The Inevitable Crash

Automation makes a strategy consistent, until something breaks. The failure points to plan for: your own machine, your broker, the exchange, code bugs, liquidity and margin.

Automated trading brings a great deal of convenience. It helps execute a strategy consistently and systematically, with little or no human intervention. There are already several commercial algo platforms and browser extensions that let the retail segment automate strategies, and anyone can subscribe to a broker’s API and build custom automated strategies.

But automation is not without its share of problems. It can crash and burn along with your capital in the blink of an eye, before you even know, or can intervene to prevent it.

Where automation breaks

• Your own infrastructure. You can run your strategy on a local PC or server, using broker APIs for the data feed and order execution. But the machine can crash, or the internet connection can fail, landing you in trouble. Keep redundant computing resources and internet connections so you can switch over when one fails, and add an alert so a manual switch is timely. If you are technically inclined, automate the switch, though line flapping can still defeat it, and both systems can fail at once, which makes the redundancy useless. A cloud server or a commercial auto-trading interface offers better reliability, ease of deployment and maintenance, but their infrastructure and connectivity can fail too. There is no running away from a technical failure at your end. It can hit anytime, and all you can do is cover the risk as far as possible.
• Stock broker infrastructure. This includes the front end, the back-end order and management system, and the leased lines connecting to the exchange that route data packets to and fro. A failure in any of these grounds your strategy. This too is inevitable, and we have seen several such instances over the years. One workaround is to keep a trading account with a second broker, to enter or hedge open trades if your main broker’s infrastructure gives way.
• Exchange infrastructure. Exchange servers and connectivity can crash, as they did on 24 February 2021. There is nothing you can do in this scenario. Since no trading happens while the exchange is down, you probably need not worry about missed entries and exits, but stay aware of the next exchange opening time and any follow-up action the broker or exchange has taken.
• Software bugs. A bug in your strategy code can lead to misfired orders, high frequency orders or oversized positions, and it can eat your capital fast. Test the strategy code in a paper-trading environment first, then live with small capital, before going all in.
• Liquidity. If many people run the same strategy in an automated environment, such as a commercial auto-trading platform, it can create a liquidity squeeze. You may not be able to enter and exit at favourable prices if there is not enough supply or demand, and the order-book depth can be shallow because others have used it already. This can happen even in highly liquid instruments if enough users follow the same strategy. For scalping and option-premium strategies, the slippage from a liquidity crunch is enough to wipe out the small per-trade profit. It is better to run automated strategies at your own end, to counter this, and to trade liquid instruments only, using limit orders wherever possible instead of market orders.
• Margin and cash. Option sellers and futures traders have to keep margin money with the exchange. As the exchange changes the requirement through the day, your margin limit can be hit and a penalty levied against your account, and your position can even be auto squared off by the broker’s risk team. Always keep capital in excess of what the strategy metrics suggest, and add an alert for when the margin requirement nears your account ceiling.
• Manual authentication. Broker APIs require you to re-authenticate your credentials every day for security reasons. Forget the process, and your auto-trading strategy will not run.