Use Algos to Fight The Algos

Every trader’s resolution for this year should include a transition to a mode of operation that is compatible with current market environment. Among other things, this includes abandoning random discretionary trading and adopting a systematic framework. When human traders start a fight with algos, chances of winning are slim. I offer some basic guidelines for this transition.

Trading is getting harder every year. When I see traders in social media making references to some trendlines, chart patterns, or some moving average levels and try to rationalize price action based on them I am nearly 100% true that this is a recreational activity for them. There are also some who try to impress using their (purchased) follower count and even offer to teach others how to trade in 21st century using these or similar simplistic random ideas. Anyone who thinks money can be made from counterparties such as Renaissance Technologies (Simons) and Point72 (Cohen), just to name two, using random patterns is more than wrong: he is incredibly naive.

Below is just one example of how harder trading has gotten in recent years. In the past, short-term traders could use simple strategies to profit in daily and weekly timeframes. One such strategy in the weekly timeframe involved buying the S&P 500 index (or a representative index portfolio or SPY ETF after 1993) if the close of the week was higher then the close of the previous week and also higher than the open. The exit was at the close after four weeks. Below is the equity curve of this strategy from 01/1950 to 12/1999 when equity is fully invested.

Although the strategy experienced a drawdown in the 1970, it recovered and was overall highly profitable with 63% win rate and Sharpe 0.75. Exposure is about 85%.

Below is the performance of the same strategy from 01/1950 to 12/2016

I hope you get the message. I have placed a vertical marker at the low of the financial crisis and where the -66% drawdown occurred. There is no way to profit now from simple ideas such as this one based on short-term price action momentum.

The transition to systematic trading

Systematic trading is nothing new but nowadays there are many more tools, most of which are for free, to facilitate a transition. The basic requirements include:

  1. The strategies must be modeled in the form of an algorithm
  2. The algorithms must be programmed in a suitable platform

1.  If the strategy involves subjectivity, such as subjective chart patterns or evaluating the impact of fundamentals, then it may not be suitable for systematic trading.

2. Programming skills are necessary for systematic trading. There is also software that generates code but there is often need for modifications. Traders who cannot program have very low chances of success in this new environment unless they partner with someone who does that. This is an algo game. Whoever tells you otherwise is either naive or a liar. Nearly 80% of exchange volume is generated by algos.

There are platforms, some for free, that provide all the tools to develop strategies, test them and deploy them for systematic trading. Amibroker, Tradestation (free with account) and NinjaTrader are three examples out of may choices that are available.

For developers who also want to participate in profit sharing trading a fund there is Quantopian. I have several strategies running on that platform and my DLPAL software generates automatically code for it but I will be honest in that I would never put there anything I will consider a major edge. This is not because I do not trust Quantopian; It is is a matter of principle because I have no control of security of strategies coded there and there is small risk of hacking. Therefore, I would never code my PSI5 mean-reversion algo in Quantopian or similar platforms. The DIAT4S3 strategy I know how to modify to regain it’s edge in case it starts losing it.

As far as learning programming, Python is the way to go now. It is not my favorite language, especially due to indentation requirements, but this is what the industry has chosen. Here is a free course.

Most traders who are addicted to visual charts find the transition hard and that reinforces the cognitive bias known as backfire effect. This causes them to even hold their views more strongly and even turn hostile towards those who try to warn them. Obviously, the transition to systematic trading comes with some pain but there is no other way. Staying at the level of chart patterns and trendlines is similar to those merchants that still used horses to transport their goods after trucks were available. That led them to their demise.

Here is an interview I gave last week about technical analysis and what has changed in recent years. Here are examples of strategies I use, including results from Quantopian.

I wish you success in the New Year!

If you have any questions or comments, happy to connect on Twitter: @mikeharrisNY

Subscribe via RSS or Email, or follow us on Twitter.

Charting and backtesting program: Amibroker


Technical and quantitative analysis of Dow-30 stocks and 30 popular ETFs is included in our Weekly Premium Report. Market signals for longer-term traders are offered by our premium Market Signals service. Mean-reversion signals for short-term SPY traders are provided in our Mean Reversion report.

Copyright Notice

This entry was posted in Trader education and tagged , , , , , . Bookmark the permalink.

Leave a Reply