Chapter map

1. 1. Item set
2. 2. Table
3. 3. Row one maximum
4. 4. Row two maximum
5. 5. Row total
6. 6. Column one maximum
7. 7. Column two maximum
8. 8. Column three maximum
9. 9. Column total
10. 10. Baseline
11. 11. Distance
12. 12. Normalize
13. 13. Competitor table
14. 14. Competitor distance
15. 15. Blind spot
16. 16. Claim

Visit the How Mathematics Works hub for the wider series and its connections to education, reasoning, science, computing, design and careers.

---

1. Item set

Here the geometry becomes operational. Align the same eight observations.

A concrete situation

n=8 enters the maximum distance.

What the mathematics does not license

Missing items need a declared policy. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Record n. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

2. Table

This step deserves a careful pause. Use [[3,1,0],[1,2,1]].

A concrete situation

Rows are P clusters and columns Q clusters.

What the mathematics does not license

Labels can be permuted. The same reasoning helps students move from an exercise to a real dataset.

Try it, check it, transfer it

Record table. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

3. Row one maximum

The useful question is not only how to calculate, but what the calculation assumes. Take max(3,1,0).

A concrete situation

First row contributes 3.

What the mathematics does not license

Use one maximum per row. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Show step. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

4. Row two maximum

A small example makes the mechanism visible. Take max(1,2,1).

A concrete situation

Second row contributes 2.

What the mathematics does not license

Ties need no arbitrary label choice for value. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Show step. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

5. Row total

This is where modelling choices begin to affect evidence. Add 3+2.

A concrete situation

Row-max total=5.

What the mathematics does not license

Do not sum all cells. The same reasoning helps students move from an exercise to a real dataset.

Try it, check it, transfer it

Record total. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

6. Column one maximum

The central mathematical idea is concrete. Take max(3,1).

A concrete situation

First column contributes 3.

What the mathematics does not license

Direction is Q to P. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Show step. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

7. Column two maximum

Here the geometry becomes operational. Take max(1,2).

A concrete situation

Second column contributes 2.

What the mathematics does not license

Keep column orientation. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Show step. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

8. Column three maximum

This step deserves a careful pause. Take max(0,1).

A concrete situation

Third column contributes 1.

What the mathematics does not license

Singletons still count. The same reasoning helps students move from an exercise to a real dataset.

Try it, check it, transfer it

Show step. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

9. Column total

The useful question is not only how to calculate, but what the calculation assumes. Add 3+2+1.

A concrete situation

Column-max total=6.

What the mathematics does not license

Audit against n. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Record total. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

10. Baseline

A small example makes the mechanism visible. Compute 2n.

A concrete situation

The baseline is 16.

What the mathematics does not license

Both directions are represented. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Show step. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

11. Distance

This is where modelling choices begin to affect evidence. Compute 16-5-6.

A concrete situation

Van Dongen distance=5.

What the mathematics does not license

Lower is closer. The same reasoning helps students move from an exercise to a real dataset.

Try it, check it, transfer it

Record value. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

12. Normalize

The central mathematical idea is concrete. Compute 5/16.

A concrete situation

Normalized distance=.3125.

What the mathematics does not license

State denominator. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Record value. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

13. Competitor table

Here the geometry becomes operational. Use [[2,2,0],[0,1,3]].

A concrete situation

Row and column totals still cover eight items.

What the mathematics does not license

Compare full tables. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Record table. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

14. Competitor distance

This step deserves a careful pause. Compute 16-5-7.

A concrete situation

Distance=4 or .25 normalized.

What the mathematics does not license

Use same convention. The same reasoning helps students move from an exercise to a real dataset.

Try it, check it, transfer it

Compare. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

15. Blind spot

The useful question is not only how to calculate, but what the calculation assumes. Move mass among nonmaximum cells.

A concrete situation

Score may stay unchanged.

What the mathematics does not license

Inspect the full contingency table. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Audit. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

16. Claim

A small example makes the mechanism visible. Name table,maxima,n,normalization and item policy.

A concrete situation

Second candidate is closer.

What the mathematics does not license

Do not claim identical error structure. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Bound conclusion. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

A seven-day practice lab

1. Day 1: define the decision, observation unit, X variables and Y or reference condition.
2. Day 2: reproduce the smallest worked example and check every dimension, unit and sign.
3. Day 3: build a transparent baseline and write what it assumes.
4. Day 4: fit the main method inside a training-only pipeline; record every parameter.
5. Day 5: challenge one assumption with a deliberately difficult dataset or operating regime.
6. Day 6: compare validation metrics with the practical workload created by errors.
7. Day 7: present one supported conclusion, one unresolved uncertainty and one safe next experiment.

Questions for a student, parent or mentor

• Which part of the model is a mathematical definition, and which part is a choice?
• What small calculation would catch a sign, scale or dimension error?
• What evidence comes from training, validation and the untouched test?
• What benign condition could look like a fault or prediction failure?
• How would the answer change for another response, asset, school project or time period?
• What should a human review before acting on the score?

A method-specific audit before deployment

Van Dongen distance compares partitions through their contingency table. In each row, keep the largest overlap between one cluster in P and any cluster in Q. Do the same for every column. These best-overlap totals reward how much of each cluster can be matched in both directions. The distance is 2n minus the row-max total minus the column-max total.

Let eight observations form the contingency table [[3,1,0],[1,2,1]]. Row maxima are 3 and 2, summing to 5. Column maxima are 3, 2 and 1, summing to 6. The unnormalised distance is 2(8)-5-6=5. Dividing by 2n=16 gives a normalized distance of 0.3125.

For a second candidate table [[2,2,0],[0,1,3]], row maxima again sum to 5, while column maxima sum to 2+2+3=7. Its distance is 16-5-7=4, or 0.25 after division by 16. Lower is better, so the second candidate is closer under this metric.

The score focuses on the best overlap in each row and column; rearrangements among nonmaximum cells can sometimes leave it unchanged. It requires the same item set and an explicit treatment of unassigned or overlapping observations. Report the full table, row and column maxima, normalization, direction, cluster sizes and any matching or filtering rules.

---

Frequently asked questions

What is Van Dongen distance?

Two times the item count minus the sum of every row maximum and every column maximum in the partition contingency table.

Which direction is better?

Lower distance is closer; identical partitions have distance zero.

Why use both row and column maxima?

They capture best overlaps in both split and merge directions.

Is it label invariant?

Yes. Permuting cluster names leaves the table maxima unchanged.

What can it miss?

Changes among nonmaximum cells can leave the score unchanged.

What must be reported?

Full table, n, row maxima, column maxima, normalization, item set, unassigned-item policy and cluster-size margins.

---

Primary source and fact-check note

Stijn van Dongen, “Performance Criteria for Graph Clustering and Markov Cluster Experiments,” CWI Technical Report INS-R0012 (2000). Definitions and core claims in this article were checked against this primary publication. Calculations shown here were independently recomputed. Examples are educational illustrations; they are not reported industrial performance results.

Useful next reading

• V-Measure homogeneity and completeness
• Cluster purity and inverse purity
• B-Cubed item-level evaluation
• How AI finds patterns without labels
• eduKateSingapore learning routes

Keep the options open. Mathematics education supports clearer questions, transferable problem-solving skills and informed choices across science, engineering, computing, AI, statistics, design and many other careers. It does not guarantee intelligence, income, admission or employment; those outcomes depend on far more than one subject or model.

Extended laboratory: partition-comparison evidence and responsible use

Data audit

Freeze the shared item set, identifiers, exclusions, missing-label rules, weights, sample size and any reference partition.

Partition audit

Record both label vectors, cluster counts, sizes, unassigned items, overlaps and whether labels are hard or soft.

Table audit

Show the full contingency table and derive a, b, c, d, row maxima, column maxima and positive-pair totals.

Direction audit

Name reference and candidate, split and merge directions, denominators, normalization and lower-or-higher selection rule.

Arithmetic audit

Recalculate combinations, pair totals, component ratios, means, raw distances and normalized distances independently.

Uncertainty audit

Repeat samples, label perturbations, cluster counts, algorithms, starts and plausible matching or missing-item conventions.

Meaning audit

Compare partition agreement with geometry, stability and domain usefulness without assuming that either partition is true.

Final claim

State the exact pair or overlap evidence and stop before causal, semantic, career or guaranteed-performance claims.

Mechanism workshop: calculate, challenge and transfer

Workshop 1: Item set

Begin with this mechanism: Align the same eight observations. Recreate the concrete case: n=8 enters the maximum distance. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Missing items need a declared policy. Complete the transfer step: Record n. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 2: Table

Begin with this mechanism: Use [[3,1,0],[1,2,1]]. Recreate the concrete case: Rows are P clusters and columns Q clusters. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Labels can be permuted. Complete the transfer step: Record table. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 3: Row one maximum

Begin with this mechanism: Take max(3,1,0). Recreate the concrete case: First row contributes 3. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Use one maximum per row. Complete the transfer step: Show step. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 4: Row two maximum

Begin with this mechanism: Take max(1,2,1). Recreate the concrete case: Second row contributes 2. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Ties need no arbitrary label choice for value. Complete the transfer step: Show step. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 5: Row total

Begin with this mechanism: Add 3+2. Recreate the concrete case: Row-max total=5. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Do not sum all cells. Complete the transfer step: Record total. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 6: Column one maximum

Begin with this mechanism: Take max(3,1). Recreate the concrete case: First column contributes 3. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Direction is Q to P. Complete the transfer step: Show step. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 7: Column two maximum

Begin with this mechanism: Take max(1,2). Recreate the concrete case: Second column contributes 2. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Keep column orientation. Complete the transfer step: Show step. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 8: Column three maximum

Begin with this mechanism: Take max(0,1). Recreate the concrete case: Third column contributes 1. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Singletons still count. Complete the transfer step: Show step. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 9: Column total

Begin with this mechanism: Add 3+2+1. Recreate the concrete case: Column-max total=6. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Audit against n. Complete the transfer step: Record total. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 10: Baseline

Begin with this mechanism: Compute 2n. Recreate the concrete case: The baseline is 16. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Both directions are represented. Complete the transfer step: Show step. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 11: Distance

Begin with this mechanism: Compute 16-5-6. Recreate the concrete case: Van Dongen distance=5. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Lower is closer. Complete the transfer step: Record value. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 12: Normalize

Begin with this mechanism: Compute 5/16. Recreate the concrete case: Normalized distance=.3125. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: State denominator. Complete the transfer step: Record value. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 13: Competitor table

Begin with this mechanism: Use [[2,2,0],[0,1,3]]. Recreate the concrete case: Row and column totals still cover eight items. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Compare full tables. Complete the transfer step: Record table. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 14: Competitor distance

Begin with this mechanism: Compute 16-5-7. Recreate the concrete case: Distance=4 or .25 normalized. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Use same convention. Complete the transfer step: Compare. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 15: Blind spot

Begin with this mechanism: Move mass among nonmaximum cells. Recreate the concrete case: Score may stay unchanged. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Inspect the full contingency table. Complete the transfer step: Audit. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.

Workshop 16: Claim

Begin with this mechanism: Name table,maxima,n,normalization and item policy. Recreate the concrete case: Second candidate is closer. Write n, full contingency table, row maxima, column maxima, raw distance, normalization and candidate comparison with identifiers and assumptions. Calculate the smallest non-trivial example by hand and reproduce it independently. Challenge this limitation: Do not claim identical error structure. Complete the transfer step: Bound conclusion. Change exactly one item, label, cluster size, contingency-table cell, pair count, direction, denominator, normalization, matching rule, missing-item policy or random seed. Explain which mathematical object changes, which score moves and which interpretation must weaken. Keep pair agreement, splits, merges, geometry, stability, uncertainty and practical usefulness on separate notebook lines. Explain once to a student and once to a parent. Add a counterexample involving all singletons, one giant cluster, zero positive-pair denominator, missing labels, unequal cluster sizes, duplicated identifiers or a nonmaximum table-cell change. State the extra evidence needed to distinguish explanations. Finish with one supported sentence, one unsupported sentence and one useful next measurement.