It's Alive! Automated Trading Strategy #104 (HMM-Dual Layer)
This strategy made $10K in two weeks on 25 trades in forward testing
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 atypical. Trading futures is extremely risky. You should only use risk capital to fund live futures accounts and if you do trade live, be prepared to lose your entire account. There are no guarantees that any performance you see here will continue in the future. I recommend using ATS strategies in simulated trading until you/we find the holy grail of trading strategy. This is strictly for learning purposes.
I’m not just developing strategies—I’m on a quest for the holy grail of automated trading. Questions? Check the FAQs or feel free to reach out directly: AutomatedTradingStrategies@protonmail.com.
October 31, 2024 - Exactly one year ago, on Halloween night, I shared Strategy 87 with you. I called it “The Trader’s Homunculus”. It was a blueprint for creating artificial trading intelligence using a nine-factor opportunity matrix. Like the alchemists of old, I was attempting to embed decision-making consciousness into code.
Strategy 87 had intelligence —optimized weights that prioritized market context over technical indicators—but it lacked one crucial element: adaptability.
The system couldn’t learn. I had to tell it what mattered. The weights were fixed, determined by historical optimization. It was, in the truest sense, a proto-AI —the form of intelligence without the spark of life.
October 31, 2025 —Tonight, exactly one year later, the homunculus opened its eyes. This Halloween, I’m sharing what happens when you stop trying to create intelligence and instead give your system the tools to DISCOVER it.
The alchemists of old failed not because transmutation was impossible, but because they were trying to impose their will on nature rather than understand nature’s hidden patterns. Isaac Newton spent decades on alchemy, but his lasting contribution came from discovering the laws that were already there —gravity, motion, light. He stopped trying to create and focused on discovery.
The Spark of Life
Last Halloween, I ended the post with this vision:
“I’d like to let the model discover its own version of the matrix, based on its assessment of which factors drive performance under different markets.”
That is the alchemical dream —not to impose structure, but to discover it. Not to tell the system what states exist, but to let it reveal them through experience.
The problem? I was still thinking like a traditional trader. I was modeling markets instead of behavior.
Then I found Leonard Baum’s principle:
“The structure exists in the data. The algorithm’s job is to reveal it, not impose it.”
And I found Jim Simons’ secret:
“Our entire premise was that human actors will react the way humans did in the past... we learned to take advantage.”
Notice he didn’t say “prices will move like they did.” He said “human actors will react.” Behavioral patterns, not price patterns.
This isn’t semantic hairsplitting. It reveals how Renaissance thinks about markets: they’re not modeling prices directly. They’re modeling behavior through its effect on performance.
Tying back to Part 1, public information suggests Renaissance Technologies:
Doesn’t predict prices — They predict STRATEGY PERFORMANCE (trade outcomes, not price movements)
Uses HMM extensively — But on order flow, trade outcomes, execution patterns
Runs 200+ strategies —Each optimized for specific performance regimes
Strategy 104: How Exactly Does It Work?
Before getting into how the strategy works, let’s look at forward test performance over the last two weeks:
As you can see the strategy made just under $10K with 25 trades. It has a win rate of 68%, max drawdown of $2K, average MAE of $500 and a profit factor of 3.90.
Here’s a breakdown of trades by day.
You can download Strategy 104 at the bottom of this page.
I’ve had a good amount of feedback from some of you saying that HMM doesn’t work. It didn’t work for me either—-at first. And, this is only a few weeks of forward testing, but there are some things you can do to make the algorithm more “automated-trading” friendly.
For this strategy I used the following algorithm, feature set and trade count requirement: