Skip to main content
谱 Su-Pu Studies · Lesson 08

Solution Space, Not Candidate Space

Why ibtree's pencil digits trace a trial solution forward, not every possible candidate in every cell.

Written by the Orbace Sudoku Editorial Team · September 11, 2026 · 8 min read · Download PDF
Su-Pu Studies Extreme Ibtree Binary Tree Cross-hatching

In this study

  1. Prepare the Ground
  2. Identify the Fork
  3. Trace the Assumption
  4. Two Kinds of Space
  5. Retrospect

01 · Prepare

Prepare the Ground

Before selecting a chain trigger digit, we want to confirm as many digits as possible using basic cross-hatching and confined group methods. This narrows down the choice of the fork itself — the fewer open cells, the clearer the binary options become.

The starting grid
The starting grid, with many givens already in place.

After placing 9 at R3C9 and 1 at R8C4, we can visually identify two confined groups: a 5-7 pair and a 4-9 pair, both in Box 8.

Confined groups identified
The confined groups in Box 8 narrow the remaining candidates.

Afterwards, we can confirm 4 at R4C9 — a straightforward cross-hatch placement.

4 at R4C9 confirmed
Placing the 4 clears one more cell and sharpens the fork ahead.

02 · Fork

Identify the Fork

The 4 at R4C9 is helpful because it presents a binary choice: digit 7 in Box 6 now has exactly two possible homes.

Binary choice for 7 in Box 6
Box 6 offers a clean fork — two cells, one digit, no ambiguity.

Are there any other binary choices? In Box 5, there are two possible homes for 3. But a quick visual inspection shows that neither 3 in Box 5 would substantially trigger any further forced placements. This selection is less ideal.

Alternative fork in Box 5
The 3 in Box 5 is also a binary choice, but it doesn't cascade forward.

In contrast, assuming 7 at R6C9 would generate a long reaction. It's a no-brainer to test it through.

03 · Trace

Trace the Assumption

Assuming 7 at R6C9, we can immediately fill Row 4 with pencil digits.

Row 4 filled under the assumption
The assumption cascades into Row 4, filling multiple cells in one step.

Furthermore, we can fill up Box 3 with pencil digits — including a confined group of 6-8 — and many other digits in Box 2 and Box 9.

Multiple boxes filled
The assumption continues to cascade: Box 3, Box 2, Box 9 all take shape.

The method used here is cross-hatching — simple, easy, and straightforward. In no time, we can get the entire grid filled with pencil digits.

Entire grid filled
The complete grid, every cell accounted for under the assumption.

While converting pencil digits to confirmed digits, we just need to cross-check each digit — horizontally (Box 1-2-3, 4-5-6, 7-8-9) and vertically (Box 1-4-7, 2-5-8, 3-6-9) — to avoid human error.

04 · Spaces

Two Kinds of Space

One important point to make here: in the candidate reduction (canred) method, a player needs to fill in all possible digits and then perform rigorous reduction by identifying what is not possible in certain cells. In a sense, canred uses pencil digits in the candidate space — every cell holds a list of every digit that could still fit.

In the ibtree (inferential binary tree) method, pencil digits serve a different role. They represent a trial solution under an assumption, tracing forward along one branch of the fork. In this sense, ibtree uses pencil digits in the solution space — a much smaller set, since each cell holds at most one digit under the current assumption, not a full list of remaining candidates.

The Orbace rule for pencil digits in ibtree: pencil marks trace a trial solution forward, not every possible candidate. A smaller space means a faster cascade and a clearer path to the proof.

Ibtree Candidate reduction
Pencil digits represent one assumed solution at a time. Pencil digits list every remaining candidate in every cell.
Solution space: one digit per cell, chosen under the current assumption. Candidate space: multiple digits per cell, reduced step by step.
Fast forward cascade: a single assumption fills the grid quickly. Systematic elimination: each technique removes one candidate at a time.
Proof by contradiction if the assumption fails; immediate confirmation if it holds. Gradual narrowing until only one candidate remains in each cell.

05 · Retrospect

Retrospect

Was it lucky to try out 7 at R6C9? Yes — the assumption held, and the grid filled cleanly. Was it guesswork? No. If the assumption had caused a conflict, we would have confirmed 7 at R4C8 by proof of contradiction. Either way, the fork resolves.

This is the discipline of ibtree: pick a binary choice, trace it forward in solution space, and let the grid tell you whether the assumption holds or breaks. The pencil digits you write are not a catalog of possibilities — they're a single trial solution, committed enough to cascade forward, light enough to erase if it fails.

茶

Try this in your next solve

Next time you face a fork, trace the assumption forward in solution space — one trial digit per cell, fast and light. Let the cascade do the work.

Start a new puzzle

Reconstructed from a real personal Su-Pu replay and checked against the recorded grid. Spotted a step that doesn't hold up? Tell us — every Journal article carries an accountable byline and an open correction channel.