Does This Futures Trading Framework Work in Every Market Regime?

Testing Across 5 Volatility Environments and 18 Years of ES Futures Data

There is a question I needed to answer before I was willing to share anything publicly. In the full scope of that research, I described the data foundation, the verification pipelines, and the scale of the permutation space. But none of that matters if the edge only works in one type of market.

Not "does this work?" That question is necessary, but it is not sufficient. A framework can work in specific conditions and still fail you when conditions change. And conditions always change.

The question I needed to answer was harder: does this work in every market regime?

Not just during the last bull run. Not just in the conditions the framework was developed in. Across every kind of market the modern era has produced.

That question required a different kind of test.

What regime testing actually means

Most retail traders have never heard the term "regime-conditional analysis." That is not a criticism. It is a reflection of how little the trading education space talks about the things that actually matter.

Here is what it means in plain language.

Markets do not behave the same way all the time. A calm Tuesday in 2017 and a circuit-breaker day in March 2020 are both "the ES futures market." But they are completely different environments. The range is different. The speed is different. The way price interacts with levels is different. The risk profile is different.

Regime-conditional analysis means splitting your results by the type of market they occurred in. Instead of one aggregate number across all conditions, you get a separate answer for each environment.

If an edge only appears in one regime, it is not a real edge. It is a coincidence that happened to show up in whatever conditions your testing period contained.

I built the framework to answer this across five distinct volatility regimes, three market direction states, and every major stress period of the last 18 years.

Five volatility environments

I classified every trading session in the 18-year ES dataset into one of five volatility regimes based on realized range relative to historical norms. A session-by-session analysis of the optimal trading hours complements this regime view from a different angle.

Calm. Sessions with compressed ranges, low volume, minimal directional movement. The kind of environment where nothing seems to happen. Many frameworks struggle here because the moves are too small to capture.

Low. Below-average volatility. Not dead calm, but muted. The environment where most mean-reversion approaches quietly do their best work and most momentum strategies slowly bleed.

Normal. Average conditions. The middle of the distribution. This is what most backtests are implicitly tuned to, because it is where most sessions fall.

High. Above-average volatility. Larger ranges, faster moves, wider candles. This is where risk management becomes more important than signal quality.

Extreme. Crisis-level volatility. The days that show up in headlines. Circuit breakers. VIX above 40. The sessions where most retail traders lose months of progress in hours.

The results across these five regimes followed a pattern I did not expect.

The U-shaped curve

I expected the framework to perform best in normal conditions. That is where most of the data lives, and it is reasonable to assume that any framework would be calibrated — consciously or not — to the environment it sees most often.

That is not what happened.

The performance curve across volatility regimes is U-shaped.

The framework performs well in calm conditions. It dips in moderate volatility. Then it gets stronger again in high and extreme volatility.

The strongest results came from the environments most people would expect to be the hardest. Extreme volatility sessions — the days that destroy most traders — produced the highest win rates in the dataset.

I spent time trying to understand why. The explanation is structural, not coincidental.

In calm conditions, price moves slowly and levels act as clear boundaries. There is less noise, less randomness, and interactions with key levels are clean. The framework reads these sessions well because the signal-to-noise ratio is high.

In extreme volatility, the dynamic is different but the outcome is similar. Extreme sessions produce larger moves, but they also produce sharper reactions at key levels. When price reaches a structural level during a volatile session, the reaction tends to be decisive. It either breaks through with conviction or reverses hard. There is less ambiguity.

The middle of the distribution — moderate volatility — is where levels are most likely to produce choppy, indecisive price action. Touches that go nowhere. Breaks that fail. Reversals that reverse again. The framework still works in these conditions. But the edge is narrower.

The important finding is not that any single regime produced a high number. The important finding is that every regime produced a positive result. 89,774 signals across all five environments. All profitable. The edge is real in every condition, not just the convenient ones.

Market direction neutrality

This was the second structural question I needed to answer.

Many trading frameworks have a hidden directional bias. They work well in trending markets and fail in flat ones, or they work well in bull markets and fail in bear markets. The bias is usually invisible in a backtest because most backtesting periods contain more up days than down days.

I split the results by market direction on the session level.

Up days: Profitable.

Down days: Profitable.

Flat days: Profitable — and the highest of the three.

The framework does not need to predict market direction to produce an edge. It works in trending and flat environments. It works when the market goes up and when it goes down.

This is not an accident. It is a consequence of how the framework is designed. The Algorithmic Suite identifies structural levels and measures how price interacts with them. Those interactions are not dependent on whether the day finishes green or red. A level rejection at a key price works the same way in both directions.

Direction neutrality means the framework does not require you to have a view on where the market is going. It is not a trend-following approach and it is not a mean-reversion approach. It is a structural approach, and structure does not care about direction.

Surviving the stress tests

Numbers in aggregate are useful. But the real test of regime resilience is whether a framework survives the specific periods that broke everything else.

I ran the framework through every major stress period in the dataset. Not to cherry-pick the ones that looked good. To find out if any of them broke the edge.

2008: The financial crisis. The most violent sustained move in modern futures history. VIX reached 80. Daily ranges in ES exceeded 100 points on multiple sessions. This is the period that destroys any framework that was only tested in normal conditions. The Algorithmic framework remained profitable through 2008.

2010: The flash crash. ES dropped nearly 60 points in minutes on May 6th, then recovered most of it within the same session. A single-event stress test for any level-based framework. The levels held. The framework held.

2018: The volatility spike. After years of historically low volatility, February 2018 produced one of the sharpest single-week drawdowns in years. The Q4 selloff added another 20% decline. A regime transition from calm to violent, which is exactly the kind of shift that exposes fragile frameworks. The results remained positive.

2020: COVID. The fastest bear market in history. A 35% decline in 23 trading days. Circuit breakers triggered multiple times in a single week. Then a V-shaped recovery that reached new all-time highs within months. This was not one stress test. It was two — the crash and the recovery — compressed into a few months. The framework was profitable through both phases.

2022: The rate hike cycle. The fastest pace of Federal Reserve tightening in 40 years. A 27% bear market. Sustained directional pressure. The highest inflation since 1981. A grinding, methodical decline that lasted most of the year. This is the opposite of a flash crash — slow, persistent, demoralizing. The framework survived it.

Across all of these periods and the full 18-year dataset, the framework produced profitable results in 18 out of 19 calendar years. Every regime. Every crisis. Every recovery. The year-by-year performance breakdown lays out every single calendar year individually, including the one that lost money.

Why this matters more than a win rate

When someone shows you a win rate, ask them one question: in what conditions?

A high win rate in a bull market is not impressive. A high win rate over the last two years is not impressive. A high win rate during the specific regime the framework was built for is not impressive.

What is impressive — what actually matters — is consistency across conditions the framework was not built for.

I did not build the Algorithmic Suite for the 2008 financial crisis. I did not build it for COVID. I did not build it for the 2022 bear market. But the research had to prove it worked in all of them, or it was not ready.

Most indicators are tested during one regime and sold during another. They work in the conditions they were built for and fail in everything else. The vendor has moved on to the next product by the time you notice.

Regime-conditional analysis is the antidote to that. It forces honesty. It forces a framework to prove itself in environments it was never optimized for.

If the edge holds across all five volatility regimes, across up days and down days and flat days, across every crisis and recovery the modern market has produced — that is a different kind of result than a win rate from the last quarter.

What I chose not to share

I want to be direct about one thing.

Not every regime produced the same result. The U-shaped curve means the moderate-volatility environment is the weakest. The edge is still positive. The framework is still profitable. But the margin is thinner.

I chose not to publish the specific numbers for every regime because numbers without context become promises, and promises are not what this is about. The aggregate tells the honest story: profitable across all conditions, with natural variation in how strong the edge is.

If you want to see the full breakdown, the Algorithmic platform gives subscribers access to the tools that make this kind of analysis possible on their own charts, with their own rules.

The real question

The question is not "does this work?" The question is "does this work in the market conditions I am about to trade in?"

Since nobody knows what those conditions will be, the only honest answer is to test across all of them.

That is what regime-conditional analysis does. And that is what 18 years of data across five volatility environments, three market direction states, and every major crisis of the modern era is for.

Not to produce a single impressive number. To produce an answer you can trust regardless of what the market does next.

The Algorithmic Suite

Midnight Grid. Quantum Vision. Turning Points.

Three indicators. One framework. Tested in every market regime the modern era has produced.

Available on your TradingView charts today.

Start Your 7-Day Free Trial

Algorithmic is charting software for decision support on TradingView. It is not financial advice. Trading involves risk. Outcomes depend on your rules, risk management, and execution. Past performance does not guarantee future results.