The bell curve is the quiet foundation under almost everything in volume analysis. When a market is in balance, its volume profile forms a normal distribution, heavy in the middle and thin at the edges, and once you see why, the value area, standard-deviation bands and mean-reversion trades all stop being arbitrary conventions and start making sense. It is also, not by accident, the shape in this site’s logo.
What a normal distribution is
A normal distribution, or Gaussian distribution, is the classic bell-shaped curve. Most observations cluster around a central average, and the further you move from that center in either direction, the fewer observations you find, symmetrically. Statisticians measure that spread in standard deviations:
- Roughly 68% of the data falls within one standard deviation of the mean.
- Roughly 95% falls within two standard deviations.
- Roughly 99.7% falls within three.
That is the entire idea. A central tendency, symmetric tails, and a predictable share of activity inside each band of spread.
Why volume profiles form a bell curve
Turn a balanced trading session into a volume profile and it takes on that same bell shape, rotated onto the price axis. The reason is the auction itself.
A market spends most of its time near the price where buyers and sellers agree, the fair price, because that is where business gets done. Move away from fair value and one side gets an incentive to push back: too high and sellers step in, too low and buyers do. Price is drawn to the center and repelled from the edges. So volume piles up in the middle and thins toward the extremes, exactly like observations clustering around a mean.
This is why a balanced day prints a D-shaped profile: a fat, symmetric bulge with a point of control near the center and thin tails top and bottom. The market found a fair price and rotated around it. When one side dominates instead, the curve skews into a P-shaped or b-shaped profile, the lopsided cousins of the balanced bell.
The value area is one standard deviation
The single most practical consequence of all this is the definition of the value area. By convention, the value area is the band of prices holding 70% of the session’s volume, centered on the point of control.
That 70% is not a random pick. It is essentially the first standard deviation of the profile’s distribution, the 68% core of the bell curve rounded to a working number. So the value area is literally the statistical heart of the auction: the prices the market genuinely accepted as fair. Everything outside it sits in the thinner tails, where the market spent little time and found less agreement.
This is what gives the value area high (VAH) and value area low (VAL) their power as reaction levels. They mark the edge of the accepted range, one standard deviation out, where price is stretched and the auction tends to pull it back. The full mechanics of trading those edges live in the volume profile hub.
The same statistics show up on the VWAP side of the toolkit, where VWAP bands are drawn at one, two and three standard deviations from the volume-weighted average, giving you the same bell-curve logic around a moving line instead of a static profile.
Trading the balance
The normal distribution is a mean-reversion framework, and it applies whenever the market is balanced.
In a clean D-shaped session, price over-extends to the VAH or VAL, into the thin tail, and then reverts toward the point of control at the center. That is the auction rejecting a price it considers unfair and returning to value. The trade is to fade the edge back toward the middle: sell tests of the VAH, buy tests of the VAL, targeting the POC.
An ES example. The session builds a symmetric bell with the point of control at 5,390, VAH at 5,404 and VAL at 5,376. Price pushes to 5,405, into the upper tail, and the footprint chart shows aggressive buyers getting absorbed with no follow-through. That is the edge of the distribution rejecting, a short back toward 5,390 with a stop above 5,408.
The key discipline is knowing when the framework applies and when it does not.
When the bell curve breaks
Markets are not always normally distributed, and this is where traders get hurt. On a trend day, price does not rotate around a center; it moves in one direction all session, and the profile stretches into a long thin shape with no meaningful central bulge. Fading the “edges” of a trend day means fighting the move, and the mean reversion never comes.
Real markets also have fat tails: extreme moves happen more often than a pure Gaussian predicts. News, liquidation cascades and gaps produce jumps that a textbook bell curve would call nearly impossible. That is why the normal distribution is a model for balanced conditions, not a law. Before you trade the bell, confirm the market is actually balancing, a symmetric developing profile, a stable point of control, price rotating rather than trending. When the distribution skews or stretches, switch playbooks.
The tell that balance is turning into trend is often the point of control itself. A POC that keeps migrating session after session says value is moving and the market is no longer rotating around a fixed center. For the wider framework that ties distribution, profile and order flow together, see the order flow trading guide.
Frequently Asked Questions
Why does a volume profile look like a bell curve?
Because a balanced market spends most of its time near fair value, where buyers and sellers agree, and only brief moments at the extremes, where one side pushes back. Volume therefore piles up in the middle and thins toward the edges, producing the symmetric bell of a normal distribution. This balanced shape is called a D-shaped profile.
How does the normal distribution relate to the value area?
The value area is the band of prices containing 70% of the session’s volume, centered on the point of control. That 70% corresponds to roughly the first standard deviation of the profile’s distribution, the 68% core of the bell curve. So the value area is the statistical heart of the auction, the prices the market accepted as fair, with everything outside it sitting in the thinner tails.
Can I always assume the market is normally distributed?
No. The normal distribution models balanced conditions only. On trend days price moves one direction all session and the profile stretches into a long thin shape with no central bulge, so mean-reversion trades fail. Markets also have fat tails, meaning extreme moves happen more often than a pure bell curve predicts. Confirm the market is balancing before trading the distribution.
How do I trade a normal distribution profile?
Trade it as mean reversion while the market is balanced. In a symmetric D-shaped session, fade the extremes back toward the center: sell tests of the value area high, buy tests of the value area low, targeting the point of control. Confirm the edge is rejecting with order flow, such as absorption on the footprint, and keep a stop just beyond the edge in case balance breaks into a trend.