Skip to content
TechniquesExpert

X-Chain

Learn how an X-Chain alternates Strong and Weak Links for one digit and eliminates that digit from cells that see both endpoints.

An X-Chain is an alternating Sudoku chain that follows one candidate digit only.

If the chain digit is 7, every chain node represents candidate 7 in a different cell.

The chain alternates:

  • Strong inference roles, usually supplied by conjugate pairs;
  • Weak inference roles, supplied by same-digit candidates that see each other.

The standard eliminative X-Chain taught here:

  • starts with a Strong Link;
  • alternates Strong and Weak roles;
  • ends with a Strong Link;
  • proves that at least one endpoint contains the chain digit.

Therefore the chain digit can be removed from every candidate that sees both endpoints.

X-Chain is the general single-digit language behind several shorter named patterns.

Quick pattern

Conceptually:

A = B - C = D - E = F

For one digit X:

  • = = Strong role;
  • - = Weak role.

Trace from A false:

A false
→ B true
→ C false
→ D true
→ E false
→ F true

Trace from F false:

F false
→ E true
→ D false
→ C true
→ B false
→ A true

So A and F cannot both be false.

At least one endpoint is X.

Sudoku visualAn X-Chain alternates strong and weak links on one digitFollow the same candidate through one-digit conjugate logic. If the endpoint digit is true at one end or the other, a target that sees both endpoints cannot keep that digit.
7A
7B
7C
7D
Sees both 7 endpoints · remove 7

Why X-Chain uses only one digit

The “X” in X-Chain represents one candidate digit followed across the grid.

That makes the chain easier to visualize than a general AIC because:

  • candidate value never changes;
  • Strong Links normally come from houses with exactly two X candidates;
  • Weak Links come from X candidates that see each other.

If the chain changes from digit 7 to digit 3 inside a bivalue cell, it is no longer a pure X-Chain. You have moved into a different chain family such as XY-Chain or a more general AIC.

Why the chain starts and ends Strong

The standard endpoint elimination requires this conclusion:

at least one endpoint must be true.

At the first endpoint, assuming it false activates the first Strong Link and starts the alternating propagation.

At the far endpoint, a Strong-ending structure makes the reverse proof work as well.

So:

A false → F true
F false → A true

Therefore A and F cannot both be false.

That endpoint truth guarantee is what makes a common-peer elimination valid.

If the path does not produce that guarantee, do not use the standard X-Chain endpoint rule.

For one fixed digit X, Strong Links usually come from conjugate pairs.

Examples:

  • exactly two candidate Xs in one row;
  • exactly two in one column;
  • exactly two in one box.

Those pairs are exhaustive:

if one endpoint is not X, the other must be X.

A conjugate pair is also Weak because both endpoints see each other for the same digit. This means a Strong relationship can sometimes serve in the Weak role needed by the alternating path.

A Weak Link exists between two candidate Xs that cannot both be true.

For same-digit X candidates, this normally means the cells see each other through:

  • a row;
  • a column;
  • a box.

Unlike a Strong Link, the source house does not need to contain exactly two X candidates.

If B is X, C cannot be X.

But if B is not X, C is not automatically forced.

That asymmetry is exactly what the following Strong segment repairs.

The endpoint elimination

Suppose a valid X-Chain proves:

A is X or F is X.

Now consider a target candidate T for digit X.

If T sees A, then T conflicts with the case A = X.

If T also sees F, then T conflicts with the case F = X.

Those two endpoint cases cover all possibilities.

Therefore:

T cannot be X.

The target does not need to see every internal chain node.

For the standard elimination taught here, it must see both endpoints.

Step-by-step example

Before using the board, identify these five facts:

  1. one fixed chain digit X;
  2. first Strong-Link segment;
  3. alternating Weak/Strong path;
  4. final Strong-Link segment;
  5. outside candidate X seeing both endpoints.

Short X-Chains and named patterns

Several familiar single-digit techniques can be expressed as short X-Chains.

Skyscraper

Two conjugate pairs form a short chain whose roof endpoints create a common-peer elimination.

Two-String Kite

A row conjugate pair and column conjugate pair are connected through a box Weak Link.

Turbot Fish

A general family of short single-digit chain structures, usually described with named geometries when possible.

Empty Rectangle

Another short single-digit pattern that can be represented through chain logic.

If a short named technique describes the structure clearly, use the name. X-Chain becomes most useful when the alternating path extends beyond one familiar shape.

X-Chain vs X-Wing

Both techniques follow one digit, but the proof structure is different.

X-Wing

Uses a Fish counting structure across two base sets and two cover sets.

X-Chain

Uses alternating Strong and Weak implications along a path.

Some candidate layouts can admit more than one description, but you should prefer the simplest proof that accurately explains the elimination.

X-Chain vs Simple Coloring

Both techniques work with one candidate digit and conjugate-pair structures.

Simple Coloring

  • builds a connected network of Strong Links;
  • assigns two alternating states;
  • searches for Color Trap or Color Wrap.

X-Chain

  • chooses one particular alternating path;
  • explicitly includes Weak connections between Strong segments;
  • uses an endpoint relationship to eliminate a common peer.

A large coloring network may contain many possible X-Chain paths.

Coloring is often a better network view. X-Chain is often a better path view.

How to find X-Chains without drawing the whole candidate graph

Use a controlled digit-first search.

1. Choose one digit

Prefer a digit with several conjugate pairs.

Rows, columns, and boxes with exactly two positions are your main anchors.

3. Choose a promising endpoint

Start from a Strong-Link endpoint near other same-digit candidates.

4. Find a Weak connection

From the true side of one Strong segment, look for another X candidate it sees that can enter a new Strong segment.

5. Continue the alternation

Strong → Weak → Strong → Weak...

6. Stop as soon as useful endpoints appear

Do not extend a chain merely because another link exists.

7. Search common peers of the endpoints

If an outside X candidate sees both endpoint cells, test the elimination.

Short, verifiable chains are preferable to long speculative ones.

Suppose two candidate 4s form a conjugate pair.

They are Strong because one false forces the other true.

They are also Weak because both cannot be true in the same house.

So if an X-Chain requires a Weak step at that point, the conjugate pair can supply it.

This flexibility can make chain diagrams look as though two “Strong segments” touch.

The correct question is not the visual line label alone.

Ask:

Which inference role is this relationship performing at this step?

A valid X-Chain can be unproductive

You may build a perfectly valid alternating path and still find no candidate X that sees both endpoints.

Then the chain, in the standard form taught here, has no current elimination.

Do not invent one.

Return to:

  • a different endpoint;
  • a different chain digit;
  • another technique.

Advanced solving depends as much on abandoning unproductive structures as on finding valid ones.

How long should an X-Chain be?

There is no useful goal of maximizing length.

Start with short chains because they are:

  • easier to recognize;
  • easier to verify in both directions;
  • less vulnerable to candidate errors;
  • more likely to correspond to a familiar named pattern.

Only extend when a shorter structure produces no result and the next link remains completely clear.

Common mistakes

Changing candidate digit inside the chain

A pure X-Chain uses one digit only.

Using a Weak-only relationship as Strong

If A is false, B must genuinely be forced true.

Alternating line styles without checking inference direction

The logic matters more than the drawing convention.

Ending with the wrong endpoint structure

The common-peer elimination requires the proof that at least one endpoint is true.

Eliminating from a target that sees only one endpoint

Then the other endpoint case leaves the target possible.

Treating every conjugate network as an X-Chain elimination

A network is not automatically a productive path.

Building extremely long chains too early

Master short, transparent proofs first.

Recognition drill

Filter digit 7.

  1. Mark every conjugate pair.
  2. Choose one pair as the first Strong segment.
  3. From one endpoint, find a visible 7 that participates in another conjugate pair.
  4. Continue once or twice.
  5. Identify the two final endpoints.
  6. Search for candidate 7s that see both.

If you cannot explain every Strong source and every Weak conflict, stop and re-check the chain.

FAQ

What is an X-Chain in Sudoku?

An X-Chain is a chain that follows one candidate digit through alternating Strong and Weak inference roles.

Does X-Chain use only one digit?

Yes. If the candidate value changes along the path, it is not a pure X-Chain.

Why does an X-Chain start and end Strong?

That structure proves that at least one endpoint must contain the chain digit, which enables the standard common-peer elimination.

Can a Skyscraper be represented as an X-Chain?

Yes. Skyscraper and other Turbot-style patterns can be represented as short X-Chains.

Is X-Chain the same as Simple Coloring?

No. Coloring manages a connected Strong-Link network with two states, while X-Chain follows one specific alternating Strong/Weak path.

Does every valid X-Chain produce an elimination?

No. In the standard endpoint form, there must be a useful target that sees both endpoints.

Is X-Chain an AIC?

It is a restricted one-digit form of alternating-inference logic. The later AIC Guide will generalize the framework beyond one digit and this specific endpoint pattern.

What to learn next

Move to XY-Chain, where the Strong Links live inside bivalue cells and the candidate value changes as the chain travels.