Structural Limits of OHLCV-Based Intraday Signals in MNQ Futures: A Systematic Falsification Study

This systematic falsification study demonstrates that no OHLCV-based intraday momentum signals for MNQ futures produce a statistically significant, cost-adjusted trading edge under strict institutional criteria, revealing the structural limits of such strategies despite the presence of isolated, sample-size-deficient anomalies.

The Gist

Imagine you are a treasure hunter looking for gold in a very crowded, high-tech mine called the MNQ Futures Market. This mine is so popular that thousands of professional diggers (institutions) and amateur hobbyists (retail traders) are all searching the same spots every day.

This paper is a report by an independent researcher named Mathias Mesfin. Instead of trying to find a "magic shovel" that guarantees gold, he set out to prove that most of the common shovels people use are actually useless in this specific mine.

Here is the breakdown of his findings in simple terms:

1. The Mission: Testing the "Common Shovels"

For years, traders have believed that if they look at a simple chart showing the Open, High, Low, Close, and Volume (OHLCV) of the market every 5 minutes, they can predict which way the price will go next. They use patterns like:

• "The Morning Rush": If the price breaks a certain level right when the market opens, it will keep going that way.
• "The Gap": If the market opens much higher or lower than yesterday, it will either keep going that way or snap back.
• "The Volume Spike": If a lot of people trade suddenly, the price will keep moving in that direction.

Mesfin took 14 different versions of these common ideas and tested them rigorously over 4 years (2021–2025) using real data.

2. The Rules of the Game (The "Real World" Test)

To make sure the results were fair, he set strict rules, like a referee in a sports match:

• The Cost of Digging: Every time you make a trade, you pay a "fee" (like a toll road). He assumed a cost of 2 points per trade (covering fees and the difference between buying and selling prices).
• No Cheating: You can't use information from the future to make a decision today.
• The "30-Trade" Rule: You need to dig at least 30 times to prove you aren't just lucky.
• The "Statistical Proof": You need to be statistically sure you aren't just guessing (a score of 2.0 or higher).

3. The Big Discovery: The "Gold Ceiling"

After testing all 14 strategies, Mesfin found a hard limit.

• The Gross Edge: Before paying the "fee," the best any of these simple chart patterns could do was earn about 0.07 to 1.50 points per trade.
• The Reality Check: Since the fee to trade is 2 points, these strategies actually lose money every time you use them.

The Analogy: Imagine you find a vending machine that gives you a candy bar worth $1.50. But, the machine charges you a $2.00 fee to press the button. No matter how many times you press it, you will always lose 50 cents. Mesfin found that these 14 trading strategies are exactly like that vending machine.

4. The "Fake" Winners and the "Real" Winners

• The One Almost-Winner: One strategy (betting that a "gap down" day will keep going down) looked amazing, earning huge profits. However, it only happened 22 times in 4 years. Because it didn't happen enough times (the rule was 30), it couldn't be trusted as a reliable method. It was like finding a rare, lucky coin flip that worked once, but you can't build a business on it.
• The "Positive Controls" (Proof the Test Works): To prove his test wasn't broken, he included two strategies that did work. The key thing about these isn't that they are "more complex" — it's that they look at a fundamentally DIFFERENT kind of signal on a LONGER time horizon. Instead of trying to read a single 5-minute candle (open/high/low/close/volume) and predict the next 5 minutes, they classify the broader "market regime" — what kind of state the market is in right now — and then hold the position across many bars (roughly an hour). They are LONG-HORIZON, STRUCTURAL signals; the strategies in the main study are SHORT-HORIZON OHLCV signals. This proves that if a real edge exists at the long-horizon structural level, his test would catch it; the main study's null result is specifically about short-horizon OHLCV patterns, not all of trading.

5. Why Does This Happen?

The paper suggests that the MNQ market is like a super-efficient vacuum cleaner.
Because the market is so liquid and watched by so many smart computers, any simple pattern (like "if the price goes up, it keeps going up") is spotted and exploited instantly. The "smart money" buys or sells so fast that by the time a regular trader sees the pattern and tries to act on it, the profit has already been taken away. The only profit left is smaller than the cost of trading.

6. The Conclusion

The main takeaway is a "Null Result" (a finding that says "nothing works here").

• Simple chart patterns (looking at 5-minute bars) cannot beat the fees in this market.
• Most retail traders who think they have a "secret formula" based on simple charts are likely falling victim to luck or bad math.
• To actually make money, you likely need to look at a different kind of signal — longer-horizon, structural ones (like the regime detectors above) — or data that regular people don't have. Squeezing more juice out of short-horizon OHLCV patterns alone probably won't do it.

In short: The paper is a systematic "falsification" study. It didn't try to find a winning strategy; it tried to prove that the popular ones don't work. And after testing 14 of them, it successfully proved that for simple 5-minute charts, the market is too efficient to be beaten after costs.

Technical Summary

Technical Summary: Structural Limits of OHLCV-Based Intraday Signals in MNQ Futures

Problem Statement
The paper addresses a critical gap in trading literature: the lack of rigorous documentation regarding strategies that fail. While retail and academic communities frequently publish successful intraday momentum strategies based on Open-High-Low-Close-Volume (OHLCV) data, there is a systematic absence of evidence regarding the vast population of strategies that do not work. The central hypothesis tested is whether short-term price patterns in Micro E-Mini Nasdaq 100 (MNQ) futures contain exploitable directional information that survives realistic execution constraints, specifically a two-point round-trip friction cost.

Methodology
The study employs a systematic falsification framework applied to 947 regular trading hours (RTH) days of MNQ data from December 2021 through August 2025. The methodology is characterized by the following strict protocols:

• Signal Universe: Fourteen distinct signal families were tested, including Opening Range Breakouts (ORB), Asia session expansion and liquidity grab reversals, gap fill/continuation strategies, volume signature signals, volatility-regime classifiers, and news-driven directional strategies.
• Execution Assumptions: All signals are computed at bar close with entries executed at the next bar open. This eliminates same-bar fill bias. A fixed two-point round-trip friction cost (approx. $4.00 per micro contract) is applied to all trades to account for spread, fees, and conservative slippage.
• Validation Structure: The study utilizes expanding-window walk-forward validation (training on prior years, testing on the subsequent year) to prevent look-ahead bias and overfitting.
• Pass Criteria: A signal is only considered valid if it simultaneously satisfies five conservative criteria:
  1. Out-of-sample (OOS) T-statistic $\ge$ 2.0.
  2. Minimum of 30 OOS trades.
  3. Positive net return after the two-point friction deduction.
  4. Consistency across all tested years (2022–2024).
  5. Permutation test significance ($p < 0.05$) where applicable.

Key Results
The study yielded a definitive null result for the primary signal universe:

• Systematic Failure: None of the fourteen OHLCV-based signal families satisfied all five validation criteria simultaneously.
• The Gross Edge Ceiling: The maximum achievable gross return for next-bar-open execution across all signal families was found to be between 0.07 and 1.50 points. This ceiling is structurally insufficient to clear the two-point friction cost, resulting in negative net returns for every tested variant.
• Specific Signal Performance:
  • ORB and Gap Strategies: Most variants failed with T-statistics near zero or negative. The "Gap Fill Fade" hypothesis was rejected, as gaps did not consistently fill.
  • Volume Signals: Both volume spike momentum and volume dry-up exhaustion hypotheses produced T-statistics indistinguishable from noise, confirming that bar-level volume magnitude does not predict next-bar direction.
  • Asia Session Signals: Expansion bars and liquidity grabs produced statistically significant negative T-statistics (e.g., $T = -10.96$ for Asia expansion), indicating that the directional move is often exhausted within the signal bar itself, making next-bar entry unprofitable.
  • The "Gap Continuation Short" Anomaly: This signal achieved a T-statistic of 3.23 and a mean net return of +14.52 points. However, it generated only 22 trades over three years, failing the minimum sample size threshold (N $\ge$ 30) and thus remaining unvalidated for deployment.
• Positive Controls: To confirm the methodology was not overly restrictive, two pre-existing signals were tested as controls. The "RTH Confluence Signal" (T = 5.83) and "London Session Signal B" (T = 5.15) both passed all criteria. These signals rely on regime classification and multi-bar hold periods (12–15 bars), distinguishing them from the single-bar directional signals tested in the main study.

Significance and Claims
The paper's primary contribution is methodological rather than the discovery of a new profitable strategy. It establishes a reproducible, documented falsification framework that prioritizes the reporting of negative results to counteract survivorship bias.

The study claims that for liquid equity index futures like MNQ, the competitive market structure has arbitrated away any gross edge derived from publicly observable OHLCV patterns at the five-minute resolution. The "edge ceiling" of ~1.50 points suggests that the marginal cost of trading (friction) exceeds the available alpha for single-bar directional predictions.

The author concludes that genuine intraday edge in this market likely requires either:

1. Proprietary data (e.g., tick-level order flow) not available in standard OHLCV bars.
2. Structural regime detection (as seen in the positive controls) that captures multi-bar market state transitions.
3. Execution advantages unavailable to retail participants.

The paper asserts that the consistent failure of retail OHLCV strategies under these rigorous standards is not a failure of the testing methodology, but an informative reflection of the market's efficiency regarding public price and volume data.