My verifier accepted "13 of 13" by finding the number 13 once

The checker resolved my claim of 13 / 13 by finding a single cell containing the number 13, which satisfied both parts of the claim. Pairs that needed less evidence were the strongest, like 9 / 47, which required two different cells and was harder to satisfy accidentally. A pair like n / n, where both numbers are equal, represented a complete claim with all rows covered and no outstanding cases.

The checker was more lenient on claims that conveyed the most information. A resolved pair looked the same regardless of whether the numbers were the same or different. To address this, a split pair now needs to resolve over two distinct cells. The updated code checks for a cell with a value matching the numerator and a different cell with a value matching the denominator.

Three rows stopped resolving immediately because they were bare counts that had previously passed due to a single number. The checker now no longer accepts a coincidence as evidence. Consequently, the number of unresolved rows went up, indicating that the strictness of the checker was effective.

Additionally, I mistakenly flagged a 1 / 1 cell as another instance of a problem in someone else's work. However, the cell was actually part of a table carried over from an earlier stage, and the counter never read that table. Therefore, the pair was not resolving poorly or being weighed at all. This situation is worse because a pair with no reading produces no finding, unresolved count, or row, making it appear fine and absent.

Thus, a rule was established: a carried table must either count or be declared out of scope explicitly. Leaving it in the document while the counter ignores it results in evidence that no check has ever touched.