“Automate the search for promising assets”
The use of trading systems means, among other things, that more markets and smaller time levels can be considered with the same team. The higher trading frequency and diversification into more markets and time levels will in turn lead to better performance and reduced risk for the investor. Philipp Kahler von Tradesignal chatted at the 143rd Hedgework from the sewing box of a technical analyst. He answers the most important questions here.
HEDGEWORK: Mr. Kahler, you represent the quantitative side of portfolio management. Has fundamental analysis become obsolete?
Philipp Kahler: Quantitative analysis deals with the creation of investment rules, which are so clearly defined that you can even teach them to a computer. Whether these rules are based on price or fundamental data is the same. The only important thing is that you can test the rules and regulations in a meaningful way and that the result is convincing.
HEDGEWORK: What are the advantages of technical analysis?
Kahler: Compared to fundamental data, technical indicators have the advantage that they are available in real time. It is not the distributions in the last quarter that are decisive, but rather what the market is doing today. My market-to-market result is then also evaluated on the basis of current prices. Whether the technical indicators are better than the fundamental analysis indicators is unclear, but as a trader, the most important thing for me as a trader is to have my risk under control NOW – and that’s easier with technical analysis.
HEDGEWORK: In your presentation, you talk about the transition from technical to evidence-based technical analysis. What’s behind it?
Kahler: I mean quantitative models that make use of classical technical analysis in the toolbox. Indicators and simple price patterns are ideal for searching markets for opportunities. However, even the best technical analysis indicator alone will hardly justify a successful trading approach. The combination of several indicators, perhaps even combined with pre-screening by fundamental analysis – this is the best basis for creating a stable, quantitatively controlled portfolio.
HEDGEWORK: In which areas are quantitative systems particularly useful?
Kahler: If the holding period of a position is somewhere between five minutes and two weeks, then there is no way around technical analysis. If your holding period is between two weeks and several months, the technical analysis will at least provide you with valuable services for timing your decisions. What you do is not decisive.
HEDGEWORK: Quantitative systems can, in a first step, bring structure to investment decisions?
Kahler: Yes, you can automate the search for promising values, improve the timing of trading activities, and you will also see an improvement in the delivered performance by avoiding emotional decisions.
HEDGEWORK: What role do FinTechs and Robo Advisor play? Has their increased market presence already made a significant difference?
Kahler: Yes and no. Of course, flash-crashs are caused by the increased use of machines, but if you look at the Dow Jones for more than 100 years, you will soon find that the market hasn’t changed. The daily/weekly/annual volatility has been almost constant for more than 100 years. These new technologies have the strongest influence not on market behaviour, but on the business model of traditional asset managers.
HEDGEWORK: What added value can an investor expect from a quantitatively controlled portfolio?
Kahler: One advantage is that the use of trading systems means that more markets and smaller time levels can be viewed with the same team. The higher trading frequency and diversification into more markets and time levels will then lead to better performance and reduced risk for the investor.
HEDGEWORK: Could you please describe how you proceed with a new product offering or a new investment strategy?
Kahler: This can be done in two ways. Either I have a trading system that I am convinced of. I then look for all the markets in which it functions and combine them into a quantitatively managed portfolio. Or I get a market given. Then I try to develop systems that work without adapting the parameters, e. g. in the hour/day and week range. Several such approaches are then combined to form a portfolio.
HEDGEWORK: Backtesting is an essential part of a product launch. These are sometimes not very reliable in retrospect. What is important to note here? What are the pitfalls?
Kahler: The baking test is not the problem. The adaptation of the strategy to the market – the parameterization and weighting of the individual components of the trading system – is critical. Since only a few adjustment screws lead to a high degree of adaptability of the trading system, there is a high risk that one adapts too much to past data, called curve fitting, without the system’s set of rules really saying anything decisive about the market.
HEDGEWORK: What method do you propose for testing the robustness of a strategy?
Kahler: On the one hand, you can first test the stability of the parameters. If a system includes the 200-day line, then it should work roughly as well with the 150-day and 250-day line. In a further step, the system must then be tested with unknown data. If a trend-following model works in Germany, for example, then it should not fail in the USA either. And finally, you should let the system disappear in the drawer for half a year and then check again to see if the real-time results were as expected.
HEDGEWORK: Maybe we can take a little more look at your sewing box. What practical tips can you give prospective quants?
Kahler: It’s easy: Learn to trade! Without a computer, with real money, so losses really hurt. Even though studying science is an advantage, I do have the experience that traders who are not blinded by numbers because of their own experience will eventually develop the more stable systems.
HEDGEWORK: Finally, perhaps a glimpse into the future?
Kahler: Computers will take on more and more tasks, the worldwide availability and comparability of brand names will in many cases make brand names unimportant. Investors no longer ask for a certain fund, but want to invest their money with manageable risk and limited correlation to other markets. Whether an algorithm or an analyst does this is information that will not reach the investor at some point. Since the computer often delivers better performance and is paid less than the classic fund manager, it doesn’t take much imagination to estimate future developments.
The interview was conducted by Ronny Kohl, automatic translation by DeepL
Philipp Kahler is Senior Quantitative Analyst at Tradesignal Ltd. and advises financial institutions and energy trading companies on the development of quantitative-based trading strategies. He gained professional experience as a trading system developer in proprietary trading at Berliner Landesbank. There, he managed a wide range of very successful, systematic trading systems for several years.