# VectorPaint Fermat check: ninety-one passes base three yet equals seven times thirteen

Language: en
Canonical: [VectorPaint Fermat check: ninety-one passes base three yet equals seven times thirteen](<https://allluckly.cn/en/guides/vectorpaint-fermat-pseudoprime-ninety-one-base-three-witness/>)
中文：[中文 HTML](<https://allluckly.cn/guides/vectorpaint-fermat-pseudoprime-ninety-one-base-three-witness/>)
English: [English HTML](<https://allluckly.cn/en/guides/vectorpaint-fermat-pseudoprime-ninety-one-base-three-witness/>)
Published: 2026-11-19T00:00:00+08:00
Modified: 2026-11-19T10:01:54+08:00
Author: Bison

Draw an implication card and modular-power ledger, checking 3^90 remainder one and base-two remainder sixty-four without mistaking a necessary condition for primality.

English translation of the [Chinese original](<https://allluckly.cn/guides/vectorpaint-fermat-pseudoprime-ninety-one-base-three-witness/>), translated on 2026-11-19. Original publication dates, attribution, code and version notes are retained.

Distinguish “prime implies this condition” from “this condition implies prime”. Use composite ninety-one as a complete counterexample: establish seven times thirteen first, then verify its base-three Fermat congruence. Recalculate integers through modular arithmetic. A “pass” label records only this check, not software primality certification or a security conclusion.

## State the direction of Fermat’s little theorem

If p is prime and integer a is coprime to p, then a^(p−1) has remainder one modulo p. Test odd n=91 at a=3 with exponent n−1=90, checking coprimality first. The theorem runs from primality to congruence, not automatically both ways. On about 2000 × 1400 pixels place implication left, powers centre and controls right. Group numbers, arrows and conclusions, then save. These abstract integers are not keys or accounts. Calculate externally; colour alone does not establish truth.

## Establish compositeness through seven times thirteen

91=7×13 has two factors strictly between one and ninety-one, proving it composite. Since 91 leaves remainder one on division by three, gcd(3,91)=1; this is not a failure caused by a shared factor. Put “known composite” first: passing a congruence cannot overturn factorisation. Replace “looks composite” with the explicit product rather than judging decimal endings or digit count. Seven and thirteen are factors, not two modular remainders, and no real cryptographic system is needed.

## Reduce the sixth power of three to remainder one

3^6=729=8×91+1, so its remainder modulo 91 is one. Since 90=6×15, 3^90=(3^6)^15 also leaves one. No full decimal expansion of the enormous power is needed. Keep exponent six and modulus ninety-one distinct; do not write 3^91 or mistake quotient eight for the remainder. Retain the 729=8×91+1 ledger and indicate the power rule with a derivation arrow, not a random state transition.

## Use the base-three pass to refute only the converse

Here composite n passes a^(n−1)≡1 for coprime base three, making 91 a Fermat pseudoprime to base three. Write “prime ⇒ pass; pass ⇏ prime”, placing 91 beside the failed converse rather than crossing out the original theorem. One counterexample refutes that converse, not a claim that every composite passes base three or that repeated passes automatically give absolute proof. Keep the base and test form explicit; an ordinary Fermat pass is not a pass of a stronger, different test.

## Switch to base two and obtain remainder sixty-four

Base two is also coprime to 91. Since 2^12=4096=45×91+1 and 90=12×7+6, 2^90 has the remainder of 2^6, namely 64, not one. By the theorem’s contrapositive, this failure proves nonprimality, agreeing with 7×13. After changing the base, retain exponent 90 and modulus 91, not twelve or six. Show base-three pass beside base-two failure: the base changed, not layer order or random numerical error.

## Use prime seven with its own exponent as a control

In a separate frame take prime p=7, a=3: 3^(7−1)=729=104×7+1, as the theorem predicts. Its exponent is six and modulus seven, not the old ninety and ninety-one. Keep it separate from the composite pass so “remainder one” does not erase differing conditions. Use standard remainders from zero to modulus minus one: 64 is valid modulo 91, not something to reduce below three. A pass/fail label without base, exponent, modulus and remainder is unverifiable; retain all four.

```text
91=7*13; gcd(3,91)=gcd(2,91)=1
3^6=729=8*91+1; 90=6*15
3^90 mod 91=1, but 91 is composite
2^12=4096=45*91+1; 90=12*7+6
2^90 mod 91=2^6 mod 91=64 != 1
prime control p=7: 3^6=104*7+1
prime => Fermat pass; Fermat pass does not imply prime
```

## Diagnose quotients, remainders and implication arrows

If the base-three result is eight, separate quotient and remainder in 729÷91. If base two gives one, retain the leftover exponent six after dividing ninety by twelve. Remove the pass-to-prime arrow and attach its counterexample without changing the true theorem. Keep the faulty draft and reconstruct each quotient×modulus+remainder, independently checking powers by repeated squaring externally if needed. Do not imagine automatic app arithmetic in place of the ledger. On restoration, check all conditions in both 91 examples and the prime-seven control, avoiding half-edited copied text.

## Reopen and reconstruct conclusions from factors and powers

Save native work and integer calculations, export SVG or PDF and open a PNG separately. At fit-to-window size inspect one-way implications; enlarge for intact exponents 90, 6, 12, moduli 91, 7 and negated arrows. Reopen editable work, hide pass labels, recompute factors and both remainders, then restore necessary-but-not-sufficient wording. Deliver equations, counterexample, prime control and pending output steps. The source supplies number-theory definitions, not an app primality tool, random probability guarantee or real cryptographic security assessment.

## Related tutorials and sources

[VectorPaint factor trees for seventy-two: retain repeated primes and the leaf product](<https://allluckly.cn/en/guides/vectorpaint-seventy-two-prime-factor-tree/>); [VectorPaint sieving up to thirty: retain each prime while removing its multiples](<https://allluckly.cn/en/guides/vectorpaint-eratosthenes-sieve-two-to-thirty-grid/>).

Method sources: [Matt Kerr number theory notes: Fermat theorem and pseudoprimes](<https://www.math.wustl.edu/~matkerr/NTCbook.pdf>). The notes establish the implication and base-specific Fermat pseudoprime terminology. All integer calculations for 91, bases three/two and prime seven are independently verified, without importing other tests’ conclusions.

Version and operation scope: checked on 19 November 2026 against the [official VectorPaint listing](<https://apps.apple.com/cn/app/id1587272880>) and [Apple public lookup](<https://itunes.apple.com/lookup?id=1587272880&country=cn&entity=software&timestamp=20261119-vectorpaint>): version 1.0.8, requiring iOS / iPadOS 17.0 or later. Draw manually with published shapes, pen, Bézier nodes, grids/guides, alignment, layers/groups, text and file output through visible installed controls. Prepare definitions and calculations externally; infer no built-in formula, statistics or graph solver. Keep editable native work and inspect SVG, PDF or PNG as needed. SVG/PDF containing imported bitmaps do not automatically become entirely vector. Canvases and samples are original teaching settings, without physical-device execution or real measurement; mark unperformed actions pending.

