Between my 4 charting platforms I must have over 1000 technical indicators available and the ability to build another 1000 or more, sometimes we get lost in these indicators and forget that one of the best indications of a change in the market is a change in the character of the market. Mass sentiment is a good reflection, but in days like this when news is so abundant and fast moving, sentiment can change literally by the moment.
I've found the same to be true with technical indicators, especially in times of chaos, they are chaotic.
In fact, last Friday I used 3C and a simple, but rare bearish candlestick pattern (a visual representation of sentiment used for hundreds of years by the Japanese) to make a call that the markets would likely fall on Monday, that is what happened, even though I spent most of the weekend looking at traditional indicators and had I gone by them, I would have never said the market would fall Monday. I have found in times like this, taking the long term view is about the only thing that really works, and indicators can work as well, but they need to be set to a longer term view.
My Trend Channel is not like other channels that are set to fixed settings, it automatically adjusts, widens and narrows according to the stock's recent volatility. I use it as a longer term stop, especially for trends, but it is also helpful in identifying changes in character. Above on a daily setting, it first showed a short position stop out in white and suggested the trend would change to up, after a volatile move outside of the channel, which I consider to be bearish in most cases, even though it looks bullish, the channel then gave a sell signal in red, weeks before the market actually crashed. Then it gave the all clear with another white (cover signal) suggesting the downtrend was done for now, which it was. It recently gave a sell signal on Tuesday of this week. Additional changes in character include this week is the first week of the last 5 weeks that we have seen a weekly close at a loss, the last 4 weeks all posted weekly gains, suggesting a change of character is underway. This is more or a broad sword then a scalpel, but it is effective for it's purpose. It will never take you out at the top of a long trend or the bottom of a short trend, but it will allow you to catch the 60-80% of the trend in the middle, which most people simply use arbitrary stops and leave a lot of profit on the table. After all, if you have a 6 month uptrend, fear may cause you to take profit in the first few weeks, the trend channel makes the decision for you based on the stock's own changes of character and allows you to have an objective stop.
Here we see price in a bear flag this week,
just like I showed you in yesterday's post, we saw a similar bear flag in 2008 at the same area. Right now price is caught between resistance of the major Head and Shoulder top's neckline resistance at the red arrow and recent consolidation support at the white arrow. There is almost always short term volatility upon the break or breakout of any important resistance or support level, so this area has the potential for volatility.
This is the S&P candlestick pattern, a bearish reversal Harami Cross, which is quite rare to see, that led me to say (a week ago from this Friday) that I thought the market would be down on Monday.
This Doji in the SPY on Tuesday shows a loss of momentum and indecision and led to to say that I thought we would see a bounce from there (this Tuesday)- the bounce formed a bear flag which Is showed you appeared in 2008 at almost the exact same time and level in an extremely similar pattern-see my last post on the market
"Follow Up 2008/2011 Market Action"
The standard technical tools, including MACD would not have suggested any type of likelihood of a Monday/Tuesday decline. The 50 day average also shows what will likely be an important support level in blue.
This 3C shows the recent weakness in the bear flag this week.
The 5 min chart shows the Friday area that looked very much like a top area, the second divergence is the bear flag from this week. The leading divergence is hitting new lows right now on this hart.
I have over the years, used StockFinder's BackScanner function to test many of the most well known technical set ups and signals and have found that they have almost no edge over the market, they are just cherry picked examples put in the newest Technical Analysis book and rarely, if ever, have back testing data to support their use. One such indication is the "Death Cross" or the "Golden Cross" which is the 50 day moving average crossing below or above (respectively) the 200 day moving average. Backtesting this strategy has shown no usefulness whatsoever. However, one thing I have observed at nearly every major and minor bull market top (changing to a bear market) is the 50 day crossing down below the 200 day, but the 200 day must be close to or already be pointing slightly down. Above is the 2008 top showing this feature and recently around late August 2011 we have seen it again. As I said above, as with almost any breach or breakout of an important support or resistance level, there is volatility nearby and you can see the bounces to the 200 day moving average in both cases after the cross has occurred (at the yellow arrows).
Looking at more tops, here is the S&P breaking down at the 2000 tech top, the direction of the 200 day average is very slight, but it has a slight downward bias to it.
The Dow-30 1981 top shows the same.
In 1976 there was a cross below the 200, but notice the trajectory of the 200 was up, in 1977 there was another cross, this time the trajectory of the 200 was slightly down and there were several tests of the 200.
In 1971 there was a false cross and 1973 a real one with the 200 starting to move down as well as several volatile tests of the 200. You don't have to use a microscope to determine the 200's direction, usually the false signals it will be clearly facing up while the real signals it will be facing slightly down or at least flat and confirm the down move shortly thereafter.
Here's the S&P 1969 top with a false and real signal, note the trajectory of the 200 in each (white and red boxes).
The Dow in 1946 was quite clear, it led to a sideways/lateral market for years, note the false signals during that lateral market.
In 1938, we have another perfect example of the false cross and the real cross.
Finally in 1929, the cross was quite obvious as was the bear market rally to the 200 day.
Determining the exact direction of the 200-day average can sometimes be challenging and leave the door open to subjective/emotional interpretations, however, I have found that is you add a 100 day moving average to the 200 day moving average (not to price), the 200 day average will often cross below the 100 day average in an environment that may be otherwise difficult to determine the direction of the 200 day average.
Here's a close up of the recent cross below the 100 day average and I have created a kind of MACD of the two averages to show the actual cross below. As to the effectiveness of this strategy...
Here's a long term view of the results, it called every top and bottom, removing the uncertainty of bear market rallies or bull market declines. The 200 top is called, the 2003 bottom is called, the 2008 top is called, the 2009 bottom and recently this week another signal of a 2011 top.
Here is my Trend Channel's performance during some of the most well known bull market tops in the last century. A setting of 5 days handles shorter term trends, a 9 day setting handles longer term trends and a monthly setting gives the trend the most leeway possible, a break below a monthly Trend Channel Setting (as far back as the last century, has always ended up being a market top.
Here we see the widest setting of the monthly trend channel, it calls the 2008 top; in the white box there was a top-ish looking area after QE1 ended, a shorter timeframe may have called it a top, but the monthly timeframe allows for this period to remain as part of the longer term uptrend, only recently in August did we see a break of the monthly trend channel. Note also how well it hold the uptrend from 2003-2008-not a single false signal.
Here's the NASDQ 200 Tech bubble top , both the uptrend before and the downtrend after are held well within the channel. While a shorter timeframe could have been used to have an earlier signal, the monthly timeframe is fairly definitive.
The minor S&P 1976 top, even though there is volatility after the signal is given, the top is still assured and this is on a 5 day basis.
The S&P 1969 and 1973 tops as well as the 1970 bottom, again, shorter timeframes, especially for the bottom would have produced earlier signals, but the monthly timeframe again throughout the century has been definitive.
Dow-30 1946 top leading to a lateral market.
Dow 1937 top, also look how well the channel held the preceding uptrend and the proceeding downtrend.
The 1929 top on a 9-day basis, it holds the uptrend for several years before calling a top and all of the downtrend.
On a monthly setting it holds 9 years of uptrend without a single false signal, the top is signalled a little later, but with the change in volatility between 1922-1928 vs 1928-1929, the channel can and should have been set to a shorter timeframe as a change was underway clearly indicated by the price trend, thus giving us the earliest signal possible.
In essence, the length or setting of the Trend channel depends on what kind of trader you are and the market conditions. I have used 5 min version for day trading, after several years of uptrend, I would choose a longer setting as a clear trend is in effect. For swing traders a 60 min to 1 day setting may be appropriate, the important thing is that the Channel adjusts its width as price volatility changes, so it is unique to each individual stock.
No comments:
Post a Comment