The Forex Automaton project was launched in April 2008 with the ambitious mission of leveraging certain algorithmic know-how to create a trading signal service useful to institutional and retail forex traders. This report highlights the noteworthy new developments that took place during the project's second year of life, from April 2009 through March 2010. The intent is to help the reader navigate what is becoming a rather complex network of research topics, concepts and results, by providing an overarching logical framework. Links to complete stories are provided.
-
The Second Annual Summary of Forex Automaton Research Progress, April 2010
- Tags:
- frontpage
April 2 2010, 3:06pm | Comments »
-
Distribution of GBP/USD daily high and low during a day
I continue documenting the distribuions of the moments of time daily high and low for the popular forex pairs are achived. Some of the motivations for such type of analysis have been outlined in the first post on the topic. Availability of temporal distributions of volatility during a day (see the one for GBP/USD here) puts this type of study in a new context: significant deviation of shapes between the two distributions can be used to challenge a random walk model to reproduce the distribution of high and low on the basis of the volatility distribution.
- Tags:
- frontpage
March 31 2010, 7:01pm | Comments »
-
Danica v0.5 forecast does not "age" intra-day, is more relevant to periods of high volatility.
So far, positive Pearson correlation coefficients between system forecasts and real forex moves on the day scale have been seen both during the back testing and in the live regime and it looks like a positive expectancy trading system has been indeed created. Among the methods of risk management, we have been focusing our effors on the issues of what forecasts to trade and how to allocate capital to trades. This report is our first research attempt to reduce risk by being selective about when to trade during the day.
- Tags:
- frontpage
March 22 2010, 7:29pm | Comments »
-
Distribution of USD/JPY daily high and low during a day
I continue analyzing the timing of daily high and low for the popular foreign exchanges rates, one exchange rate at a time. Some of the motivations for such type of analysis have been outlined in the first post on the topic.
- Tags:
- frontpage
March 17 2010, 2:03pm | Comments »
-
February performance review for Danica-9am algorithmic system
During the month of February, the second month of real-time documented performance, the system kept running on complete autopilot, with no code upgrades or parameter changes. This document consists of a summary section followed by 14 subsections, dedicated to the individual exchange rates tracked by the system. Those contain color-coded charts of the performance and details pertinent to the specific currency pairs. Usage strategies and effects of various approaches to selecting the forecasts to trade are discussed. For comparison with the previous month, you may want to take a look at January review.
- Tags:
- frontpage
March 3 2010, 12:00pm | Comments »
-
Progress towards algorithmic implementation of Kelly Criterion
The "Kelly Criterion" in quant folklore is based on the exposition and development of Kelly's work done by Edward Thorp. Acknowledging Thorp's contribution, I find the original article by Kelly conceptually more relevant to the realities of algorithmic trading as developed by ForexAutomaton. Our forecasting algorithm, as any forecasting tool, can be very naturally considered a case of the hypothetic communication channel discussed by Kelly, and the related mathematical objects, such as joint and conditional probability density distributions for the communicated "symbols" (forecasts and real quotes), are being already monitored here, despite the fact that the Kelly Criterion for capital allocation to trades remains to be coded.
- Tags:
- frontpage
February 26 2010, 3:49pm | Comments »
-
Intermarket analysis vs markets-in-isolation on day scale.
I am revisiting the issue of the intermarket analysis on the day scale. The conclusion from the previous report on the subject was that, the rest of the algorithm being the same, intermarket analysis gives no advantage on this time scale and simpler analysis of the isolated markets should be preferred. In this report, data on the predictability of high and low are added and a bug related to the estimation of statistical precision of the data is fixed. Nevertheless, the conclusions remain the same.
- Tags:
- frontpage
February 18 2010, 12:24pm | Comments »
-
Day-range strategy is improved by removing profit target. Profit distribution may lose second moment.
I continue developing ways of using Danica's output for profitable trading. The strategy of combining protective stop with a profit target, setting these at the extremes of the previous trading day and trading in the direction of the forecast (referred to as the Day Range strategy and reported in the previous post is modified to remove the profit target. Much better results are obtained.
- Tags:
- frontpage
February 12 2010, 3:21pm | Comments »
-
Further analysis of the day-range strategy. Selecting the forecasts to trade.
In the January performance review for the Danica trading system, an idea of a day-range trading strategy capitalizing on the high quality forecasts for the direction of daily high and low was expressed and a set of what-if charts for the forex pairs tracked was provided and discussed. This post is a deeper and more quantitative discussion of the strategy. Which forecasts should be acted upon? What is the expected profit per trade? How are profits/losses distributed statistically? -- these are the questions addressed.
- Tags:
- frontpage
February 9 2010, 2:45pm | Comments »
-
January performance review for Danica-9am algorithmic system
For the first time I am able to discuss performance of a ForexAutomaton system without the "benefit of hindsight" caveat: the results for January have been obtained in real time, hence no hindsight. Executive summary In the absense of a capital allocation and trade-idea disrimination system, the main figure of merit is the Pearson correlation coefficient between real and forecast logarithmic returns in day high, low and close. In January, these figures of merit appear to remain in line with longer range historical performance. A type of a trade strategy specifically designed to take advantage of the superior forecasting quality for daily high and low, while minimizing exposure to the forecast for close, is discussed, with an attempt to evaluate performance using the recorded output of the system in January.
- Tags:
- frontpage
February 3 2010, 7:24pm | Comments »
-
From forex forecasts to forex signals. Effect of the forecast move magnitude.
In the course of the trading system optimization, best returns have been seen to be obtained by ignoring forecast moves below a certain threshold, and acting on those above that threshold (that threshold was also called entry parameter). A small fraction of large returns has been seen to account for much of the positiveness of the Pearson correlation coefficient between actual logarithmic returns and their forecasts. Each Danica output contains forecasts for 14 forex rates, and a natural question is: how do I pick the ones to place trades? One might expect that the odds of success can be improved by selecting those markets where the next move is forecast to be large. I take version 0.5 of Danica forex system and study dependence of the correlation between the forecast and real logarithmic returns in day close on the relative strength of the forecast move.
- Tags:
- frontpage
January 29 2010, 6:29pm | Comments »
-
From forex forecast to forex signal. Level 0 forecast discrimination.
Some opportunities for analysis are offered by the fact that the forex trading system such as Danica gives not just a forecast for the next close, but a combination of next high, low and close. This report is the first attempt to develop a selective approach to the forecasts, a discrimination algorithm of sorts, such that a decision to give the forecast further consideration or ignore it would be based on the information contained in the forecast itself.
- Tags:
- frontpage
January 20 2010, 7:53pm | Comments »
-
Explaining Danica v0.5 optimization choices
This report gives a more detailed discussion of the optimization trade-offs made for Danica, as compared to the initial announcement. At the same time, it establishes a reference point for future studies of past performance.
- Tags:
- frontpage
January 14 2010, 3:45pm | Comments »
-
Danica the daily forex forecaster is rolled out
It's now official: from now on, a forex forecast of low, high and close for the next 24-hour period will be posted on this site daily at 9am Eastern time. The system is named Danica following the naming convention where first names starting with D are assigned to systems with daily decision-making scale.
- Tags:
- frontpage
December 31 2009, 3:37pm | Comments »
-
Explaining the output of Danica trading system
This document explains the output of Danica -- an experimental free (payless but closed-source) ForexAutomaton day-scale forex forecasting system.
- Tags:
- frontpage
December 24 2009, 4:21pm | Comments »
-
Prediction quality for high, low is improved by tying the four components of a day candle together.
In this week' update, I demonstrate an improvement in the prediction quality for day low and high in six major forex pairs -- EUR/USD, USD/JPY, GBP/USD, USD/CHF, AUD/USD, and USD/CAD -- by imposing the obvious constraints of low being below high and day's open (which is assumed in forex to coincide with previous day's close) being between the day's low and high. As before, I use Pearson correlation coefficient between the real and predicted logarithmic returns, as a figure of merit to gauge the prediction quality. Contrary to my expectation, no visible improvement for close is obtained by such technique. The results have to be compared with the previous report. Fig.1. Pearson correlation of predicted and actual day-scale logarithmic returns in low, high and close as a function of the forecasting parameter nicknamed Fred. The shaded bands indicate a measure of uncertainty, their boundaries mark one standard deviation (among the forex pairs considered) distance to the points. Back-testing simulations give the forecasing engine no access to the future data, direct or indirect. Significantly positive (and ideally, large) values correspond to quality forecasting. Note that the quantities at different Fred are not quite statistically independent, therefore the error bands should be understood as describing the uncertainty of the position of the curve at large rather than that of individual points.
- Tags:
- frontpage
December 18 2009, 1:57pm | Comments »
-
Predicting next low and high looks much easier than next close.
Just like logarithmic returns can be defined and analyzed for daily close, they can be defined for daily high and low. Japanese candlestick charting techniques, believed to have predictive power, study patterns formed by open, low, high and close as the time series progresses. In this report I extend application of my newly developed forecasting figure of merit, Pearson correlation coefficient between the real and predicted logarithmic returns, to the daily high and low, taking another look at the dependence of the prediction quality on the magnitude of a forecasting parameter nicknamed Fred. As the prediction quality is seen to be much better for the next high and low than it is for next close, I am contemplating ways of improving quality for close. Fig.1. Pearson correlation of predicted and actual day-scale logarithmic returns in low, high and close as a function of the forecasting parameter nicknamed Fred. The shaded bands indicate a measure of uncertainty, their boundaries mark one standard deviation (among the forex pairs considered) distance to the points. Back-testing simulations give the forecasting engine no access to the future data, direct or indirect. Significantly positive (and ideally, large) values correspond to quality forecasting. Note that the quantities at different Fred are not quite statistically independent, therefore the error bands should be understood as describing the uncertainty of the position of the curve at large rather than that of individual points.
- Tags:
- frontpage
December 10 2009, 5:01pm | Comments »
-
Intermarket analysis seems to make no dramatic difference to forex prediciton quality on day scale
As we are gearing up towards the launch of a real-time day scale forecasting service, I am using the newly developed forecasting figure of merit, Pearson correlation coefficient between the real and predicted logarithmic returns, to make a choice of the operating mode for the test mode of the service. I am taking another look at the dependence of the prediction quality on a day scale on the magnitude of a forecasting parameter nicknamed Fred for two distinct pattern recognition modes, namely that of completely independent analysis of the different exchange rates (so-called Step One), and the one in which the time series for the indvidivual exchange rates are considered jointly in order to utilize potential inter-market patterns (Step Three). Fig.1. Pearson correlation of predicted and actual day-scale logarithmic returns as a function of the forecasting parameter nicknamed Fred. The vertical bands indicate a measure of uncertainty, their boundaries are plus/minus one standard deviation to the points. Back-testing simulations give the forecasing engine no access to the future data, direct or indirect. Significantly positive (and ideally, large) values correspond to quality forecasting. Note that the quantities at different Fred are not quite statistically independent, therefore the error bands should be understood as having to do with the curve at large rather than with individual points.
- Tags:
- frontpage
December 4 2009, 6:50pm | Comments »
-
Forecasting optimization: an overlooked parameter fix improves quality while "efficient market" Monte Carlo supports the results.
Following up on the topic of our forex prediction quality measurements, I've decided to conduct the same analysis on the simulated data, unpredictable by construction. As before, I am tracing the dependence of the Pearson correlation coefficients between predicted and actual logarithmic returns in day close value on the magnitude of the forecasting parameter nicknamed Fred. Fig.1. Pearson correlation of predicted and actual day-scale logarithmic returns as a function of the forecasting parameter nicknamed Fred. The vertical bands indicate a measure of uncertainty, their boundaries are plus/minus one standard deviation to the points. Back-testing simulations give the forecasing engine no access to the future data, direct or indirect. Significantly positive (and ideally, large) values correspond to quality forecasting. Shown in red is the standard deviation band associated with simulated data (see text). Note that the quantities at different Fred are not quite statistically independent, therefore the error bands should be understood as having to do with the curve at large rather than with individual points.
- Tags:
- frontpage
November 27 2009, 12:27pm | Comments »
-
Mikhail Kopytine gives an interview to Forex Hunter Blog
An excerpt from the interview: - How do you see the future of trading and markets in the XXI century? I see the XXI century as an age of crowds of kaleidoscopically diverse individuals, empowered and inter-connected by technology, interacting in more and more complex, conditioned and mediated ways -- and at the same time more and more differentiated within personal "virtual realities." I anticipate further inter-penetration between the concepts of information and capital.
- Tags:
- frontpage
November 18 2009, 4:11pm | Comments »