Automated Trading Portfolio Filters Ranked: From $319K to $696K (2025 Forward Test)
Here’s what I learned from a year of forward testing, and how I'm going to apply it in Q1
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. 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 on a quest for the holy grail of automated trading. Questions? Check the FAQs or feel free to reach out directly: AutomatedTradingStrategies@protonmail.com.
Last week I shared updates with you on the:
VPS experiment
LLM-created strategy performance
New model for strategy alerts that will start in January.
Forward Test Completion: 75% complete; resuming in spring.
I’m currently building the Q1 2026 Forward Test, preparing for a Live Test starting in January, and developing an army of Automated Portfolio Managers (APMs) that require performance logic. All three benefit from a rigorous analysis of this year’s Forward Test results. For those who are new to automated trading, a forward test uses live data on simulated accounts. A backtest uses historical data on a simulated trading engine, so there’s a lot that can go wrong.
This is an analysis of 2025’s YTD Forward Test data. Preliminary analysis identified a portfolio of Core Strategies that performed fairly well over the year, but we can do better.
Structural vs Rotational Filters
Individual automated trading strategies fail—that’s inevitable. Portfolios diversify that risk, but they can’t protect you from regime shifts that hit everything at once. That’s where portfolio filters come in. They’re not about picking better strategies; they’re about knowing when to step back from all of them.
Strategy selection asks which one? Portfolio filtering asks should I be trading at all?
Over the last quarter we’ve examined numerous portfolio filters: time of day, day of week, session-based, momentum, HMM states, and contrarian logic. Today, I’m sharing how these filters performed over the entire year and how I’ll be changing the Forward Test portfolio in Q1 of 2026 based on what I’ve learned.
First, let me give you the headlines, then we’ll dig into the details.
Filters fall into two categories: structural and rotational. Structural filters are tied to institutional money flows and remain persistent across different market regimes. Rotational filters (like day-of-week) flip with algorithmic discovery and crowding.
The Q4 drawdown was a regime shift, not a calendar effect. Using Hidden Markov Model analysis, I discovered that 76% of Q4 days were in a DANGER state versus only 14% pre-Q4. The market regime changed. The calendar didn’t.
Overnight entries dominate. Trades entered between 12-3am CST produced $417/trade—the highest of any session.
One filter stands above all others. Here’s a comparison of the baseline Forward Test performance which is currently up $319K YTD (gray), compared to the use of Calendar based filters (blue), HMM-based filters (pink), and Momentum based filters (green). Momentum beats.