← Back to blog

Gann Square of Nine for Silver: A Worked Example

Type one silver price into the Square of Nine and the app returns a full ladder of levels in a single step. From the 16 July 2026 closing low of 55.898, the cardinal levels run 49.00, 56.25, 64.00, 72.25, 81.00 and on up. This post shows how to run it, how to read it, and how silver's closes sat against the grid, including the misses.

The useful part comes first. The spiral arithmetic is exact and the app does it instantly. You can then test the levels against real closes, and work out the odds of a lucky fit, before you give them any weight. That order is the point of the exercise.

What the Square of Nine does

The Square of Nine is a number spiral. Take the square root of a price, add a quarter or half a turn, and square the result. Each step gives a price that shares an angle on the spiral with your starting price. The app marks the 0, 90, 180 and 270 degree levels as cardinal, and the 45 degree levels in between as diagonal.

For the background, see our Square of Nine guide. This post is the worked example for one market.

Step 1: choose the anchor from a stated rule

The anchor is the one input that is up to you, so pick it by a rule you write down first. Otherwise you can try ten anchors and keep the one that looks best.

Our rule: the lowest daily close of silver from 2 January 2026 to 30 September 2026. We stopped at 30 September because a low needs some weeks to show it was a low. We pulled the closes from Yahoo Finance (ticker SI=F, silver futures) on 29 October 2026. The lowest close in that range is 55.898 on 16 July 2026, and we entered it as 55.90.

Some context from the same data:

  • The highest close in the whole window is 115.08 on 26 January 2026, so the low sits 51.4% below the high.
  • Silver then fell fast. It closed at 114.04 on 29 January and 78.29 on 30 January, a one day drop of 31.3%.
  • The latest close, on 28 October 2026, is 53.665. That is under our anchor, and a new low for the year. It is only days old, which is why the rule left it out.

(51.4% is 1 - 55.898 / 115.08. 31.3% is 1 - 78.29 / 114.037. The 28 October close is 4.0% below the anchor: 53.665 / 55.898 - 1.)

A caution on the data. A futures series built from rolling contracts has quirks. Some days in the file show the open, high, low and close all identical with tiny volume, which looks like a stale print. We used the closes as Yahoo supplies them and did not adjust for contract rolls. Treat the exact decimals as indicative.

Step 2: run the module

It costs 1 compute unit, and the Starter plan includes 50 units a day.

  1. Sign in at crohamhurst.app and open the dashboard.
  2. Choose Gann Square of Nine from the sidebar.
  3. Type 55.90 into the anchor price box.
  4. Leave the step at 45 degrees and the rings at 4.
  5. Press Run Square of Nine.

Crohamhurst Gann Square of Nine form with the anchor price 55.90 entered, 45 degree step and 4 rings

Step 3: read the ladder

The app places 55.90 at 85.8 degrees on the spiral, close to the 90 degree north axis. That is a feature of this price, not something we chose. The app draws the spiral with the anchor in green:

Square of Nine spiral for silver with the anchor 55.9 marked in green near the 90 degree north axis and rings of price levels around it

The cardinal levels around the anchor from that run:

Level Type Distance from the anchor
42.25 Cardinal (South) -24.42%
49.00 Cardinal (East) -12.34%
56.25 Cardinal (North) +0.63%
64.00 Cardinal (West) +14.49%
72.25 Cardinal (South) +29.25%
81.00 Cardinal (East) +44.90%

The diagonal levels sit between them: 52.5625, 60.0625, 68.0625 and 76.5625. Silver trades at a low price, so the grid is wide. The gap between neighbouring cardinal levels is about the square root of the price plus a quarter, which is roughly 7.7 dollars at 56. That is about 14% of the price. Hold that thought for the chance check below.

What the closes did against the grid

We checked three points, chosen by simple rules before looking at the levels:

  • The lowest close in the window up to 30 September (the anchor, 16 July 2026).
  • The highest close in the window (26 January 2026).
  • The highest close after the anchor up to 30 September (21 August 2026, 69.466).

We counted a "fit" as a close within 0.1% of a cardinal level. For the other two points we ran the module again, so the nearest levels come from the app too.

Point Close Nearest cardinal level Distance Fit?
Window low, 16 Jul 2026 55.898 56.25 0.63% No
Window high, 26 Jan 2026 115.08 110.25 4.20% No
High after the low, 21 Aug 2026 69.466 72.25 4.00% No

Each distance is the level minus the close, divided by the close. For the low: (56.25 - 55.898) / 55.898 = 0.63%. For the January high: (110.25 - 115.08) / 115.08 = -4.20%. For the August high: (72.25 - 69.466) / 69.466 = 4.00%.

The result is zero fits from three points. One near miss is worth noting: the January high of 115.08 sat 0.42% below the diagonal level at 115.5625. That is outside our band, and we did not widen the band after seeing it.

How often would a random price fit?

This is the check the app lets you run before you trust a level. A band of 0.1% either side is 0.2% of the price wide. Divide that by the gap between the two cardinal levels that enclose the price and you get the chance of a random close fitting:

Point Enclosing cardinal levels Gap Band width Chance a random close fits
55.898 49.00 and 56.25 7.25 0.112 1.5%
115.08 110.25 and 121.00 10.75 0.230 2.1%
69.466 64.00 and 72.25 8.25 0.139 1.7%

The three chances add to 5.4%, so luck alone would give about 0.05 fits from three points. We got none, which is exactly what luck predicts and tells us nothing for or against the method. Three points cannot do that.

What the table does show is useful. At silver prices the cardinal grid is sparse, so a tight band is hard to hit by accident. A fit here would be more notable than the same fit in bitcoin, where the levels crowd together and a fixed band catches most random prices. The sparse grid works in your favour as a tester: a real pattern would show up against a low chance baseline. If you add the diagonals you roughly halve the gaps and double the chance of a lucky hit, so count them separately.

Where the latest close sits

For reference only, silver closed at 53.665 on 28 October 2026. On the same grid that is between the diagonal level at 52.5625 (2.1% below) and the cardinal level at 56.25 (4.8% above). Those are reference marks, not forecasts. We have not tested whether silver reacts to them, and a level can be ignored as easily as respected.

Common mistakes

  • Choosing the anchor after seeing the chart. Pick it by a rule, as we did here.
  • Counting only the hits. Note every level the market ignored, and compute the fits chance would have given you.
  • Moving the band after the result. Fix the band before you look.
  • Mixing closes and intraday prices. Choose one and keep to it. Silver's intraday range can be wide: on 30 January 2026 it traded between 76.03 and 117.79 and closed at 78.29.
  • Trusting every decimal. Check the data source for stale prints and contract rolls.

What this does not tell you

The ladder gives price levels, not direction, and not whether a level will hold. Three points cannot confirm or reject a method. A stronger test would score many swing points over many years, or a written call with a date and a rule, as we do in our gold example. The positive side is that the app makes this check fast and repeatable for any market. The same steps for WTI crude take a few minutes, and our post on what the app does well shows the other modules.

This is research and education, not financial advice.

Try it yourself

Create an account at crohamhurst.app, open Gann Square of Nine, and enter a swing low you chose by a rule. The Square of Nine guide walks through every field, and the free calculator works without an account.

Run this analysis yourself

Crohamhurst computes the Square of Nine, Bradley siderograph, planetary ephemeris, cycle scans and historical analogs from your own price data, in your browser. Plans start at $49 a month.

Create an account