Look at the market action through cold and alien eyes that know no fear or greed -- the eyes of Forex Automaton™ .

Recent:

Our primary goal is to create a public information service providing financial markets forecasts, based on our proprietary forecasting tools: an automated trading system -- a Forex Automaton™. Our secondary goal is to quantify and monitor the very existence of sustainable opportunities for arbitrage profit-making. Or simply put, to monitor the degree to which these markets are more predictable than a "fair game" -- to a trader without access to insider information.

Once you know something, such as the past history, pretending that you do not is very difficult. The ways a trader can delude him-/herself while backtesting a trading system vary in degree of subtlety. A crude one is to develop a set of trading rules by observing the past history of the market, and apply the algorithm on paper to the same past history. A more subtle one has to do with the choice of adjustable parameters. It is common to abstract the algorithm, and split it into the "artificial intelligence" (AI) core, capable of "learning", and adjustable parameters which may control the learning process, the application of its effects, or both. A developer would then prevent the AI from learning the future, and run a trading simulation program on historical data, having the AI algorithm fixed, for a range of parameter values and then pick the value that gives the best result. By not letting the AI learn the future, the developer reduces the level of self-delusion somewhat. The best simulated performance would still incorporate "the benefit of hindsight" albeit in a refined way -- through the choice of the adjustable parameters. While the AI did not learn from the future directly, its way of learning would incorporate the benefit of hindsight, and the results might contain a survivorship bias of sorts. The result? In case of a poor AI, the result will be a trading strategy with a hidden "peso problem".

Below I demonstrate the quality of the Forex Automaton™ AI directly, by removing the selection step and thus the survivorship bias it brings in. I look simultaneously at the entire range of possible "ways of learning", refusing to reject  those inside the range. Nor do I reject any of the possible money management styles, for the same reasons.

Fig.1:Return vs risk for paper trading in real AUD/JPY and simulated reference random walk markets. Vertical axis: annualized return, calculated as an average of independent monthly statistics, 1=100% per annum. Horizontal axis: standard deviation (RMS) of such annualized return from the same monthly statistics, same unit. Each point represents an instance of a trading system, a simulated trader; points differ by trading strategy which is subject to optimization. Black points: trading 4 fantasy markets with volatility of AUD/JPY, random by construction (no predictability or patterns). Red points: real AUD/JPY. Not all outcomes are included: in some cases, the trading did not last even a month therefore no RMS could be calculated -- there would be no other month of trading performance to compare with. In such cases the system is programmed to record a zero RMS, and such outcomes did not make it into the figure. The only other requirement is that of excluding the mean annualized returns above 5 (500%) and RMS above 20 -- happy as I am to include higher returns, they would make the picture difficult to analyze visually.

While it is true that past performance does not indicate the future, the only reliable information we have is about the past. A few important things make a difference between unbiased trading-system testing and self-delusion. Here I summarize my current understanding.

This article begins a series of analysis reports investigating a degree of predictability in the LIBOR rates, a popular capital cost indicator. The analysis is based on historical LIBOR interest rate data released by the British Bankers Association. I continue with the same technique proved useful in the predictability analysis of forex exchange rates, as our interest in the interest rates in general is in part provoked by the results of the latter analysis, namely:

• sometimes, one forex exchage rate can "show the way" to a number of others, or in other words, foretell (in a probabilistic or statistical sense) their movement.
• when that happens, it is usually the exchange rate with a large interest rate differential showing the way to the ones with lower interest rate differentials.

Obviously, when exploring these "loopholes" or market inefficiencies for wealth generation, an algorithmic trader or a forex trading system (an automated decision making algorithm such as the one being built here on Forex Automaton™ site) must be mindful of the picture of LIBOR rates and its evolution, albeit in a somewhat different context than a long-term money manager. Being able to predict events, even in a weak statistical sense, is even better than merely following. Besides being useful via their implications for forex forecasting, LIBORs form an underlying indicator for derivatives of their own. LIBOR futures contracts and options on such contracts are traded on the CME. How does the predictability of LIBORs compare with that of currencies? Which one, LIBOR or forex, is more attractive to trade? Answering these questions, or providing a technical analysis framework to approach the answers, while leaving the fundamentals and event-driven trends aside, this series of articles about correlation features in LIBORs will serve as a useful compliment to our set of forex correlation analysis notes. I start this new series of articles with the all-important US Dollar LIBOR.

The negative correlation peak of Pound Sterling/US Dollar and US Dollar/Swiss Franc, when studied with hour time scale resulution, looks symmetric and wider than one bin. Therefore, either exchange rate may have predictive influence on the other. Of course, this is a time-integrated picture and the time evolution may have more complex structure: the symmetric peak may be composed out of a number of asymmetric ones.

Pound Sterling/Japanese Yen is weakly positively correlated with US Dollar/Swiss Franc and, what is an independent piece of information, predictive correlations seem absent in this pair of exchange rates with this (hour time scale) resolution. The necessary caveat is that this is a time-integrated picture.

The two exchange rates with the same sign of the interest rate differential, Euro/Japanese Yen and US Dollar/ Swiss Franc turn out to be negatively correlated. Studied on the hour time scale, the correlation looks highly asymmetric, making EUR/JPY look like the leader and CHF/USD (not USD/CHF, due to the negativeness of the correlation) -- the follower. The necessary caveat is that this is a time-integrated picture.

There is quite a shortage of good historical LIBOR interest rates data charts with a selection of currencies and maturities on the net. ForexAutomaton.com has a new section for just that: find historical LIBOR charts sorted by currency, loan duration and year on our site. These are essentially graphical "dumps" of our SQL database, based on the official BBA LIBOR archive. Use the Filter feature of the list to find the data you need.

The two exchange rates with the opposite signs of the interest rate differential, Euro/Pound Sterling and US Dollar/ Swiss Franc and negatively correlated. The correlation, when studied on the hour time scale, seems too tight to offer forecasting potential. Such simple one-peak structures centered at the zero time-lag are called "trivial correlations" in this series of analysis notes. The necessary caveat is that this is a time-integrated picture.