the biggest IPO U-turn set-ups at the time were eBay, Inc. (EBAY) in October of 1998 and Ubid, Inc. (UBID) in December of 1998. The infamous IPO U-Turn was not only a first-stage base set-up that we observed during that dot-com bubble period, but would also serve as a later-stage base as hot Internet IPOs would run up rapidly, then correct very sharply before making a sharp U-turn right back up to new highs. In August of 2004, another Internet IPO, Google, Inc. (GOOG) formed its own first-stage IPO U-turn (Figure 9.40).
Figure 9.40. Google, Inc. (GOOG) weekly chart, 2004–2005.: Chart courtesy of eSignal, Copyright 2010
When we first mentioned GOOG's IPO U-Turn formation to Bill O'Neil, he wasn't buying it. However, once it was explained in terms he could understand, he took another look at it. The primary characteristic of an IPO U-Turn is not the "U" in the formation, it is actually the "invisible flag pole" that forms when the stock opens up for trading on the first day at a price far above its initial offering price. In GOOG's case, it had come public via a Dutch auction, which priced the stock's initial public offering at $85. GOOG opened up for trading on August 19, 2004, at exactly $100 a share, pulled down a couple of bucks, and closed at 100.33. Over the next two days, the stock ran up to an intraday high of 113.48 before pulling back down to the 100 price level and then pulled a U-turn back to the upside. By including the "invisible flag pole" that GOOG formed by opening up at $100 a share on its first day of trading, a full 15 points higher than its $85 a share IPO price, we can see that GOOG's pattern was actually a "high, tight flag." In this manner, one of the primary requirements of an IPO U-turn formation is that it should generally form well above the stock's initial public offering price.
9.4.3. Institutional Sponsorship Is Key
Another characteristic of GOOG at the time that also influenced our decision to purchase the stock at around the $111–$112 price level, initially, was the fact that a 13-D filing had been released and was reported by news service Briefing.com as indicating that Fidelity Management & Research had taken a 13 percent position in GOOG stock. In most cases, when a large mutual fund known to have outstanding research capabilities buys a stock like GOOG in such size, they are not planning to flip the IPO, but are in fact accumulating stock with the idea of holding it for three to five years. This made GOOG a "big stock" right out of the gate, and its IPO U-Turn base worked quite well.
9.4.4. The Short Stroke
Occasionally we might witness the genesis of a new base pattern, such as was the case with the "short-stroke" formation, although we didn't know it at the time. Three days before Christmas, December 22, 2003, O'Neil called to ask what we think about Research in Motion (RIMM). We noted that the base was very tight, and that the company would be reporting earnings that day after the close. We debated whether to put the stock on our institutional stock ideas list. We ourselves could not buy it for 48 hours once it was put on the ideas list, so that our institutional clients were able to act on the information first. We decided to wait for the earnings to come out that afternoon, and once the news was out that RIMM had handily beaten the estimates, the stock gapped up out of a very tight, six-week flat base, shown in Figure 9.41. This was a very powerful move, but the stock was immediately extended from its buy point by over 20 percent, hence the idea of trying to buy the stock up that far away from a proper buy point carried a great deal of risk with it.
The quandary here was where to jump on what we thought was likely to become a very hot-moving stock. The next week, the stock traded in a very tight price range, as Figure 9.41 shows, with the stock closing at the peak of the weekly range as trading volume dried up precipitously. This sort of action indicated that the story here was simply too big from a fundamental perspective, and institutions were not interested in selling any RIMM stock that they may have held. This stinginess with the stock shows up in the very tight weekly range, the close right at the weekly peak, and the lack of trading volume as sellers fail to show up. Later this chart pattern began to appear as the "short-stroke" formation in O'Neil's workshops. The term "short-stroke" comes from the musical notation that indicates where the bow of a string instrument is to briefly touch the strings in a short stroke. This makes complete sense when you understand that O'Neil was a trumpet player in his youth and likely played in orchestras with string instruments. Statistically, it is not clear whether this pattern has any real validity, but it is something of a corollary to a high, tight flag formation where the flag is a single, very tight and "short" one-week flag.
Figure 9.41. Research in Motion (RIMM) weekly chart, 2003–2004.: Chart courtesy of eSignal, Copyright 2010
9.4.5. Labels Can Be Misleading; Context Is Key
One of the drawbacks to trying to label every type of base formation is that it puts one in the position of assessing bases or chart formations on the basis of their "fit" to these base "templates." In the process, many of the chart phenomena that O'Neil goes to great lengths to measure, label, and categorize are easily misinterpreted. For example, in October of 2004 we were playing very heavily in Apple, Inc. (AAPL), which was riding a wave of iPod sales. On October 15th, Apple announced earnings and the stock gapped up through the 20 price level on the chart in Figure 9.42. The huge-volume breakaway gap to the upside had many thinking that the stock was putting in a "climax top," failing to understand that the position of the stock within its overall move and the market cycle is critical in determining when a stock is making a climactic topping move. Because the general market had only begun a new rally phase by follow-through to the upside in August of 2004, Apple's breakout in the last week of August was a breakout from a first-stage base. During their price moves, stocks will generally put in three to four later-stage bases after initially breaking out at or around the same time as the general market also begins an upside rally phase. The first breakout that coincides with the market's recovery and follow-through to the upside is the first-stage base. Climax tops generally occur after a third-stage or fourth-stage or later base is formed in a stock's overall price run.
Figure 9.42. Apple, Inc. (AAPL) weekly chart, 2004–2005.: Chart courtesy of eSignal, Copyright 2010
The interesting thing about this was that at the same time a lot of individuals at the firm were thinking Apple was putting in a climax top, we were buying the stock aggressively for our own accounts. We had actually bought most of our position on the earnings gap-up on October 16, 2004, recognizing that the "line of least resistance" had been broken (Figure 9.43). In Chapter 4, we discussed buying gap-ups like this, because they are buyable, particularly when they have a "volume signature" as Apple did on that day, which is clearly visible on the daily chart in Figure 9.43 as a big upside volume spike.
Figure 9.43. Apple, Inc. (AAPL) daily chart, 2004. The gap up in October was a clearly buyable move.: Chart courtesy of eSignal, Copyright 2010
Apple was the last big, winning stock we played as internal portfolio managers at William O'Neil + Company, Inc., and O'Neil was in it very heavily at the time as well. The last classic maneuver we saw him pull occurred in our last month at the firm, when Apple announced earnings in October 2005. By that time Apple had already had a huge price run, and following this October 2005 earnings announcement the stock began to sell off, finally gapping down and dipping just below its 50-day moving average. O'Neil was shaken out of a lot of Apple stock at that point, and the stock ended up closing just above its 50-day moving average, as we see in Figure 9.44. The next day saw Apple lift off the 50-day moving average on huge volume, something that we know today as a pocket pivot buy point. O'Neil saw this action and, forgetting about whatever stock he had sold at lower prices, immediately began buying the stock back, most likely buying back even more, as this was often his policy. He would often urge us that if we found ourselves shaken out of a big, winning stock, and the stock then showed overwhelming power as it recovered and turned back to the upside, we should consider buying back our entire position and then some. This was the type of decisive and flexible reaction that O'Neil displayed which is very uncommon among average investors. O'Neil was not price-sensitive; he did not care what price he had sold a stock at and whether he was buying it back a few points higher. The point in buying the stock is to catch the "fat" part of the price move—whether one bought it a few points higher or lower, or sold it lower before buying it higher, is simply not a factor. The key is whether the stock is signaling that the line of least resistance has been broken, and a big move is now starting.
Figure 9.44. Apple, Inc. (AAPL) daily chart, 2005.: Chart courtesy of eSignal, Copyright 2010
Buying back Apple in this manner is typical of O'Neil's aggressive style in correcting mistakes in a decisive manner that quickly compensates for the prior mistake of being shaken out of the stock. As Figure 9.44 clearly shows, had O'Neil worried about having to buy Apple stock a few points higher than where he was shaken out around the 50-day moving average he would have missed the parabolic move in Apple that ensued. Notice also that AAPL gave investors two pocket pivots.