What Is a Gann Box?
Michael Whitman
·
Trading since 2017, full-time since 2020
·
YouTube creator, 137K subscribers
·
Published 2026-09-15
·
Updated
Gann box is a rectangle drawn between two points on a chart and subdivided at fixed proportions such as a half, a third and an eighth, producing a grid of horizontal and vertical levels. Every level depends entirely on the two points chosen, so the tool formalizes a judgment rather than measuring the market.
Video: Michael Whitman on YouTube, a tutorial on the Gann box and Gann square. Watch it on YouTube.
Covered on this page: TradingView.
A Gann box divides a price move into fractions and draws lines at each one. What it is not — and this is the part worth getting straight — is a way of discovering where those lines should go.
How it works
Pick two points. Usually a significant low and the high that followed, or the reverse — the box spans that move in both price and time.
The box is then divided at set proportions. Halves, thirds, quarters and eighths are the conventional set, drawn as horizontal lines across price and vertical lines down time.
Those divisions are where attention goes. The halfway line in particular is treated as the most important level in the box.
The dependence nobody states clearly
The two points are a judgment. Which low counts as significant, and which high — reasonable people pick differently on the same chart, and each choice produces a different grid.
So the levels are downstream of the judgment. The box does not locate support; it divides a range you nominated into parts, and calls the parts levels.
Which is exactly what Fibonacci retracement does with a different set of fractions. The two tools share their structure entirely and differ only in which numbers they divide by.
A worked example
Take a stock whose ordinary daily bar spans about $0.50, with busy days reaching $1.10.
Draw a box over a move of about $5. The eighths are then $0.625 apart — wider than an ordinary bar, narrower than a busy one.
So an ordinary bar spans most of the distance between two divisions. A level that a typical bar can cross in one move is not a level in any useful sense — it is a coordinate.
The grid gets meaningful only when the box is large relative to the bars. On a move of $40 the eighths are $5 apart, which is ten ordinary bars, and a reaction at one of them is at least distinguishable from normal movement.
Why it might work at all
Self-fulfilment is the honest mechanism. Halves and quarters of a recent move are watched by enough people, on enough platforms, that orders cluster near them.
Round proportions are a natural way to think. “Half the move back” is how people describe a pullback regardless of what tool they use, so the halfway line has attention on it whether or not anyone drew a box.
And the tool enforces a discipline. Marking levels before the move rather than after is a real improvement on the habit of explaining a chart in hindsight.
What it is not is a property of the market. There is no mechanism by which a market knows where you put a rectangle, and any explanation that implies one is describing a belief rather than a process.
The original data
Put a box over a real move: the S&P 500’s fall from 1,565.15 on 9 Oct 2007 to 676.53 on 9 Mar 2009, a decline of 56.8%. The box is 888.62 points tall, so its eighths sit about 111 points apart and its halfway line lands near 1,120.84. Where price later reacted inside that grid is visible only in hindsight, which is the tool’s standing problem.
The size of the move against the size of an ordinary bar sets the smallest division worth drawing. A grid whose spacing is below a normal day’s range produces levels that ordinary movement crosses constantly, which is indistinguishable from no levels.
And each level acted on costs the spread. A dense grid produces many touches, each one a decision, and since every round trip pays the spread and any commission, the arithmetic punishes treating every line as tradeable.
How to use one without fooling yourself
Draw it once, from the two most obvious points. If you have to hunt for the pivots that make the grid fit, the grid is fitting the past rather than framing the future.
Treat the halfway line as the only one that matters. It has the most attention and the clearest rationale; the eighths are mostly clutter.
Require confirmation at the level, not arrival at it. Price reaching a line is not information; price reacting there is at least an observation.
And leave it in place once drawn. Redrawing until levels line up with what already happened is the single most common way this tool is misused, and it converts a planning aid into a hindsight generator.
Who Gann was, and why it matters here
W. D. Gann was a trader and author working in the first half of the twentieth century, who published a body of methods relating price movement to geometry and to time.
The claims made about his own results are not verifiable. Figures circulate about his trading record that trace to promotional material of the period rather than to audited accounts, and repeating them as fact is exactly the sort of unverified performance claim this site does not publish.
Which leaves the tools to stand on their own. A box that divides a move into halves and quarters is useful or not on its merits, regardless of what its originator did or did not earn.
Judge them as chart organization. They impose structure on a move and force levels to be marked in advance - both genuinely helpful habits - and neither depends on the market obeying any geometry.
How this connects
One of three Gann tools that share an assumption: that price and time move in fixed proportion.
The box divides a range into equal parts - which is arithmetic anybody can do, and the Fibonacci ratios are often overlaid on the same divisions.
The proportion has to be chosen. Gann fixed his charts explicitly, so many points to so many days; modern platforms rarely do, which makes the drawing depend on your scaling.
The gann fan and gann square apply the same idea with different geometry, and all three move if the chart is rescaled.
And a level you cannot write down without also specifying your screen is a weak basis for an order.
When it fails
The characteristic failure is redrawing. Price ignores the grid, so the box is adjusted to a different pair of pivots, and the new grid explains recent action neatly. It will — any two points produce a grid, and a grid fitted after the fact always describes what happened. The tool has been quietly converted from a forecast into a narrative, and because the levels still look principled, the change is hard to notice from inside.
A second failure is drawing divisions finer than the bar range, producing levels ordinary movement crosses constantly.
A third is treating the grid as a market property rather than a division of a move you chose.
A fourth is running several boxes at once, which covers the chart in lines and makes any outcome explicable.
And a fifth is expecting precision. These are regions where attention clusters, not prices at which anything is obliged to happen.
Related
Gann fan covers the angular version of the same idea. Gann square covers the attempt to relate price to time directly. And Fibonacci retracement covers the same construction with different fractions.
I use these and I am precise about what they are. A Gann box does not find levels, it divides a move you selected into fractions. That can be genuinely useful for organizing a chart, and it is not a measurement of anything the market did on its own.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money. Some links on this page earn a commission if you buy through them. It costs you nothing and it does not decide what appears here or in what order — how these pages are made is set out in our methodology.