A New Technical Indicator and its Probability of Being Useful Ian Aberdeen Retired agricultural scientist Fellow of the Australian Institute of Agricultural Science 30+ years share trading Creator of The A-Line Market Index Iain McLean BSc Eng GradDipPsych Fenestration Engineer UK, EU & USA
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
A New Technical Indicator and its Probability of Being Useful
Ian Aberdeen Retired agricultural scientist
Fellow of the Australian Institute of Agricultural Science 30+ years share trading
Creator of The A-Line Market Index
Iain McLean BSc Eng GradDipPsych
Fenestration Engineer UK, EU & USA
The A-Line Market Index The A-Line Market Index is a leading indicator that predicts trend developments and reversals in the Australian stock market. A proprietary algorithm creates a quantified reading of market mood. The resultant data is used to create the A-Line Market Index.
The A-Line Market Index
The Scientific Method The classic method originated with Aristotle and was refined by Galileo before Rene Descartes established the modern method which allows for heuristics to harness chance, which in turn leads to discovery. Heuristics Definition:
“Exploratory problem-solving techniques that utilise self-education techniques (as the evaluation of feedback) to improve performance.” -Oxford Dictionary
Probability What is probability? Definition:
“The extent to which an event is likely to occur, measured by the ratio of the favourable cases to the whole number of cases possible.” -Oxford Dictionary
XAO & The A-Line
The A-Line Newsletter
The Signal & The Noise “Most statistical models are built on the notion that there are independent and dependant variables, inputs and outputs, and they can be kept pretty much separate from one another. When it comes to the economy, they are all lumped together in one hot mess.”
-Nate Silver
Sir Ronald A. Fisher FRS
1919 Began using statistics for agricultural research & design.
1925 Pub: Statistical Methods for Research Workers.
1929 Fellow of The Royal Society.
1933 Galton Professor of Eugenics at University College, London.
1957 Emigrated to Australia. Senior Research Fellow at CSIRO.
(1890-1962)
Experimental Design Fisher pioneered the use of the term ‘variance’ in statistics & experimental design to eliminate variance.
His work gave the modern scientific method, developed in the 17th century, a new level of power in that experiments were mathematically analysed and could be peer-reviewed.
Statistical Significance When Fisher developed ANOVA he created the notion of a 5% probability level commonly called the alpha (α) level - the probability of making a false positive decision. In essence: If we run an ANOVA and find <5% probability that the result is attributable to the null hypothesis (H0) we describe the result as statistically significant and supporting of our hypothesis (H1).
Dr Helen Turner AO OBE (1908 -1995)
Australian geneticist – sheep breeding & wool production. 1931 Clerical then Technical Officer / Statistician at CSIRO.
1956 Senior Research Scientist (genetics) at CSIRO.
Thomas Bayes (1701-1761)
English statistician, philosopher & Presbyterian minister. 1736 Pub: Introduction to the Doctrine of Fluxions defending Newton’s calculus. 1743 Fellow of the Royal Society.
1763 Pub: An Essay towards solving a Problem in the Doctorine of Chances.
Bayesian probability is used today in science, engineering, law, psychology and medicine to strengthen the case as more evidence is acquired. We learn about the universe through approximation, getting closer to the truth as we get more evidence.
Bayes’ Theorem
Does The A-Line leave you holding shares when the XAO is in downtrend? Our instinct is ‘YES’ because: a) It predicted the GFC and stayed depressed throughout the following dead cat bounce. and b) It became prosperous prior to the XAO upturn in 2009.
Research Question
Gather Data XAO & A-Line weekly charts. 5 & 10 period EMA’s crossings to show trends. Hypothesis Test Support / refute the hypothesis: The A-Line doe not leave you holding shares in a bear market.
Method
Samples & Populations A normally-distributed population
Samples & Populations A normally-distributed population & sample. Inferences on sample will reflect the population.
Samples & Populations Normally-distributed population & biased samples Inferences on biased samples will skew results
The Data Weekly charts: Trends confirmed with 5 & 30 period EMA’s
This is our population – all the available data.
What hypothesis test is most suitable for the A-Line data?
Data structure = correlational Data type = quantitative discrete
Measurement level = ordinal
Pearson’s r: Data type will only give Phi Coefficient Phi-Coefficient: Already verified by visual check on data Spearman’s rho: Already verified by visual check on data Bayes’ Theorem
Hypothesis Testing
Takes a prior probability and re-evaluates it using new evidence to provide a revised probability. “Bayes’ theorem is to the theory of probability what Pythagoras’s theorem is to geometry.”
-Sir Harold Jeffreys FRS
Bayes’ Theorem
Bayesian Convergence
Extended form
Pr = Revised / Posterior Probability = Prior Probability. 5% - Very low. Initial estimate of how likely an A-Line event predicts a market trend reversal. = Probability hypothesis (H1) is true.
50% - Boolean. ‘Yes’ or ‘no’. The A-Line keeps us in or takes us out of the market.
= Probability null hypothesis (H0) is true. 20% - Very high. Raw data or data entry error.
Bayes’ Theorem
Event Definition An event is when the A-Line signals a depression and moves to take profit when the XAO turns down: From the chart above we see there is a positive relationship but we want to quantify the probability of that relationship continuing to protect our capital.
Starting out with a very low expectation (5%) we return a probability of 11.63% doubling our expectations for future events being profitable.
It is not statistically significant but we can use this revised probability as our prior for the next event.
First Event
Within 6 events the probability of the A-Line protecting capital is 92.43%.
6-Event Analysis
XAO Retracements There are 6 events when both XAO & A-Line turn down. There are also an additional 2 depressions in the A-Line during significant XAO corrections in 2004 & 2006.
XAO Retracements
All Ordinaries (2006) with 5 & 30 EMA’s
By the 7th event the probability of the A-Line protecting capital is 96.64% - statistically significant under Fisher’s 5% alpha level criterion.
8-Event Analysis
Changing the values to: = 5% initial estimate probability = 33% chance H1 is true = 35% chance H0 is true There is no statistical significance.
Alternatively…
High Probability Trades Whenever a high probability strategy is presented always look at the back tested data. If none is provided then consider that strategy as untested.
“If you can’t make money from unleveraged intermediate term trades using weekly and daily data, you are going to get cleaned out fast making short-term trades with highly leveraged markets.”
-Robert Miner High Probability Trading Strategies
Conclusion Using Bayesian inference there is a statistically significant positive relationship between A-Line signals and Australian stock market trend reversals
There is a high probability that A-Line will protect capital by signalling future depressions in the Australian stock market.
No indicator can predict the future with 100% accuracy but data analysis can isolate trends for profit.
Knowing when the trend begins is not as important as knowing when the trend is ending.