pixel
DIY Trader
- Joined
- 3 February 2010
- Posts
- 5,359
- Reactions
- 345
The following is the transcript of an email discussion I had - spread over several sessions - with a Trader, who had been "dabbling" in some T/A, but wanted to get his head around practical applications of a software package. In the transcript, which I also prepared for posting outside this Forum, I refer to "the software". For those "in the know": I'm talking about the Market Analyser.
Q:
I have read a technical analysis book but didn't have faith in it. I am open to changing my mind. I am looking at holding from a week to 3 months if I get the balls to do this type of training with confidence and accuracy. Do you find the software useful for long term investment as well?
A:
For a computer-illiterate person or a newbie that lacks basic T/A know-how, the learning curve is probably daunting. When I bought the package, the deal included a six-week course in Technical Analysis and a number of follow-up seminars; I know of several people, who attended those sessions several times. But there were others, who "knew it all", had made up their mind, and still make the most basic mistakes interpreting a chart.
Q:
A 34 day swing Horizon isn't very long in my view so what do you call "Long"? I think Warren Buffet views at least 5 years!
A:
When you wonder about 34 days being a "long" time, let me explain the rationale in more detail:
I am looking for a 34-day event that establishes a likely bottom reversal in a time horizon that may take up to a month to develop. We start with a significant Low, say the lowest price in a month; then we wait for primary resistance to establish itself within a few days or couple of weeks. The primary resistance pushes the trading back to a support level, which I prefer to be at a Fibonacci level of 50 or 68.1%. Taking that support as the first Higher Low, we wait (in theory) for the primary resistance to be broken, assuring us of a Higher Low - Higher High sequence within about a month or less. Having back-tested "a few" configurations of that kind, I've found a high correlation between certain price-volume-momentum moves and a successful breakout.
In some cases - especially in a trending Market - the ensuing rise can go on for months or years. In such cases I'm happy to sit it out, even switch to weekly charts with different parameter settings and higher tolerances. But as you may appreciate, if I see a "reverse pattern" at the top, I'd feel silly to insist on holding on to a stock that gives a diminishing return.
In addition, I have come up with a concept for a volatility-based trailing stop that uses the number of average days' price moves "against me". The way it works: I accept a reduction of my asset's theoretical value (i.e. overall: my net worth including paper profits) by, say, 1 or 2 days adverse moves in my disfavour. As an example, let's say ANZ had been trading around $20 with an average trading range of 50c per day. If I had a 1-day "risk" in mind, I'd set my stop at $19.50.
(It's a little more involved, smoothing peaks and troughs and intraday swings, but that's the gist.)
Q:
A very subjective question: How much should I invest per trade? How much you are prepared to lose? Or can you put a $figure on it? Let's assume a portfolio of $100,000. I was thinking about 10%, but would have done a lot of back testing before entering the market.
A:
Your question about sizing a trade is really a difficult one. Scaled to a $100K portfolio, I'd say a blue chip position could be built to about $10K per share; lesser lights half that or, if "speccie rules" apply, possibly only $3000. If it's running well, one can always average up; if it moves against us, caution limits the size of the loss.
Q:
http://rettmer.com.au/TrinityHome/Trinity/index.htm has exactly what I need to ground my feet and learn a different aspect of trading. I am a bit shocked at how much there is, but one step at a time. At least now I can go back over my Technical Analysis books with a different slant and more confidence.
A:
As for being "shocked at how much there is" to learn: You're only scratching the surface of what I had to do in order to condense hundreds of indicators and concepts into one "Trinity" that avoids false readings due to double-counting.
(Just as an aside.
Some people take 2 or 3 - sometimes even more - indicators that measure the same factor, e.g. price momentum; then they jump on board because they get 3 or more ticks from all those indicators, and that, they think, outweighs the one caveat, e.g. falling volume. Whereas with Trinity, I want all three: price direction, volume, and momentum to tick their boxes before I allow a buy order.)
Q:
What I cannot get my head around is that you are applying strict mathematical formulae on the tendencies of fragile, determined, emotional, greedy, cautious, wise, foolish humans who enter and leave the market for emotional and external financial reasons, as well as trying to make a profit. Individuals do not behave mathematically and are raw illogical. Humans are not robots.
The market is imperfect as not all participants have the same ability to use common data, sense and knowledge. In addition, participants misread and misunderstand market conditions as well. What am I missing here?
A:
You've hit the nail on the head: People are not robots. People are not acting logically either.
That's the reason why I don't calculate how much a company or a share "is really fundamentally worth", but I let the chart, via statistical analysis, tell me how all people in aggregate perceive this company that they take an interest in.
Whether I take a short-term view of mood swings that last minutes, or consider macro-economic mood swings of "The Market": as long as I find a sufficient number of quasi-independent events, I can apply statistical analysis to make an educated guess about the relative odds of each of three outcomes. These outcomes are of course
(1) people love it and buy,
(2) people want to get rid of it and sell, or
(3) people don't really know and wait.
Under the right conditions, there comes a point in a share's life cycle when the odds for (1) are so much higher than for the other two that I can safely buy. If I'm wrong and an outcome with lesser probability prevails after all, I activate Plan B: Stop out with no regret.
Assuming I buy on odds of 80%, I may have to stop out of one trade in five, but will reap a profit in the other 4. Which means, even if my average profit only equals the average loss, I can expect an overall net winning margin of 3 in 5 (the one loss cancels the 4th win.)
As to your assumption that I need to investigate each stock over many years before I can analyse its potential: That is not the case. I am using statistical analysis as a stock-independent tool that deals with "crowd behaviour". I'm almost tempted to call it "mass psychology". And given similar stimuli that make people buy stock A, their reaction will be sufficiently similar when we deal with stock B.
In that regard, T/A is far more economical in its application than studying the fundamentals of a myriad stocks. With the software, I can run the same analysing script over each and every stock in any of the supported world markets, and the resulting table, if any, will show me exactly those shares that at this point in time match the conditions that I've been filtering for.
As I have also written corresponding scripts that display the matching conditions for a particular stock on its chart, I can look back into a particular stock's history and check, how well it has met the underlying expectation.
Let me explain by an example:
Let's say I run an analysing script that checks for stocks that meet the classic Trinity conditions: Price decline, followed by consolidation, followed by a price rise sufficient to make the momentum turn positive. Let's say I find 3 stocks, for which the setup is true today. I then load the matching template into a chart and bring up each of the three stocks, looking back in history whether a similar setup has occurred before, and if so, whether the price has risen afterwards. If the answer is affirmative, I can safely assume that this stock will likely give me also this time a profit. If the answer is negative or so-so, I either stand aside, or take a smaller initial position, or drop it from the watchlist altogether.
Once in a trade, I simply apply a trailing stop formula, based on the same criteria and time frame that I used when selecting the entry price. If the profit builds up and I see a precedent that appears to increase my profit expectation, I may average up, or I may take part profit. If a situation evolves, where I have to be concerned that my (paper) profit may be eroded, I don't wait for the trailing stop to be hit, but take the best exit on offer, while it is on offer. (I can always buy back in if I sold too early; but I can't sell at the high point if the price continues to fall.)
Q:
I have read a technical analysis book but didn't have faith in it. I am open to changing my mind. I am looking at holding from a week to 3 months if I get the balls to do this type of training with confidence and accuracy. Do you find the software useful for long term investment as well?
A:
For a computer-illiterate person or a newbie that lacks basic T/A know-how, the learning curve is probably daunting. When I bought the package, the deal included a six-week course in Technical Analysis and a number of follow-up seminars; I know of several people, who attended those sessions several times. But there were others, who "knew it all", had made up their mind, and still make the most basic mistakes interpreting a chart.
Q:
A 34 day swing Horizon isn't very long in my view so what do you call "Long"? I think Warren Buffet views at least 5 years!
A:
When you wonder about 34 days being a "long" time, let me explain the rationale in more detail:
I am looking for a 34-day event that establishes a likely bottom reversal in a time horizon that may take up to a month to develop. We start with a significant Low, say the lowest price in a month; then we wait for primary resistance to establish itself within a few days or couple of weeks. The primary resistance pushes the trading back to a support level, which I prefer to be at a Fibonacci level of 50 or 68.1%. Taking that support as the first Higher Low, we wait (in theory) for the primary resistance to be broken, assuring us of a Higher Low - Higher High sequence within about a month or less. Having back-tested "a few" configurations of that kind, I've found a high correlation between certain price-volume-momentum moves and a successful breakout.
In some cases - especially in a trending Market - the ensuing rise can go on for months or years. In such cases I'm happy to sit it out, even switch to weekly charts with different parameter settings and higher tolerances. But as you may appreciate, if I see a "reverse pattern" at the top, I'd feel silly to insist on holding on to a stock that gives a diminishing return.
In addition, I have come up with a concept for a volatility-based trailing stop that uses the number of average days' price moves "against me". The way it works: I accept a reduction of my asset's theoretical value (i.e. overall: my net worth including paper profits) by, say, 1 or 2 days adverse moves in my disfavour. As an example, let's say ANZ had been trading around $20 with an average trading range of 50c per day. If I had a 1-day "risk" in mind, I'd set my stop at $19.50.
(It's a little more involved, smoothing peaks and troughs and intraday swings, but that's the gist.)
Q:
A very subjective question: How much should I invest per trade? How much you are prepared to lose? Or can you put a $figure on it? Let's assume a portfolio of $100,000. I was thinking about 10%, but would have done a lot of back testing before entering the market.
A:
Your question about sizing a trade is really a difficult one. Scaled to a $100K portfolio, I'd say a blue chip position could be built to about $10K per share; lesser lights half that or, if "speccie rules" apply, possibly only $3000. If it's running well, one can always average up; if it moves against us, caution limits the size of the loss.
Q:
http://rettmer.com.au/TrinityHome/Trinity/index.htm has exactly what I need to ground my feet and learn a different aspect of trading. I am a bit shocked at how much there is, but one step at a time. At least now I can go back over my Technical Analysis books with a different slant and more confidence.
A:
As for being "shocked at how much there is" to learn: You're only scratching the surface of what I had to do in order to condense hundreds of indicators and concepts into one "Trinity" that avoids false readings due to double-counting.
(Just as an aside.
Some people take 2 or 3 - sometimes even more - indicators that measure the same factor, e.g. price momentum; then they jump on board because they get 3 or more ticks from all those indicators, and that, they think, outweighs the one caveat, e.g. falling volume. Whereas with Trinity, I want all three: price direction, volume, and momentum to tick their boxes before I allow a buy order.)
Q:
What I cannot get my head around is that you are applying strict mathematical formulae on the tendencies of fragile, determined, emotional, greedy, cautious, wise, foolish humans who enter and leave the market for emotional and external financial reasons, as well as trying to make a profit. Individuals do not behave mathematically and are raw illogical. Humans are not robots.
The market is imperfect as not all participants have the same ability to use common data, sense and knowledge. In addition, participants misread and misunderstand market conditions as well. What am I missing here?
A:
You've hit the nail on the head: People are not robots. People are not acting logically either.
That's the reason why I don't calculate how much a company or a share "is really fundamentally worth", but I let the chart, via statistical analysis, tell me how all people in aggregate perceive this company that they take an interest in.
Whether I take a short-term view of mood swings that last minutes, or consider macro-economic mood swings of "The Market": as long as I find a sufficient number of quasi-independent events, I can apply statistical analysis to make an educated guess about the relative odds of each of three outcomes. These outcomes are of course
(1) people love it and buy,
(2) people want to get rid of it and sell, or
(3) people don't really know and wait.
Under the right conditions, there comes a point in a share's life cycle when the odds for (1) are so much higher than for the other two that I can safely buy. If I'm wrong and an outcome with lesser probability prevails after all, I activate Plan B: Stop out with no regret.
Assuming I buy on odds of 80%, I may have to stop out of one trade in five, but will reap a profit in the other 4. Which means, even if my average profit only equals the average loss, I can expect an overall net winning margin of 3 in 5 (the one loss cancels the 4th win.)
As to your assumption that I need to investigate each stock over many years before I can analyse its potential: That is not the case. I am using statistical analysis as a stock-independent tool that deals with "crowd behaviour". I'm almost tempted to call it "mass psychology". And given similar stimuli that make people buy stock A, their reaction will be sufficiently similar when we deal with stock B.
In that regard, T/A is far more economical in its application than studying the fundamentals of a myriad stocks. With the software, I can run the same analysing script over each and every stock in any of the supported world markets, and the resulting table, if any, will show me exactly those shares that at this point in time match the conditions that I've been filtering for.
As I have also written corresponding scripts that display the matching conditions for a particular stock on its chart, I can look back into a particular stock's history and check, how well it has met the underlying expectation.
Let me explain by an example:
Let's say I run an analysing script that checks for stocks that meet the classic Trinity conditions: Price decline, followed by consolidation, followed by a price rise sufficient to make the momentum turn positive. Let's say I find 3 stocks, for which the setup is true today. I then load the matching template into a chart and bring up each of the three stocks, looking back in history whether a similar setup has occurred before, and if so, whether the price has risen afterwards. If the answer is affirmative, I can safely assume that this stock will likely give me also this time a profit. If the answer is negative or so-so, I either stand aside, or take a smaller initial position, or drop it from the watchlist altogether.
Once in a trade, I simply apply a trailing stop formula, based on the same criteria and time frame that I used when selecting the entry price. If the profit builds up and I see a precedent that appears to increase my profit expectation, I may average up, or I may take part profit. If a situation evolves, where I have to be concerned that my (paper) profit may be eroded, I don't wait for the trailing stop to be hit, but take the best exit on offer, while it is on offer. (I can always buy back in if I sold too early; but I can't sell at the high point if the price continues to fall.)