Premium Market Analysis

Price Action Strategies, Quantitative trading, Trader education

A Few Practical Principles of Quantitative Trading

Many traders would like to transition to quantitative trading but think it is overwhelmingly complicated.  However, there are many levels of quantitative trading and if the right tools are used, the complexity level decreases significantly to the point that any trader who knows simple arithmetic can do it.

It is easy to just look at a chart for promising formations or have software to scan the market for patterns. The problem with this approach is that it works only for the few that have a good understanding of price action, good intuition and can control their wishful thinking. The rest, about 95% of the traders, in the best case end up with small losses. The reason for that is known for many years: the trading signals charting methods generate are mostly random and their expected return is negative due to trading cost.

Even those who use backtesting tools in an attempt to evaluate the potential of trading systems based on chart patterns, lagging indicators, or other ambiguous signals, such as support & resistance and trendlines, to name a few, and although they have made a step towards quantitative trading, they often fail to accomplish their goal mainly due to the pitfalls of this methodology. In a nutshell, this methodology fails because of a combination of small samples, curve-fitting and data-mining bias. The latter grows uncontrollably high when new ideas are tested on the same historical data more than once.

The extra steps that are required to test for the presence of bias and determine whether a trading method has true potential are taken by few. The following principles are known to reduce the bias inherent in the trading system development process:

  • Focus on just a few patterns or indicators with proven potential
  • Avoid indicators with optimized parameters
  • Avoid the use of complex exit rules that take the place of entry rules
  • Use the out-of-sample only once and then start with a fresh idea
  • Draw conclusions only if the samples are sufficiently large
  • Use portfolio backtesting on comparable securities to increase trade samples
  • Try to increase the win rate as much as possible. Trading is all about winning, do not let anyone convince you otherwise.

The justification of the above often gets quite complex but to some they are intuitively true.