Friday, December 30, 2011
Thursday, December 29, 2011
Wednesday, December 28, 2011
As for the Covestor Model Portfolio, I'm quite loaded with various long positions. Some of them (GCI, FTK) acted weak during the recent rally, so I will probably close them and wait for better opportunities. I still maintain some shorts (HPQ, GMCR, BBY, SLV). Best Buy is the stock I will possibly cover soon: there was no follow through after the earnings miss and prices seem to bounce from an important support level.
Don't take my long exposure as a bullish sign. It is window dressing season and Europe is in hibernation mode. I can't believe this topic is solved, expecting negative headlines to continue soon.
Thursday, December 22, 2011
I added another long position yesterday, KeyCorp (NYSE: KEY). I like the relative strength of US regional banks. The market could run into major resistance soon around 1260, so I wouldn't add new long positions at this point and possibly shift into profit taking mode.
"I know you focus on technical analysis only, but wondering about the MA thesis. Do you view it as a Financial play or a Technology play? If Financial, do you worry about BAC going down and implications on credit and debit card volume? If Tech, do you worry about ORCL-like results reflecting slower growth?"
My short answer is: no, I don't worry.
Actually, it doesn't matter if I worry about anything related to the markets. What matters is to figure out if the market will worry about the worries if you will. The term "worry" should not be part of the vocabulary of a short-term investor/trader anyways. I cannot trade based on emotions. It can only work based on a whatsoever analysis that you know gives you an edge in the market. Since I'm very short-term oriented, fundamentals don't matter anyways, but let's assume they would: I guarantee you that hundreds of bank/hedge fund analysts thought about ORCL earnings yesterday and what they mean for the economic outlook. Some, maybe many, of these analysts are much smarter than at least me. Plus: these institutions have access to much better information than I do. The result of these analyses is the current stock price (if you believe that markets are efficient, if you don't, fundamental analysis doesn't matter anyways). I cannot have an edge over these institutions by performing a better fundamental analysis or have a smarter thesis.
Another point: yes, ORCL earnings were bad, but I guarantee you that this didn't come as a surprise to some of the big guys. Just check how ORCL has been underperforming (even the Tech sector) since November. There were some massive insider sales in the last months, so chances are whoever sold yesterday, was the last one to sell and that's usually the retail crowd.
With respect to MA: look at the long term correlation with BAC: both are inversely correlated, so obviously. there hasn't been any implication for the last 12 months at least. Can that change? Of course. But the key to technical analysis is to use it reactively and not in a predictive way, so if correlations change, there will be enough time to sell or even go short. I doubt this will happen in the next two weeks, because that's the time frame of my trade.
Wednesday, December 21, 2011
Going forward, we need to see if a new uptrend can shape up. The action yesterday was just a pop higher. Stocks now need to record a higher low in order to establish a trend. 1225 should offer support.
I did a couple of trades in the Covestor Model Portfolio and bought some momentum names. One example: Mastercard, which I intend to hold until $395:
Tuesday, December 20, 2011
Closing or at least scaling out of short positions seems to be a prudent move. On the other side, I had to close a long trade in DY yesterday with a small loss because the stock hit my stop price. My shorts in XLF, BBY, GMCR, FXE and HPQ are doing well, but as I mentioned before I'll need to skim exposure this week. Don't forget that US stocks are driven by the Euro and there is a record amount of open short positions in the market. So any rumor could shoot the currency to the moon , which would also be bullish for the S&P.
Monday, December 19, 2011
Sunday, December 18, 2011
Friday, December 16, 2011
Historically, Platinum has been more expensive than Gold for most of the time. Here is a chart of the long-term picture:
As for December 16, the downtrend is still intact. However, prices are reaching an important level: should the S&P break 1225, the sequence of lower lows and lower hights gets broken and the market could work on a short-term reversal. In that case, a inverse head and shoulder could mark a intraday bottom. One wants to become more offensive and possibly add to longs on any pullback.
Any failure of 1225 would be a good opportunity to add to short positions as indicated yesterday.
Thursday, December 15, 2011
He, let's call him Demetrios, is running an engineering company with ten employees. Yesterday, he received a letter from the Greek tax office and they asked him to pay 50,000 Euros in taxes for some profits his company made five years ago. The problem is: he doesn't have 50,000 Euros right now. Demetrios actually has to let some of his guys go to be able to keep paying the bills.
He also told me about a "new" government initiative to raise some cash by increasing/collecting property tax. What happens now in Greece is that people own real estate and suddenly have to pay taxes on it. OK, they probably should always have paid taxes a while ago, but the issue is that some of them don't have jobs anymore and are struggling to pay any bill.
Conclusion: politicians are dreaming if the think that they can balance their budget by raising taxes in a screwed up economy. Sooner or later, Greece has to default to reduce debt in my opinion. Color me bearish.
The following chart shows the 30-day relative performance of major sectors vs. the S&P 500. While yesterday was a good day for bank stocks, it is difficult to see a larger underlying trend. It'll need more strong days before I'm ready to agree that Financials are out of the woods. What's rather eye-catching is the continued strong relative performance of defensive sectors such as Consumer Staples, Healthcare and Utilities.
Also interesting: relative performance of Energy and Basic Material stocks has been deteriorating for weeks. So yesterday's sell-off shouldn't have come as a surprise:
Disclosure: Covestor Model Portfolio is long SKF.
If we see a one or two percent rally to the declining 5 day moving average, I might consider adding to the short side as long as downside momentum doesn't diverge: keep an eye on MACD. The divergence end of November indicated the upcoming change in trend and aggressive traders would have established long positions. So far, there is no divergence. Stay with the trend until it changes.
Another way to look at MACD is to watch out for extreme readings. In the last three months, the indicator's most negative values turned out to be below -1.5. Currently MACD is at -0.7, so there is potential for further declines. Also 120 seems to be an important area of support, so watch out for divergences at that level:
Wednesday, December 14, 2011
Today's episode: the 200 day moving average. Institutions are watching this indicator, which is why he is so important. Just like in June 2008, prices were touching the declining moving average and sold off afterwards. If equities keep copying the 2008 price action, keep your seat belts fasten. I keep focusing on short positions in the meantime.
The latest price action of the S&P 500 is a good example. There is no trend on the daily chart, so no higher highs/higher lows or lower lows/lower hights:
The 30 min intraday chart looks entirely different with respect to trending conditions. Prices established an uptrend from Nov 26 to December 5, then moved sideways until December 6 and have established a downtrend since then. Very clean and textbooklike action:
The second step is to figure out how to trade the trend. Since markets are currently trending down a swing trader wants to establish short positions at the upper channel boundary and take profits at the lower trend line. Swing trading on an intraday chart would basically mean to not hold positions overnight. It would have meant to go short yesterday morning and cover in the afternoon. However, a swing trader, who operates on daily charts (like me) becomes a trend trader on the intraday charts.
Tuesday, December 13, 2011
A recent UK paper gives a good overview of the topic. While the authors expect computer trading to increase in the next ten years, they also envision that human traders will still play, a role, although a smaller one, in the markets. One key aspect that comes to mind when studying the publication is time frame. It seems like the entire algorithmic trading discussion is centered on very short-term methods, such as high-frequency trading, and recently "news analytics" (a strategy, where algorithms scan the news for headlines and rumors and automatically act upon these information). So if the machines push into ultra-short time frames, humans have to move to lower trading frequencies where they face less computer competition. Intraday traders will probably have a hard time competing because they fight against one of the key strength of a computer: speed.
Computers became quite fast in recent years and academics are discussing when hardware will become as powerful as the human brain.
A quick comparison of processing power and memory: IBM's Watson, who competed in the "Jeopardy" game show this year, can operate 80 trillion floating point operations per second (80 Teraflops). Theoretically, the human brain is much faster with an estimated 100 quadrillion flops (100 Petaflops). The big difference lies in parallelization: the computer can use most of his processing power for a single task, the human is not only running more complex jobs (breathing, eating, typing,...) in parallel, he has also the unique capability to learn while he executes. Computers can't really do that (yet). In terms of memory, computers have been catching up: Watson has 15 terabytes of RAM, while estimates for the brain range from 750 GB - 6 TB.
One of the areas, where computers still have a hard time with is pattern recognition, although progress has been made in that field. I believe that the human trader has to focus on patterns when he wants to outperform machines. Technical analysis can be part of it. Patterns can occur in news flow and the behavior of market participants. Markets always change, so recognizing this change (which is essentially a process of learning) leverages another strength of the human brain.
Here is an example: have you noticed the crazy price action in Silver this year?
This chart is very interesting and in my opinion perfectly shows how traditional technical analysis principles evolved in the light of algorithmic trading. When recognizing these changes, it is still possible to make money with Silver.
The first "technical event" was the vertical run-up in April, marking a classical blow-out top. It was easy to identify but a trader, who was long needed to sell into strength and couldn't wait for the eventual top to occur. In the new age of algorithmic trading, it is more difficult to recognize the final top itself. A technical trader would maybe watch out for some reversal candlestick pattern. In the case of Silver, you never got such a pattern. Instead, prices simply collapsed for three days. Classical technical analysis suggests waiting for a weak rally after the initial decline for a good shorting opportunity. In the age of algorithmic trading, there is no more time to do that, because every trader (computer) is trying to get out at the same time. You can still go short, but can't wait for the rally.
Then there was the second dump in September, when prices declined by 25% in just two days. The same pattern: no time to wait for a weak rally. You were either short or you were not. Technically, it was possible to recognize the weakening price structure days before the decline. In a computer-driven environment you can't chase anything anymore because risk/reward ratios become negative immediatly.
So how can a human trader win against the algos?
1) By lengthen the time frame
2) By focusing analysis on pattern recognition and learning mechanisms
3) By understanding how computer trading is changing traditional patterns
Wednesday, December 7, 2011
The McClellan Oscillator is becoming my favorite tool to evaluate overbought/oversold levels. Extreme readings tend to indicate major turning points (necessary, but not sufficient condition). +/-80 has been a level, where you want to consider to at least close long/short positions. Note how low readings set the stage for the November rally:
Current indicator level is around 20, so markets are far from overbought even despite recent gains. News from Europe later this week will be key for another leg up in US equities.
It seems hard to believe but guess who has been leading the market higher during the last two months: US Homebuilder stocks. Note how ITB, the ETF of that sector, has been outperforming the S&P 500:
Price-volume action has been extremely bullish since higher prices came in on above average volume.
Housing prices are still declining in the US, which can also mean that foreclosures are finally getting sold and bring down inventories. That's a positive for the market. It's important to keep in mind that housing stocks tend to bottom well before housing prices start to flatten out. As usual: the market is a discounting mechanism. Buy the rumor, sell the fact.
As for the Covestor Model Portfolio, I closed my TOL long position before the earnings report. However, I'm looking for getting back into the stock on pullbacks.
Monday, December 5, 2011
Bearish sentiment of Euro/Dollar currency speculators has recently reached extreme levels: Large traders accumulated a very high number of Euro short positions. The last time, Commitment of Traders (COT) data showed these readings was early Summer 2010. The Euro then reversed sharply and rallied from 1,20 to 1,40 in five months. I wouldn't be surprised to see a similar move if Europe can get their act together this week. Downside is probably limited due to the negative sentiment:
Price/volume action has been negative in recent months: higher prices in TLT came in on declining volume. Note how similar divergences let to intermediate term price reversals in recent history (Jun-Sep 2010, Jan - Mar 2011). A classical chartist would recognize a potential "double top" chart pattern in the making.
Last but not least, the latest COT action hints that large traders didn't buy into the recent November bond rally and actually increased their bearish bets: