My Perspective on Algorithmic Trading (Part 3)
Posted by Mark on October 8, 2019 at 11:57 | Last modified: April 13, 2020 15:09Today I will finish discussing advantages to algorithmic trading and then move on to some disadvantages.
Algorithmic trading simplifies portfolio diversification. Automation allows concomitant management of multiple accounts and strategies. This allows risk to be spread across different markets in addition to hedging against losses by establishing low or negative correlation. A computer can scan multiple markets, generate orders, and monitor positions in fractions of a second. The human trader is much slower and can easily make mistakes.
One final advantage to algorithmic trading is the opportunity to manage strategies 24/5. Futures markets are open nearly 24 hours during the week. Computers can execute trades when their human counterparts are sleeping.
Despite all the algorithmic trading benefits I have discussed, it’s important to understand the disadvantages as well.
Like all technology, machines are susceptible to failure. “Install the software, program the rules, and let the computer trade” obscures what is actually a very sophisticated process. When trade orders are stored on personal computer rather than exchange servers, malfunction can prevent orders from being sent. Even slight delays in order transmission can result in discrepancy between theoretical and real trades (the “backtested expectancy” mentioned in this fifth paragraph).
Because technology is not infallible, algorithmic trading requires routine monitoring. I can’t just enter the program, walk away from the computer, check my brokerage account monthly, and expect to see linear growth. Potential problems such as power disruption, internet connectivity issues, and system quirks can result in missing or duplicate orders. Errant orders can also arise due in large part to faulty programming. Especially for the non-professional, slopping coding errors are not uncommon. What’s most worrisome to me is how easily such errors can go unnoticed (a good reason to develop trading strategies in groups as mentioned in the second-to-last paragraph here).
Because technology is not infallible, backups are always prudent. I should have accounts with different brokers in case something goes wrong with one specific account or a brokerage website. I should have a backup computer available in case something breaks. I should have backup power or internet available—or some solution for what to do in case either one fails (e.g. nearest coffee shop or library with free WiFi).
The technology of algorithmic trading has a learning curve. Trade in small size as you become familiar with the process.
Perhaps surprisingly, I think the biggest disadvantage to algorithmic trading is the challenge of finding viable trading strategies. This topic can fill volumes.
I will start next time with a single post.
Categories: System Development | Comments (0) | Permalink