How to Use a Dead-Man Timer in Your Trading
By Dave Mabe
One thing I learned in my Trade-Ideas CTO days was the value of a simple but powerful concept: a dead-man timer.
Imagine you have an important process that needs to complete each day.
It's reliable, so it only fails every once in a while.
But when it does fail, someone needs to know immediately.
You could add a step to the process that sends you a notification when it completes, but you will quickly become numb to those notifications.
You could add a step that only notifies you when there's an error.
This is better, but what happens if the notification step fails? Or your server went down so it can't actually send the notification?
The right way to set up a notification in this situation is to have a separate service - outside your infrastructure - that is set up to notify by default.
Then have your important process ping the service upon completion each day.
This way, no matter what the reason for the failure, you get notified when something DOESN'T happen.
This is called a "dead-man timer" and it's a much more reliable way to be notified of issues.
How can you use this concept in your trading?
With automation, certain processes need to be up and running.
Nothing is more frustrating than your software not running for some reason and you miss a bunch of profitable trades.
An obvious way to implement a dead-man timer is to have a process make sure things are running properly at a certain time, and send a ping to the dead-man timer.
If it doesn't report for whatever reason, you can get a high priority notification in time to do something about it.
The service I use for this type of alerting is Better Uptime (now called Better Stack).
Do you use this type of alerting for your trading infrastructure?
If not, imagine how painful it would be to miss an entire day of profitable trades.
-Dave
P.S. Does it feel like your strategy works for everyone but you? Make it work for your data with MabeKit. Get Instant Access to MabeKit