Goldbach Levels Explained: The ICT Framework Built on Prime Numbers

Most ICT concepts are reactive. You wait for the order block to form, you wait for the fair value gap to print, and then you mark it. Goldbach levels invert that. They claim you can draw all of those zones in advance, at fixed percentages of the range, before price gets there.

That is a big claim, and it comes wrapped in number theory, which makes it either compelling or suspicious depending on your temperament. This guide covers what the levels actually are, where the numbers come from, how they sit on top of concepts you already know (dealing ranges, power of three, IPDA), and, just as importantly, what the framework does not and cannot tell you.

You will also see the same idea circulating under other names. Huddleston levels and IPDA levels refer to the same percentage set. The source material is a text called Demystifying ICT by Hopiplaka, and most TradingView scripts carrying "Goldbach" in the title are implementing that mapping.

What Are Goldbach Levels?

A Goldbach level is a fixed percentage position inside a defined price range.

Take any range with a low and a high. Call the low 0 and the high 100. Goldbach levels are these twelve interior positions, plus the two boundaries:

0, 3, 11, 17, 29, 41, 47, 53, 59, 71, 83, 89, 97, 100

They are symmetrical around 50. Every level below the midpoint has a partner above it, and each pair sums to exactly 100. That symmetry is the entire structural idea: the range has a mirror, and the same zone that acts as support on the way up is expected to act as resistance on the way down.

The framework then assigns each level an ICT name. The 11 and 89 levels are the order block. The 17 and 83 levels are the fair value gap. The 3 and 97 levels are the rejection block. Rather than hunting the chart for those patterns, you draw the grid and read off where they should be.

Where Do the Numbers Come From?

This is the part worth understanding properly, because it is both simpler and weaker than it sounds.

Goldbach's conjecture, from 1742, states that every even number greater than two can be written as the sum of two primes. One hundred is an even number, so it can be written as a sum of two primes. In fact it can be written that way in exactly six distinct pairs:

• 3 + 97
• 11 + 89
• 17 + 83
• 29 + 71
• 41 + 59
• 47 + 53

Those twelve numbers are the Goldbach levels. That is the whole derivation. There is no market data in it, no order flow, no volume. It is a property of the number 100, and 100 was chosen because percentages are convenient.

Check it yourself: no other pair works. 7 and 93 fails because 93 is 3 times 31. 13 and 87 fails because 87 is 3 times 29. 23 and 77 fails because 77 is 7 times 11. Only six pairs survive, and they produce a level set that happens to be dense near the extremes (3, 11, 17), sparse through the middle (29, 41), and tight again around equilibrium (47, 53).

That distribution is the framework's real appeal. It puts the most levels exactly where ICT already teaches that reactions cluster: near the range boundaries where liquidity rests, and around the midpoint where price is fair. Whether that is insight or coincidence is a question the number theory cannot answer for you.

How Do Goldbach Levels Map to ICT PD Arrays?

Each level is the trigger price for a zone that runs from the previous level up to it. The mapping taught in the framework:

Two honest caveats on this table. First, implementations disagree about the 41 to 59 band: some scripts label 41 and 59 as the liquidity void trigger, others treat that band as the start of the breaker. If you use two different indicators you will see two different labels on the same level. Second, the ranking of which zones "matter most" (order block strongest, mitigation block weakest) is a claim from the source material, not something derived from the primes.

The mapping to order blocks and fair value gaps is an assignment, not a discovery. Somebody decided the 11 level should be called the order block. Nothing in the arithmetic says so.

What Is the PO3 Dealing Range?

Percentages need a range to be percentages of, and this is where the framework connects to the power of three model.

The claim is that price delivers in ranges sized by powers of three: 27, 81, 243 and 729 units, where a unit is a pip, a point or a handle depending on the instrument. Rather than defining your range from discretionary swing highs and lows, you snap it to a PO3 box:

• 27 for intraday scalping windows
• 81 for a normal session range
• 243 for a daily to multi-day range
• 729 for weekly and higher

You pick the PO3 size that reasonably contains the current move, anchor the box so it brackets price, and then lay the Goldbach percentages inside it. On a 243-point box, the 11 percent order block level sits about 27 points from the low, and the 17 percent fair value gap level about 41 points from the low.

This is the mechanical difference from ordinary ICT. A standard dealing range is drawn between two swing points you selected. A PO3 box is drawn to a fixed size regardless of where the swings are. That removes discretion, which is the selling point, and it also removes the market's input into where the range boundaries belong, which is the cost.

How Do You Place Goldbach Levels on a Chart?

The workflow, in the order the framework teaches it:

1. Choose the PO3 size for your timeframe. 81 for a session, 243 for a day. Do not mix: pick one and keep it for the whole analysis.
2. Anchor the box. Most implementations anchor to a session open or a daily open and extend the box up and down so the current price sits inside it.
3. Draw the twelve levels. Any of the free TradingView Goldbach scripts will do this once you set the range size, or you can place them manually with a Fibonacci tool using custom values.
4. Identify the direction of the draw. The levels tell you where the zones are. They do not tell you which way price is going. That still comes from your read of the draw on liquidity and whether the current target is internal or external range liquidity.
5. Wait for price to reach a level and react. The level is a location, not a signal. No reaction means no trade.
6. Enter on a lower-timeframe confirmation. Drop a timeframe, wait for displacement or a change of character in the direction of your draw, and enter on the retracement.
7. Set risk against the next level, not the current one. If you are long from the 11 level, your invalidation is below the 3 level, because that is the next defined zone beneath you.

Step 4 is the one people skip, and it is the one that matters. A grid of levels with no directional bias is a grid of places to lose money in both directions.

A Worked Example

Take a 243-point PO3 box on an index future, anchored so the box runs from 20,000 to 20,243.

The levels land here:

• 0 percent: 20,000 (range low, sell-side liquidity)
• 3 percent: 20,007 (rejection block)
• 11 percent: 20,027 (order block)
• 17 percent: 20,041 (fair value gap)
• 29 percent: 20,070 (liquidity void)
• 47 to 53 percent: 20,114 to 20,129 (breaker band around equilibrium at 20,121)
• 71 percent: 20,173
• 83 percent: 20,202 (fair value gap)
• 89 percent: 20,216 (order block)
• 97 percent: 20,236 (rejection block)
• 100 percent: 20,243 (range high, buy-side liquidity)

Now the read. Price opens near equilibrium, sells off through the morning, and trades down to 20,026, which is the 11 percent order block level, then stalls. Two things must be true before this is a setup rather than a coincidence.

First, the draw has to be upward. Check what is untouched: if the 100 percent boundary at 20,243 still has buy-side liquidity resting above it and the low side has already been swept, the external target is above. That gives you a direction.

Second, price has to actually react at 20,027. A wick that tags the level and closes back above it, followed by displacement on the 1-minute chart, is a reaction. Price grinding sideways through the level for twenty minutes is not.

If both hold, the trade is long from the reaction, stop below the 3 percent level at 20,007, first target the equilibrium band at 20,114 to 20,129, and the runner target the external liquidity above 20,243.

Notice what the levels contributed and what they did not. They gave you a pre-marked location and a pre-defined stop placement. They did not tell you the direction, the timing, or whether to take the trade. Everything that made this a setup came from the liquidity read layered on top.

What Goldbach Levels Cannot Tell You

This section matters more than the rest of the guide.

The prime numbers are not evidence. There is no mechanism by which the primality of 11 or 89 influences an order flow engine. Institutions are not partitioning ranges by Goldbach pairs. The conjecture is a genuine unsolved problem in mathematics and it has nothing to do with markets. Any explanation you read that treats "these are primes" as the reason the levels work is asserting a connection that does not exist.

A grid of twelve levels across any range will get hit. With levels every few percent, price cannot travel through a range without touching several. Reactions at Goldbach levels are guaranteed by density, not by significance. This is the same reason a chart covered in enough moving averages always shows one "holding."

The framework is unfalsifiable as usually presented. If price reacts at 11, that confirms the order block. If it does not and continues to 3, that confirms the rejection block. If it blows through both, the PO3 box was the wrong size and you should have used 729. A model that explains every outcome after the fact predicts none of them in advance.

Fixed boxes ignore actual structure. A 243-point box anchored at a session open will frequently place its boundaries in the middle of nowhere, with no swing high, no old high, and no resting liquidity at either edge. When that happens, the "external liquidity" at 0 and 100 is imaginary.

Nobody has published out-of-sample results. The circulating material is explanatory, not empirical. There is no published test of whether the 11 level reacts more often than the 13 level, which would be the obvious control and would take an afternoon to run.

None of this makes the framework useless. Used as a pre-marking scheme, it does something real: it forces you to define a range in advance, it gives you consistent invalidation points, and it stops you redrawing levels to justify a position you already hold. That is discipline, and discipline is worth something. Just be clear that the value is coming from the structure it imposes on you, not from a property of the market it has uncovered.

If you want the version of this idea with an actual mechanism behind it, read the IPDA framework, which grounds its reference points in real prior price action rather than in arithmetic.

Common Goldbach Mistakes

Trading the level instead of the reaction. The level is a place to pay attention. Price arriving is not a signal. No displacement, no change of character, no trade.

Changing the PO3 size after price invalidates you. If you are stopped out on an 81 box and immediately redraw at 243 to keep the idea alive, you have stopped using a framework and started rationalising.

Ignoring the higher timeframe draw. The grid is direction-agnostic. Without a bias from liquidity, you will take the 11 level long and the 89 level short on the same day and lose both.

Assuming the labels are load-bearing. The 11 level is not an order block. It is the 11 percent level, which somebody named after an order block. If a real order block prints at 14 percent, trade the real one.

Stacking multiple PO3 sizes at once. Three boxes on one chart produce thirty-six levels. At that density every candle touches something, and the framework has told you nothing.

Final Takeaway

Goldbach levels are a partitioning scheme dressed as a discovery. The arithmetic is real and easy to verify, the mapping to ICT arrays is an assignment somebody made, and the claim that price respects the levels because they are prime does not survive a moment's scrutiny.

That does not mean you should ignore it. If pre-marking a range with fixed, symmetrical, non-negotiable levels stops you from moving your invalidation and forces you to name your target before you enter, the framework has earned its place on your chart. Just hold it at the right altitude: it tells you where to watch, never whether to trade or which way.

For the parts that carry the actual decision, work through the dealing range that defines your boundaries, the draw on liquidity that names your target, and internal versus external range liquidity that tells you how far the move should travel.