Important: There is no guarantee that ATS strategies will have the same performance in the future. I use backtests and forward tests to compare historical strategy performance. Backtests are based on historical data, not real-time data so the results shared are hypothetical, not real. Forward tests are based on live data, however, they use a simulated account. Any success I have with live trading is untypical. Trading futures is extremely risky. If you do trade live, be prepared to lose your entire account. This is for educational purposes only. I recommend using ATS strategies in simulated trading until you/we find the holy grail of trade strategy.
To Trade, Or Not To Trade—That Is No Longer The Question
Trading is evolving at a rapid pace and it’s doing so with the help of AI. We’re no longer using simple “if / then” statements to automate a command structure. Now we have the ability to transform a simple crossover strategy into something that can also add market context and machine precision. To trade, or not to trade—that is no longer the question. Markets aren't rigid. They flow, they adapt, they surprise us.
That's where our journey into agentic workflows begins. We're moving from the world of strict binary decisions into a realm where our strategies can think, adapt, and even explain themselves.
In Part 1 and Part 2 of this series, we explored how AI agents are proving to be a democratizing force—giving a single person the ability to have a team of analysts. We also looked at some example workflows. Today, we’re going to get personal. We’re going to see how we might be able to apply AI agents and agentic workflows to ATS strategies. In particular,
What is the best way to use machine learning (ML) in the command structure workflow?
What is the best use of LLMs and generative AI in the command structure workflow?
What about a hybrid approach?
Before you listen to Part 3 of the Podcast, I want to give a quick primer using ATS Strategy 1 as an example.
Strategy 1: The Evolution of Alpha
Strategy 1 was the first strategy to be published on ATS on Jan 14, 2021. Given the performance milestone reached in 2024, it seems only fitting that we revisit it now.
Strategy 1 is a simple strategy that uses the Linear Regression Indicator (LinReg) and the Volume Weighted Moving Average (VWMA).
Enter Long - When LinReg crosses above VWMA.
Enter Short - When LinReg crosses below VWMA.
Admittedly, Strategy 1 is simple, but it gives us: a solid foundation, a slight edge, and a nice sample size of trade opportunities to work with.
Here’s what the workflow for Strategy 1 looks like now:
Signal: "LinReg crossed VWMA? Send it!"
Decision Making: Pure if/then, no questions asked
Risk Management: As basic as a broker's weekend wardrobe
What happens if we add ML? ML takes our crossover and turns it into a probability playground. Here’s how the workflow might look if we add ML:
