Fair Value Gaps in Academic Research: The Gap-Fill Evidence, Honestly

What the Academic Gap Literature Actually Shows

A real body of research on price gaps exists. Most of it studies opening gaps — the difference between yesterday's close and today's open in equity markets, driven by overnight news, earnings surprises, or index re-weightings. Some of those gaps do tend to partially close within the session, especially smaller gaps during quieter macro periods. That much is documented.

The problem is the strength of the conclusion. Fill rates in the academic literature vary considerably depending on gap direction, gap size, the asset class, and the prevailing trend. The studies do not converge on a single high fill-rate number that would support a rule like 'this gap must fill.' What the evidence actually supports is closer to: under some conditions, some gaps partially close more often than not. That is a much weaker statement than what circulates in retail SMC communities, and it is worth keeping that gap — pun intended — clearly in mind.

How the SMC FVG Definition Drifts From the Research

The Fair Value Gap as retail traders use it — a three-candle imbalance where the wicks of the first and third candles do not overlap the body of the middle candle — is not what academic gap research measures. Academic studies examine opening gaps between sessions: a price jump caused by news arriving while the market was closed. Those are structurally different from intrabar imbalance zones derived from candle anatomy.

Applying fill statistics from one structure to the other is a category error. There is no published peer-reviewed evidence that three-candle imbalances across 1-Hour, 4-Hour, Daily, or Weekly bars fill at the rates retail SMC educators claim. When you hear 'research backs FVGs,' it is almost always opening-gap studies being borrowed from a different context and presented as if they apply directly to SMC zone trading. They do not map cleanly.

What Our Own Data Shows About SMC Indicators

Across 660,005 backtests covering 903 assets on 1-Hour, 4-Hour, Daily, and Weekly timeframes, we tested every major indicator category including SMC-derived signals. The result: not a single SMC-category indicator beat buy-and-hold across our universe. That is a consistent finding, not a one-off.

The pattern SMC indicators tend to produce is high win rates paired with poor expectancy. The SMC: Liquidity Sweep signal, for example, posts a median win rate of 71.2% — which sounds compelling — but only about 8% of assets actually beat buy-and-hold using it. A strategy that wins frequently on small amounts and loses infrequently on large amounts can look like edge in a highlight reel. Over a full distribution of outcomes, it usually isn't.

None of this means price never reacts to an FVG zone. It does sometimes. The issue is that 'sometimes' does not constitute a reliable, repeatable edge after accounting for realistic transaction costs and the full range of outcomes — and our data reflects that across hundreds of assets.

What Actually Holds Up, by Asset Class

Our backtests identify which indicators genuinely clear buy-and-hold — and the leaders are not SMC tools. In Forex, the Fisher Transform comes out on top across the most assets. In Stocks, Fibonacci Pivots, Projection Bands, and the Intraday Momentum Index rank highest. In Crypto, MA Envelope and Delta Volume Rising (a CVD proxy) lead. In ETFs, QQE leads the list. These results span the same 1-Hour, 4-Hour, Daily, and Weekly timeframes.

The broader picture: only about 26% of all indicator-asset combinations beat buy-and-hold in our tests. That number matters because it establishes the default — most combinations fail — and it means finding something that actually works requires evidence rather than a compelling narrative about institutional order flow. The indicators that do hold up tend to be quantitative and rules-based. Narrative-driven zone concepts have not been among them.

These Are Hypothetical Backtests, Not Advice

Every result on this site, including every figure cited in this article, comes from hypothetical backtests with realistic transaction costs applied. Backtest results are not live trading performance. They do not capture all real-world slippage, broker-specific spread variation, emotional execution errors, or future market regime changes. Nothing here is financial advice. No result is a promise or guarantee of anything.

The purpose of publishing this data is to give you a more honest starting point than you would get from a YouTube strategy video — so you can decide what is worth your time to study and what probably isn't. What you do with that is entirely your decision.