Speculative Bubbles in DIY Trading Algorithms

WSJ‘s Austen Hufford profiles a group of retail investors who use DIY trading algorithms to try and extract alpha. Some comments:


1. Hufford’s interviewees sound very similar to dotcom era day traders at the peak of the 1995-2000 speculative bubble — particularly about their positive expectancy of financial profits.


2. The choice of asset class (forex) and markets (S&P500 and Nasdaq Composite) will likely mean that Hufford’s retail traders are picked off by high-frequency traders.


3. Hufford’s article has some typical anecdotes on how traders lose money early on in the trade development process and how coding errors can lead to unprofitable trades. On the upside the group of traders now has a daily, actionable routine  to deal with financial markets.


4. Aspects emphasised by institutional traders — pre-trade analysis, market microstructure, transaction cost economics, and tax implications — are not considered by Hufford’s retail traders.


5. Hufford mentions Interactive Brokers whose own history of algorithmic trading is featured in Scott Patterson’s book Dark Pools. I took part of Tucker Balch’s course but also compared it to other known research such as Andrew W. Lo’s studies. I’ve also looked at Quantopian and Rizm.


6. $200,000 in account size for equities / forex depends also on other factors such as leverage, position size, and regularity of trading.


7. Computer-driven hedge funds use very sophisticated programming.


8. Twitter and academic journal research uses sentiment analysis.


9. Hufford’s trading strategy example uses a moving average indicator in technical analysis. This is a basic strategy for retail traders that is now gamed by high-frequency trading algorithms.


10. Agile software development practices have insights for how to refactor code.


For a comparison of methodology see Alvaro Cartea, Sebastian Jaimungal and Jose Penalva’s Algorithmic and High-Frequency Trading (New York: Cambridge University Press, 2015).