Gann analysis: angles, squaring and time cycles
The answer
Gann analysis draws geometric angles and squares on a price chart and treats certain price levels and dates as significant because of their numerical relationship to a starting point. There is no accepted mechanism by which geometry governs prices, the angles change when you change the chart's scale, much of the published material cannot be proved wrong, and the well-known claims about Gann's own trading returns are not documented.
Why this costs you money
The financial cost of this subject arrives in 3 forms, and the third one is the largest.
You pay for the material. Gann courses and software are among the most expensive products in retail trading education. His original course was sold at a price that was very large for its era. The modern equivalents are sold the same way, with the same argument: the method is secret, difficult, and reserved for the serious.
You lose the time. Gann material is deep. There are angles, squares, the Square of 9, the hexagon chart, the circle of 360, planetary cycles, master charts and anniversary dates. A determined reader can spend 2 years learning it. Those 2 years are not free. They are 2 years not spent on position sizing, costs, tax treatment and the arithmetic of drawdown, all of which are boring and all of which are the things that actually decide whether an account survives.
And you acquire a false sense of precision. This is the expensive one. A Gann projection gives you a price to 2 decimal places and a date to the day. That specificity feels like knowledge. It produces large positions taken with high confidence at a named level, and the confidence is not supported by anything. Precision and accuracy are different properties. A method can produce an exact number that is exactly wrong, and a number that looks exact is far more dangerous than a vague opinion, because you will size a position against it.
How it works
William Delbert Gann was born in Lufkin, Texas in 1878 and died in 1955. He traded shares and commodities, published market letters and annual forecasts, and sold courses. His books include Truth of the Stock Tape in 1923, The Tunnel Thru the Air in 1927, How to Make Profits in Commodities in 1941 and 45 Years in Wall Street in 1949.
His central premise was that price and time are related, and that markets move in proportion to geometric and numerical relationships that repeat.
Gann angles
A Gann angle is a line drawn from a significant high or low at a fixed rate of price change per unit of time.
The most important is the 1x1, meaning 1 unit of price for 1 unit of time. Others form a fan: 2x1, 3x1, 4x1 and 8x1 are steeper, and 1x2, 1x3, 1x4 and 1x8 are shallower.
The interpretation is that price above the 1x1 is strong, price below it is weak, and each line of the fan acts as support or resistance. When price breaks one line, it is said to move to the next.
Here is the problem, and it is not a matter of opinion. The 1x1 is usually described as a 45-degree line. A 45-degree line on a screen depends entirely on how many rupees or dollars of price occupy the same distance as 1 day of time. That ratio is set by your chart software, by the height of your window, by the zoom level and by whether the price axis is arithmetic or logarithmic.
Change any of those and the 45-degree line points somewhere else. The same chart on a phone and on a wide monitor produces 2 different 1x1 lines from the same starting point.
Practitioners answer this by fixing the scale. You decide that 1 unit of time equals a specific price increment, for example 1 point per day, and you draw the 1x1 at that rate. That answer is coherent and it moves the problem rather than solving it, because the choice of increment is now the input that decides the output, and there is no rule for choosing it. Different Gann analysts use different scales on the same instrument, and they get different levels.
Squaring price and time
Gann held that a trend changes when price and time come into balance, which he called squaring.
The simplest version: if a market rises 120 points, watch for a change 120 units of time later. Units may be days, weeks or months. A second version squares a high or a low with the calendar, so a market that topped at 360 is watched at 360 days from the top.
A third version uses anniversary dates. The 1-year, 2-year and 5-year anniversaries of a major high or low are treated as dates where a change is more likely, and the fractions of a year — 30, 45, 60, 90, 120 and 180 days — are watched too.
The Square of 9
The Square of 9 is a spiral of numbers arranged outward from 1, in which each complete rotation increases the value. Numbers that fall on the same ray from the centre are said to be related. Because the spiral is built so that the squares of odd numbers sit on 1 diagonal, moving 360 degrees around the spiral from a number n takes you to roughly the next value whose square root is 1 greater.
In practice a user takes the square root of a price, adds or subtracts a fraction of 1 corresponding to 45, 90, 120 or 180 degrees, and squares the result to get a target. Adding 0.25 to the square root and squaring corresponds to 90 degrees.
This is an arithmetic operation, and it works. What is not established is why the number it produces should matter to buyers and sellers.
Time cycles
Gann used cycles of many lengths, including 10 years, 20 years, 30 years, 60 years, 90 years and shorter counts of 30, 45, 60, 90 and 144 units. Later material, including Nature's Law style writing and the astrological sections of his work, tied market turns to planetary positions.
That last part is stated openly in the source material and there is no reason to be shy about it. Gann's later published work includes explicit astrological content, and The Tunnel Thru the Air is a novel that its author said contained his most valuable secret in coded form.
The part of Gann that is ordinary and defensible
Set the geometry aside and a real technician remains.
Gann published lists of trading rules — commonly given as 24 rules, with a different set of 28 in another book. They include instructions that any modern risk manual would repeat: use stop orders on every trade, never overtrade, never average down on a losing position, trade only active markets, avoid taking a profit that is smaller than the loss you risked, and do not trade merely because you are impatient.
None of that requires an angle. It is competent, conventional risk discipline written down clearly in the 1940s. If Gann made money, this is a far more likely explanation than the Square of 9, and it is the part of his work that a reader should keep.
What it tells you, and what it does not
Gann tools tell you where round numbers, anniversaries and fixed rates of change fall on your chart. That is a real thing to know, and a small part of it has an ordinary explanation.
Round numbers do attract orders, because people place limit orders and stops at round prices. Anniversaries of a major low do coincide with reporting cycles, index rebalancing and annual portfolio decisions. A fixed rate of price change per day is just a trendline with a slope you chose in advance. Each of those effects is small, real and available without any of the geometry.
The tools do not tell you that geometry governs price. There is no mechanism. Buyers and sellers do not know where your 1x1 line is, and the level is different for every analyst because every analyst chose a different scale.
They do not give you a repeatable measurement, because 2 competent users get different answers from the same chart. A measurement that depends on the operator's arbitrary choice is not a measurement.
And they do not have an evidence base. I know of no peer-reviewed test of Gann angles or the Square of 9 in a serious journal. That is unusual even by the standards of technical analysis, where head and shoulders patterns and moving average rules have both been tested many times. The absence is not proof of failure. It does mean nobody claiming this method works can point you to anything.
The decision rule
Treat every Gann level as a line somebody drew, not as a level the market knows about.
If you want to use a Gann angle, first write down the price-per-unit-of-time scale you chose and why. Then draw the same angle from 2 other significant pivots on the same chart. If the 3 projections are far apart, you have learned what the tool actually gives you, which is a family of possible levels rather than a level.
Unless a Gann level coincides with something that has an independent reason to matter — a prior high, a prior low, a round number that many orders sit at, a moving average many participants watch. In that case act on the independent reason, and note that the angle added nothing.
Never size a position on the precision of a Gann number. The decimal places are produced by arithmetic, not by evidence.
Try this now
Five minutes. This is the test that settles the question for most readers, because it uses their own chart and needs no argument.
- Open a daily chart of an index or a liquid share you follow, over about 2 years. Find your platform's Gann fan tool. Most free charting platforms have one.
- Find the lowest low on the chart. Anchor a Gann fan there. Note where the 1x1 line sits today. Write down that price.
- Now find a second significant low — a clear swing low several months later or earlier, the kind any analyst would mark. Anchor a second Gann fan there. Write down where its 1x1 sits today.
- Do it a third time from a significant high, drawing the fan downward. Write down the third number.
- Compare your 3 numbers. Calculate the distance between the highest and the lowest, as a percentage of the current price.
- Now change 1 setting and nothing else. Switch the price axis from arithmetic to logarithmic, or resize the chart window so it is half as tall. Look at your first fan again. Write down where the 1x1 sits now.
What you should see. Steps 2 to 5 usually produce 3 projections spread across a range of 10% to 30% of the current price, and sometimes much more. That spread is the tool's actual output. It is not a level. It is a wide band, and the narrower number you would have quoted came from your choice of starting pivot, which was a choice, not a fact.
Step 6 is the one people remember. The 1x1 line moves when you change the axis or the window size, because a 45-degree angle is a property of the picture and not of the prices. Nothing about the market changed between the 2 screenshots. If a level moves when you resize a window, the market cannot be respecting it.
A fair note before you conclude. A defender will say you should have fixed the price-per-day scale first, and they are right that this is the correct procedure. So do that too: pick a scale, write it down, and repeat steps 2 to 4. The 3 starting pivots will still give you 3 different projections, because the choice of anchor is a second free parameter and fixing the first one does not remove it.
Three real cases
1. The Ticker and Investment Digest interview, December 1909 — the founding performance claim, and its source The best-known claim about Gann's trading comes from an article in The Ticker and Investment Digest, a magazine edited by Richard Wyckoff, in December 1909. It reported that Gann, observed by a representative of the magazine, made 286 trades over 25 market days in October 1909, of which 264 were profitable, a success rate of about 92%, and that he multiplied his capital roughly 10 times.
Consider what this source is. It is a promotional article in a trade magazine published in 1909, describing trades observed by 1 person, with no broker statements, no independent audit and no ability for anybody to verify it now. It is the single most repeated fact about Gann and it has never been confirmed.
That does not prove it false. It means it carries no evidentiary weight, and anybody offering it as proof that the method works is offering a magazine article from more than a century ago.
2. Alexander Elder's account of Gann's estate — the claim about the fortune, and what was actually reported Alexander Elder wrote in Trading for a Living, published in 1993, that he had looked into Gann's record. He reported that Gann's son, John L. Gann, who worked as a broker, said his father had not been able to support the family by trading and made his living selling courses and instructional material. Elder reported that a person who knew the estate said it was worth approximately $100,000 including the house, at Gann's death in 1955.
Set that against the figure of $50 million that circulates in Gann marketing material. The 2 claims cannot both be right, and only 1 of them has a named source who spoke on the record.
The reason this case belongs in an article about method rather than about a man is simple. The commercial case for learning Gann rests almost entirely on the assertion that Gann himself became extremely rich by using it. If that assertion is unsupported, the commercial case has nothing under it, and what remains is the method's own evidence, which is close to none.
3. The Tunnel Thru the Air, 1927 — how a retrospective prophecy is manufactured Gann published a novel in 1927 titled The Tunnel Thru the Air, or Looking Back from 1940. He stated that it contained his most valuable discovery in concealed form. The book contains a war between the United States and Japan.
Followers cite this as a prediction of the attack on Pearl Harbor in December 1941, and cite passages in the book as forecasts of specific market movements .
This is the clearest available example of a general problem. A long, ambiguous text produces many statements. After an event, readers search the text and find the statements that fit. The statements that did not fit are not counted, because nobody is looking for them. The result feels like prophecy and is produced entirely by selection.
The test that would settle it was available and was never run: write down, before an event, the specific passage and the specific prediction, and record it. Nobody did. A prediction identified after the outcome is not a prediction, and this applies with equal force to a novel from 1927 and to a wave count published yesterday.
The question that resolves it
A novice asks: where is the Gann level?
An expert asks: what did I have to choose in order to get that level, and what would the level be if I had chosen differently?
Every Gann projection has at least 2 free choices inside it: the anchor pivot and the price-per-unit-of-time scale. Some methods add a third by letting you pick which of 9 angles to emphasise. A person who lists the choices they made and the range of answers those choices produce is using the tool honestly. A person who quotes 1 number is presenting a choice as a measurement.
This question is not special to Gann. Ask it of any tool with free parameters, including moving averages and Fibonacci retracements. Gann is where it bites hardest because it has the most parameters and the least evidence.
What would make this wrong
This article says Gann analysis has no established mechanism and no evidence base. Both statements can be overturned and here is exactly how.
The evidence half would be wrong if somebody specified a Gann rule completely — the anchor selection rule, the scale rule, which angle, what constitutes a touch, what the entry and exit are — and then tested it across many instruments and many periods, out of sample, after costs, reporting all results including the failures. That study is entirely possible. The data is free, the rules can be programmed, and any believer with a laptop could do it in a month. The fact that a century of enthusiasm has not produced one is a piece of evidence in itself, but it is not proof. If such a study appeared with a positive result, this section is where the article would have to change.
The mechanism half would be wrong if somebody described why market participants would collectively respond to a geometric relationship they cannot observe and do not agree on. A mechanism does exist for round numbers, because orders genuinely cluster there and you can see the clustering in order book data. No comparable mechanism has been offered for angles or for the Square of 9.
The strongest defence of Gann, stated fairly. A disciplined trader using Gann levels as arbitrary but consistent decision points, with a fixed stop and a fixed size, may well do better than an undisciplined trader with a better method. That is a real effect and it is not a joke. Structure improves behaviour even when the structure is arbitrary. But notice what it concedes: the benefit comes from having a rule, not from the rule being true. Any consistent set of levels would deliver the same behavioural benefit at a much lower price.
And the honest limit of this criticism. It would be wrong to say every Gann-influenced idea is worthless. Anniversary dates overlap with real annual cycles in reporting and allocation. Round numbers do attract orders. A fixed rate of advance is a trendline. Those pieces survive. They survive without the geometry, and they are available free.
In India
Gann material has a large and active following in India, and the market conditions here make it worth being specific.
The teaching is heavily commercial and often unregistered. Gann courses, seminars and software are sold widely in India, frequently at prices between tens of thousands and several lakhs of rupees. Many are sold by people who are not registered with SEBI in any capacity.
The rule to know is straightforward. A person who gives advice on specific securities to clients for consideration generally requires registration as an Investment Adviser under the SEBI (Investment Advisers) Regulations, 2013, and a person who issues research reports or makes recommendations generally requires registration as a Research Analyst under the SEBI (Research Analysts) Regulations, 2014. Teaching a method as education is different from advising on securities, and the boundary is genuinely contested. What is not contested is that nobody may promise or guarantee a return.
Price bands and the geometry do not sit well together. A Gann projection gives a specific price. An individual Indian share may be prevented from reaching that price on a given day by its daily price band. The projection cannot account for an administrative limit, and a backtest that ignores bands will record fills that could not have occurred.
The data history is short for long cycles. Gann's longer cycles run to 30, 60 and 90 years. The Sensex series begins with a 1978-79 base and the Nifty 50 with a 1995 base. A 60-year cycle applied to an index with under 50 years of data is being applied to something that does not exist in the record.
Anniversary dates land on a different calendar. The Indian financial year runs from April to March. Advance tax dates, quarterly results, index rebalancing and the Union Budget in early February all create genuine annual patterns. A practitioner using American anniversary conventions on an Indian instrument is applying the wrong calendar to a real effect.
In the United States
The United States is where the material originated and where the longest test of it has been available.
The data is long, clean and free. American commodity and equity series run back well over a century. Gann worked on wheat, cotton, corn and the Dow. Every one of those series is available to anybody who wants to test a rule. This is the best conditions any market method has ever had for verification, and after approximately 100 years there is still no published test.
The commercial industry is old and well established. Gann courses, software packages and seminars have been sold continuously in the United States since his lifetime. Original Gann materials trade as collectibles.
The regulatory position separates 2 activities. Selling educational material about a method is generally permitted. Giving personalised investment advice for compensation generally brings a person under the Investment Advisers Act of 1940, with registration required at the federal or state level depending on size . Publishing a general-circulation newsletter has a long-standing exclusion following Lowe v. SEC, 472 U.S. 181 (1985). The Commodity Futures Trading Commission and the National Futures Association regulate commodity trading advisers separately, and promotional claims about trading system performance are covered by rules requiring specific disclosures about hypothetical results.
Performance claims are the enforced boundary. American enforcement in this area concentrates on false or unsubstantiated claims of profitability rather than on the content of the method. A seller may teach geometry. A seller may not claim a documented return they cannot document.
Where they differ, and what that tells you
The method travels unchanged. The protection around the sale of it does not.
The United States regulates the claim. India regulates the person. American enforcement generally asks whether a specific performance statement was true and substantiated. Indian regulation adds a prior question: is this person registered to be giving this advice at all, and if not, the activity itself is the violation regardless of whether the advice was good. For a reader that produces different first steps. In the United States, ask what evidence supports the claim. In India, ask that too, and first check the registration list.
Scale changes the size of the problem. India has a very large and rapidly grown retail audience reached through video platforms and messaging groups, learning from material that is often paid, often anonymous and often unregistered. The same content in the United States reaches a smaller retail audience with more competing free material and a longer institutional memory of the same claims. The method is equally unsupported in both places. The number of people who will pay for it before finding that out is not equal.
The verification opportunity is better in the United States and the incentive to verify is stronger in India. American data is longer and freer, so the missing test is easier to run there. Indian readers are the ones being sold the method most aggressively right now. That mismatch is worth naming, because it means the useful work — actually testing a specified Gann rule on Indian index data, and publishing the failures — has an obvious audience and nobody doing it.
Carry this
- A 45-degree line is a property of your screen, not of the prices. Resize the window and the level moves.
- Every Gann projection contains at least 2 free choices: the anchor and the scale. Different choices give different answers, and there is no rule for choosing.
- The $50 million fortune has no documented source. The one named on-the-record account came from Gann's son and said the income came from selling courses.
- Round numbers and anniversary dates have small ordinary explanations. Keep those and drop the geometry, and you lose nothing.
- Gann's written risk rules — always use a stop, never average down, never overtrade — are worth more than every angle in the fan.