Relative Volume Spikes: Signal or Noise? We Ran the Backtests

What RVOL Is and Why Traders Use It

Relative Volume compares current volume to a historical average for the same period. When volume is running two or three times above its baseline, most traders read it as confirmation: real participation is behind the move, the breakout is legitimate, the setup is worth taking. In short-term trading circles, a volume spike is treated as a prerequisite, not an optional filter.

The appeal is intuitive. If an asset typically trades a certain amount in a given hour and today it's already at triple that level, something is happening. The harder question — the one the backtest exists to answer — is whether that something translates into risk-adjusted profit once spread, commission, and slippage are deducted.

The Test Setup

IndicatorEdge ran 660,005 backtests across 903 assets spanning stocks, ETFs, forex, crypto, commodities, and indices on four timeframes: 1-Hour, 4-Hour, Daily, and Weekly. Every test deducts realistic costs. The bar is Sharpe ratio above buy-and-hold — not raw win rate, not gross return.

Across the full dataset, only 26% of indicator-and-asset combinations cleared that bar. Even the single best indicator for a given asset produced a median Sharpe of 0.62. A test without costs isn't a test — it's a highlight reel.

What Volume Signals Actually Delivered

Relative Volume did not surface as the top-performing indicator for any of the seven asset classes in our results. That's a meaningful finding given how prominently it features in trading education.

Volume-composite indicators showed a similar pattern. Money Flow Index (MFI), which blends volume with price direction, posted a median win rate of 72.2% — a number that looks compelling in any backtest screenshot. But only 9% of assets where MFI was tested actually beat buy-and-hold after costs. Winning 72% of trades while still trailing a passive approach is the volume trap distilled to a single number.

The one volume-adjacent signal that did appear in the winners was Delta Volume Rising (a CVD proxy), which ranked among the top indicators for crypto across 4 assets. CVD is a cumulative measure of whether buyers or sellers are winning the order-flow battle — structurally different from a spike detector. It captures directional pressure over time, not a momentary surge in activity.

Why Spikes Fool You

A volume spike confirms that a lot of trading happened. It does not confirm which direction wins next. After a large spike triggered by news, earnings, or a stop hunt, price frequently consolidates or reverses as the event resolves. Entering on the spike often means entering at peak uncertainty, not at confirmed direction.

There is also a parameter problem. RVOL is defined relative to an average, and that average is sensitive to lookback period. A 10-session baseline and a 30-session baseline produce different signals on identical price bars. Without a fixed, rigorously tested parameterization, you are selecting whatever setting looked best in hindsight — which is overfitting by another name.

Short signals compound the difficulty. Across the full dataset, short-side signals showed an edge only 17.4% of the time. Volume spikes are frequently used to time shorts on failed breakouts, but that is precisely where the data suggests timing is least reliable.

These Are Backtests, Not a Trading System

Everything above reflects hypothetical backtest results on historical data with modeled costs. Real execution differs — fills may be worse, live spreads vary, and no historical edge guarantees future performance. Nothing here is financial advice. What the data provides is an honest baseline: where volume-based signals have historically shown edge, and where they have not. Use it as a calibration tool, not a signal.