# Geometric Insights into the Goldbach Conjecture

Frank Vega\
*Information Physics Institute, 840 W 67th St, Hialeah, FL 33012, USA*\
[e-mail]\
ORCID: 0000-0001-8210-4126

---

## Abstract

The binary Goldbach conjecture states that every even integer greater than 2 is the sum of two primes. We analyze a variant of this conjecture, positing that every even integer $2N \ge 8$ is the sum of two distinct primes $P$ and $Q$. We establish a novel equivalence between this statement and a geometric construction: the conjecture holds if and only if for every $N \ge 4$, there exists an integer $M \in [1, N-3]$ such that the L-shaped region $N^2 - M^2$ (between nested squares) has a semiprime area $P \cdot Q$, where $P=N-M$ and $Q=N+M$. We define the set $D_N$ of all such valid $M$ values for a given $N$. The conjecture is equivalent to there existing an $M \in D_N$ with $N-M$ prime. We conduct a computational analysis for $N \le 2^{14}$ and define a gap function $G(N) = \log^{2} (2N) - ((N-3) - |D_N|)$. Our experimental results show that the minimum of $G(N)$ is positive and increasing across intervals $[2^m, 2^{m+1}]$. This empirically-derived result, $G(N) > 0$, provides strong computational evidence that $|D_N| > (N-3) - \log^{2} (2N)$. Under this computationally-supported bound, the pigeonhole principle on the cardinality of $D_N$ and the number of primes $P < N$ (corresponding to squares $S_P$) implies $|D_N| \ge 1$ for all $N \ge 4$, yielding a conditional proof of the conjecture. While an analytical proof of this bound remains an open problem, our work establishes a novel geometric framework and demonstrates its viability through extensive computation.

**Keywords:** Goldbach conjecture; geometric construction; semiprimes; pigeonhole principle; computational number theory

**MSC:** 11P32, 51M15, 11A25, 11Y11

---

## 1. Introduction

The Goldbach conjecture, one of the most enduring unsolved problems in number theory, posits that every even integer greater than 2 can be expressed as the sum of two prime numbers. The strong form of this conjecture, often referred to as the binary Goldbach conjecture, remains unproven despite extensive computational verification for numbers up to very large magnitudes. In this note, we explore a slight modification of this conjecture and provide a geometric interpretation that links it to properties of squares and semiprimes.

Specifically, consider the following variant: *Every even integer greater than or equal to 8 is the sum of two distinct prime numbers.* This adjustment excludes the cases $4 = 2 + 2$ (the only sum of identical primes) and $6 = 3 + 3$, aligning with our geometric construction which requires distinct factors $P \neq Q$ and $N \geq 4$ (implying $2N \geq 8$).

We establish an equivalence between this variant and a geometric statement involving nested squares. This connection not only offers a novel visualization but also highlights the interplay between arithmetic progressions and the factorization of differences of squares.

---

## 2. Geometric Construction

Consider a square $S_N$ with integer side length $N \geq 4$, so its area is $N^2$. Inscribe within $S_N$ another square $S_M$ with integer side length $M$, where $1 \leq M \leq N-3$, such that $S_M$ shares one corner with $S_N$ (without loss of generality, the bottom-left corner). The region between $S_N$ and $S_M$ forms an L-shaped annulus with area $N^2 - M^2$.

The difference of squares factors as

$$N^2 - M^2 = (N - M)(N + M).$$

Define $P = N - M$ and $Q = N + M$. We analyze the constraints on $P$ and $Q$ imposed by the bounds on $M$:

- $M \ge 1 \implies P = N - M \le N - 1$ and $Q = N + M \ge N + 1$.
- $M \le N - 3 \implies P = N - M \ge N - (N - 3) = 3$.

Thus, we have $3 \le P \le N-1$ and $Q \ge N+1$. Since $M \ge 1$, it is clear that $P < Q$.

The sum of these factors is

$$P + Q = (N - M) + (N + M) = 2N,$$

which is an even integer $\ge 8$ (since $N \geq 4$).

The difference of the factors is $Q - P = (N + M) - (N - M) = 2M$, which is also even.

Since $P$ and $Q$ have an even sum ($2N$) and an even difference ($2M$), they must have the **same parity**. For $P$ and $Q$ to both be prime with $P \ge 3$, they must both be **odd primes**. Therefore, $P$ and $Q$ must be distinct odd primes.

The area $N^2 - M^2 = P \cdot Q$ is a semiprime (a product of two primes) if and only if both factors $P$ and $Q$ are prime. Our construction requires them to be distinct odd primes.

The variant of Goldbach's conjecture is therefore equivalent to the following geometric assertion:

**Theorem (Geometric Goldbach Variant):** For every integer $N \geq 4$, there exists an integer $M$ with $1 \leq M \leq N-3$ such that the L-shaped region between the squares $S_N$ and $S_M$ (sharing a corner) has area equal to a semiprime $P \cdot Q$, where $P = N - M$ and $Q = N + M$ are distinct primes.

This equivalence holds bidirectionally: If $2N = p + q$ for distinct primes $p < q$, we can set $N = (p + q)/2$.

- Since $2N \ge 8$ and $p, q$ are distinct, they must be distinct odd primes.
- Since $p$ and $q$ are odd, their sum and difference are even, so $N = (p+q)/2$ and $M = (q-p)/2$ are integers.
- We must check that $M$ satisfies $1 \le M \le N-3$.
- $M \ge 1$: $(q-p)/2 \ge 1 \rightarrow q-p \ge 2$. Since $p, q$ are distinct odd primes, this is true.
- $M \le N-3$: $(q-p)/2 \le (p+q)/2 - 3 \rightarrow q-p \le p+q - 6 \rightarrow -p \le p - 6 \rightarrow 6 \le 2p \rightarrow 3 \le p$. Since $p$ is an odd prime, $p \ge 3$, so this is always satisfied.

Thus, any Goldbach partition $2N = p+q$ (with distinct odd primes) corresponds exactly to a valid geometric construction with $P=p$, $Q=q$, and $M=(q-p)/2$, yielding the semiprime area $p \cdot q$.

Conversely, given a valid geometric construction with $P = N - M$ and $Q = N + M$ both prime, then $P + Q = 2N$, so $2N$ is the sum of two distinct primes.

To illustrate, Figure 1 depicts the construction for a generic $N$ and $M$.

## Geometric Construction


**Figure 1:** Geometric construction illustrating the L-shaped semiprime region between nested squares of sides $N$ and $M$ sharing the origin corner $O$. The horizontal extension of length $P = N - M$ and vertical extension of length $Q = N + M$ bound the region of area $P \cdot Q = N^2 - M^2$. For $N=5$, $M=2$, $P=3$, $Q=7$ (both prime), area $25-4=21=3 \cdot 7$, and $3+7=10=2 \cdot 5$.

---

## 3. Deeper Analysis and Implications

The above equivalence provides a bridge between additive number theory and geometric dissections. For a fixed $N \geq 4$, the conjecture is true if there exists at least one $M \in \{1, 2, \dots, N-3\}$ such that $P = N-M$ and $Q = N+M$ are both prime.

Since $Q = N + M = N + (N-P) = 2N - P$, this is precisely the statement that for a given $N \ge 4$, there exists a prime $P$ in the range $[3, N-1]$ such that $2N-P$ is also prime.

This is exactly the Goldbach partition for $2N$ into two primes $P$ and $Q$. As $N \ge 4$, $2N \ge 8$. As shown, $P$ and $Q$ must be distinct odd primes. The geometric view, therefore, recasts the search for a Goldbach partition as a search for an integer $M$ that defines an L-shaped "frame" with a semiprime area.

Computational evidence strongly supports the conjecture: For $N$ up to $2 \times 10^{17}$ (corresponding to even numbers up to $4 \times 10^{18}$), such a partition has always been found. Analytically, the Hardy–Littlewood conjecture estimates that the number of such partitions $g(2N)$ grows as $g(2N) \sim 2 C_2 \frac{2N}{(\log(2N))^2}$, suggesting not only that at least one partition exists, but that the number of them grows with $N$.

---

## 4. Novel Approach on this Perspective

From the previous analysis, we have $P = N - M$ and $Q = N + M$. The area $P \cdot Q = (N-M)(N+M)$ admits a natural geometric interpretation via the L-shaped region in Figure 1.

This region consists of a vertical rectangle of dimensions $M \times P$ (the top arm) and a horizontal rectangle of dimensions $N \times P$ (the right arm, accounting for the full height $N$), yielding total area $MP + NP = P(N + M) = P \cdot Q$.

Alternatively, the dashed square $S_P$ (of side $P = N - M$, positioned from the top-right corner of $S_M$ to the top-right corner of $S_N$) highlights a complementary decomposition: the L-shaped region equals the area of $S_P$ (namely $P^2$) plus two adjacent rectangles each of area $P \times M$ (one horizontal along the bottom-right and one vertical along the top-left, excluding the overlap covered by $S_P$). This confirms $N^2 - M^2 = P^2 + 2PM$, which expands to $(N-M)^2 + 2M(N-M) = N^2 - M^2$, verifying the identity tautologically.

The key geometric insight, however, derives from relating $P$ and $Q$ directly to the side lengths $N$ and $M$. Observe that $P = N - M$ measures the extension beyond $S_M$ along either arm of the L-shape (horizontal width or vertical height). The full side $N$ of $S_N$ thus decomposes as the inner side $M$ plus this extension $P$, so $N = M + P$. Meanwhile, $Q = N + M$ extends this by adding the inner side $M$ once more—geometrically, $Q$ spans the full height $N$ of $S_N$ plus the height $M$ of $S_M$, evoking a "doubled" vertical traversal from the shared corner $O$ outward and inward.

Subtracting these lengths yields

$$Q - P = (N + M) - (N - M) = 2M,$$

and solving gives the explicit formula

$$M = \frac{Q - P}{2}.$$

For $N \geq 4$ (ensuring $2N \geq 8$), any distinct primes $P < Q$ with $P + Q = 2N$ (both odd, hence $Q - P$ even) produce an integer $M \geq 1$ satisfying $M \leq N-3$, as $Q \leq 2N - 3$ (next prime after $P \geq 3$) implies $M \leq (2N - 6)/2 = N - 3$. This bidirectionally links the arithmetic partition to the geometric embedding. Instead of focusing on the semiprime area $P \cdot Q$, we now focus on the set of distinct values of $M$ generated by straddling prime pairs.

For a given $N \ge 4$, let $D_N$ be the set of all integers $M$ such that

$$D_N = \left\{M = \frac{Q - P}{2} \mid 2 < P < N < Q < 2N, \text{ and } P, Q \text{ are prime}\right\}.$$

Note that we implicitly consider $Q < 2N$ to ensure finite cardinality relevant to the scale of $N$, aligning with the geometric bounds. The Goldbach variant requires that there exists $M \in D_N$ with $N - M$ prime (ensuring $P = N - M$ prime and $Q = N + M$ prime via the partition).

The question becomes: How many distinct values $M \in \{1, 2, \dots, N-3\}$ are in the set $D_N$? For example, $|D_4| = 2$ since $D_4 = \{(5 - 3)/2, (7-3)/2\} = \{1, 2\}$.

### 4.1 Experimental Results

We conducted a computational experiment to evaluate the size of $D_N$ for every $N$ between $4$ and $2^{14}$. The experiment was performed on a standard workstation (11th Gen Intel i7, 32GB RAM) using a Python 3.12 implementation with the Gmpy2 library.

#### Code

The computational verification is available in this repository: [Experimental Results on Goldbach's Conjecture](https://github.com/frankvegadelgado/goldbach). Run `python experiment.py` to reproduce the results in Table 1.

#### Data

For each $N$ in the range, we compute $|D_N|$.

We then calculate a "gap value" $G(N)$ defined as:

$$G(N) = \log^{2} (2N) - ((N-3) - |D_N|),$$

where $|D_N|$ is the count of distinct $M$-values for $N$. The experiment is deterministic and yields the results summarized in Table 1.

**Table 1: Minimum $G(N)$ values in intervals $[2^m, 2^{m+1}]$**

| Interval ($m$) | Range ($[2^m, 2^{m+1}]$) | Minima at $N$ | Min $G(N)$ Value |
|----------------|---------------------------|---------------|------------------|
| 2 | [4, 8] | 5 | 4.301898 |
| 3 | [8, 16] | 11 | 7.554543 |
| 4 | [16, 32] | 17 | 10.435219 |
| 5 | [32, 64] | 61 | 14.078618 |
| 6 | [64, 128] | 73 | 17.836335 |
| 7 | [128, 256] | 151 | 20.608977 |
| 8 | [256, 512] | 269 | 23.537165 |
| 9 | [512, 1024] | 541 | 28.812111 |
| 10 | [1024, 2048] | 1327 | 33.154668 |
| 11 | [2048, 4096] | 2161 | 35.081569 |
| 12 | [4096, 8192] | 7069 | 42.329014 |
| 13 | [8192, 16384] | 14138 | 44.057758 |

The results in Table 1 show that $G(N)$ is consistently positive. This implies that the number of "gaps" (values of $M$ not in $D_N$, approximated by $(N-3) - |D_N|$) is less than $\log^{2} (2N)$.

More importantly, the minimum value of $G(N)$ within each successive power-of-two interval is strictly increasing. This empirical finding is the basis for the following lemma.

---

## 5. Ancillary Results

This is a key finding based on the experimental data.

**Lemma (Key - Computational):**
For $m \geq 2$ and $N \le 2^{14}$, the minimum value of $G(N)$ in the interval $[2^{m}, 2^{m + 1}]$ is strictly less than the minimum value of $G(N)$ in the interval $[2^{m + 1}, 2^{m + 2}]$.

**Proof:**
This lemma is an empirical observation from the computational data in Table 1, verified for all $N$ in the range $[4, 2^{14}]$. The data shows the minima of $G(N)$ strictly increase across dyadic intervals for $m \geq 2$. We provide a heuristic justification for why this trend should continue.

A concise heuristic justification relies on the incremental growth of $|D_N|$, the cardinality of the set of valid $M$-values (i.e., the number of distinct prime pairs $(P, Q)$ with $P < N < Q$ and $P + Q = 2N$).

**Local increment from $N$ to $N+1$**

When $N$ advances to $N+1$, almost all previously realized $M$-values persist, but two specific candidates can newly appear:

- **Case 1 (Boundary shift via $Q=2N-1$):**

$$
P=3,\quad Q=2N-1,\quad M=\frac{Q-P}{2}=N-2.
$$

This contributes a new $M$ if $2N-1$ is prime.

- **Case 2 (Twin-prime emergence):**
If $(N, N+2)$ are both prime, then

$$
P=N,\quad Q=N+2,\quad M=\frac{Q-P}{2}=1,
$$

which contributes a new $M=1$.

Thus,

$$
|D_{N+1}| \geq |D_N| + \delta, \qquad \delta \in \{0,1,2\},
$$

where $\delta$ counts how many of the above cases might occur.

**Growth over larger scales**

Over ranges like $[N,2N]$, primes $Q>n$ enter $(n,2n)$ with regularity guaranteed by refinements of Bertrand’s postulate. Specifically:

- **Nagura’s theorem:** For $n \geq 25$, there is always a prime in $(n, (6/5)n]$, implying at least $\gtrsim 5$ primes in $(n,2n)$ on average.
- **Dusart’s refinement:** For $n \geq 3275$, primes exist in intervals as short as $(n, n+n/(2\log^2 n)]$, yielding $\sim 2\log^2 n$ primes in $(n,2n)$.

Each new $Q$ can pair with many prior primes $P<n$ to realize new $M=(Q-P)/2$. With $\pi(n-1)-1 \sim n/\log n$ available $P<n$, and new $Q$ appearing at density $\sim 1/\log n$, the expected influx of new pairs per $n$ is $\gtrsim n/\log^2 n$. Conservatively, Bertrand-type results ensure at least $\Omega(\log n)$ new incorporations per $n$.

Over $[N,2N]$, the total new $Q$ in $(N,4N)$ number $\pi(4N)-\pi(N)\sim 3N/\log N$, each potentially adding up to $\sim N/\log N$ pairs, distributed across $\sim N$ values of $n$. This yields an average $|D_n|$ growth of $\Omega(N/\log^2 N)$ per step. Consequently, gaps $(n-3)-|D_n| \sim n-\Omega(n/\log^2 n)$ increase nearly linearly, but the logarithmic term ensures the deficit grows slower than $\log^2(2n)$, so $G(n)$ trends upward.

**Extension to dyadic intervals**

Extending to dyadic intervals $[2^m,2^{m+1}]$ as $m\to\infty$, each such interval contains a power of two anchoring the empirical minima observed in Table 1. The refined Bertrand bounds amplify with $m$: new $Q$ in $[2^m,2^{m+1}]$ number $\sim 2^m/m$, each incorporating $\sim 2^m/m$ prior $P<2^m$, adding $\sim 2^m/m^2$ per $n$ on average. Cumulatively, $|D_n|\gtrsim c\cdot 2^m/m^2$ for some $c>0$, with gaps $\leq 2^m(1-c/m^2)$. The increase in $\log^2(2n)$ across $[2^{m+1},2^{m+2}]$ is $O(m)$, outpacing the relative gap shrinkage $O(1/m^2)$, ensuring the minimum $G(n)$ rises, consistent with the data.

---

**Remark:** This empirical pattern strongly suggests the inequality holds universally, though a rigorous analytical proof remains an open problem. □

This is a main insight, derived from the data.

**Corollary (Computational Evidence):** For all $N$ in the range $[4, 2^{14}]$, we have $|D_N| > (N-3) - \log^{2} (2N)$. The empirical data strongly suggests this inequality holds for all $N \geq 4$.

**Proof:** This is a direct consequence of the experimental data presented in Table 1. The data shows that $G(N) > 0$ for all $N$ tested (specifically, the minimum value in each interval is positive and increasing). By definition, $G(N) = \log^{2} (2N) - ((N-3) - |D_N|)$. The condition $G(N) > 0$ directly implies:

$$\log^{2} (2N) - ((N-3) - |D_N|) > 0$$

Rearranging gives:

$$|D_N| > (N-3) - \log^{2} (2N)$$

The Key Lemma shows this bound not only holds but strengthens as $N$ increases within our tested range. While this provides compelling evidence for the universal validity of the inequality, a rigorous analytical proof extending beyond our computational range would constitute a significant theoretical advance. □

**Conjecture:** The inequality $|D_N| > (N-3) - \log^{2} (2N)$ holds for all $N \geq 4$.

---

## 6. Main Result

**Theorem (Conditional on Computational Bound):** If the inequality $|D_N| > (N-3) - \log^{2} (2N)$ holds for all $N \geq 4$, then the variant Goldbach conjecture is true: every even integer greater than or equal to 8 is the sum of two distinct prime numbers.

Our computational verification for $N \le 2^{14}$ confirms $|D_N| \ge 1$ in all cases, with the bound consistently satisfied and strengthening over successive intervals.

**Proof:** As established in Section 2, the conjecture is true if and only if for every $N \geq 4$, there exists at least one pair of distinct primes $(P, Q)$ such that $P+Q=2N$ with $P < N < Q$. This is equivalent to there existing a prime $P < N$ such that $M = N - P \in D_N$ (since then $Q = 2N - P = N + M$, and the geometric construction holds).

The candidate $M$ values are $M_P = N - P$ for each prime $P \in [3, N-1]$, giving $\pi(N-1) - 1$ candidates in $\{1, \dots, N-3\}$. The "good" $M$ are those in $D_N$, while the "bad" $M$ (with no straddling prime pair of difference $2M$) number fewer than $\log^2(2N)$ **if** the computational bound holds.

By the pigeonhole principle, if the number of candidates $\pi(N-1) - 1$ exceeds the number of bad $M$, then at least one candidate $M_P$ must be good, i.e., $M_P \in D_N$. Known lower bounds give $\pi(N) > \frac{N}{\log N + 2}$ for $N \ge 6$ (this can be deduced from results in~\cite{dusart1998autour}), and $\frac{N}{\log N + 2} > \log^2(2N)$ for $N \ge 328$. Thus, **assuming the bound**, the strict inequality holds for $N \ge 328$, proving the conjecture for $N \ge 328$.

For the base cases $4 \le N \le 12$, we verify manually (additional examples beyond $N=8$ are included for illustration):

- **$N=4$** ($2N=8$): Candidates $P=3$; $M=1$. $D_4=\{1\}$, so candidate good. Partition: $3+5$. Holds, $|D_4|=1$.
- **$N=5$** ($2N=10$): Candidates $P=3$; $M=2$. $D_5=\{2\}$, so $M=2$ good ($P=3$). Partition: $3+7$. Holds, $|D_5|=1$.
- **$N=6$** ($2N=12$): Candidates $P=3,5$; $M=\{3,1\}$. $D_6=\{1, 2, 3\}$, so all good. Partition: $5+7$. Holds, $|D_6|=3$.
- **$N=7$** ($2N=14$): Candidates $P=3,5$; $M=\{4,2\}$. $D_7=\{3, 4\}$, so $M=4$ good ($P=3$; $Q=11$ prime). Partition: $3+11$. Holds, $|D_7|=2$.
- **$N=8$** ($2N=16$): Candidates $P=3,5,7$; $M=\{5,3,1\}$. $D_8=\{2, 3, 4, 5\}$, so $M=3,5$ good ($P=5,3$; $Q=11,13$ prime). Partitions: $3+13$, $5+11$. Holds, $|D_8|=4$.
- **$N=9$** ($2N=18$): Candidates $P=3,5,7$; $M=\{6,4,2\}$. $D_9=\{2, 3, 4, 5, 6\}$, so $M=2,4,6$ good ($P=7,5$; $Q=11,13$ prime; note $P=3$, $Q=15$ not prime but others are). Partitions: $5+13$, $7+11$. Holds, $|D_9|=5$.
- **$N=10$** ($2N=20$): Candidates $P=3,5,7$; $M=\{7,5,3\}$. $D_{10}=\{2, 3, 4, 5, 6, 7\}$, so $M=3,5,7$ good ($P=7,5,3$; $Q=13,15$ not, $17$ prime). Partitions: $3+17$, $7+13$. Holds, $|D_{10}|=6$.
- **$N=11$** ($2N=22$): Candidates $P=3,5,7$; $M=\{8,6,4\}$. $D_{11}=\{3, 4, 5, 6, 7, 8\}$, so $M=4,6,8$ good ($P=7,5,3$; $Q=15$ not, $17,19$ prime). Partitions: $3+19$, $5+17$. Holds, $|D_{11}|=6$.
- **$N=12$** ($2N=24$): Candidates $P=3,5,7,11$; $M=\{9,7,5,1\}$. $D_{12}=\{1, 3, 4, 5, 6, 7, 8, 9\}$, so $M=1,5,7,9$ good ($P=11,7,5$; $Q=13,17,19$ prime; $P=3$, $Q=21$ not). Partitions: $5+19$, $7+17$, $11+13$. Holds, $|D_{12}|=8$.

For $13 \le N \le 327$, the conjecture holds by direct computational verification (included in our analysis up to $N=2^{14}$).

Thus, **under the assumption that the computationally-observed bound extends to all $N$**, the conjecture holds for $N \ge 4$. □

**Computational Verification:** Direct computation confirms the theorem's conclusion for all even integers up to $2 \times 2^{14} = 32{,}768$.

---

## 7. Conclusion

We have presented a novel geometric perspective on a variant of the Goldbach conjecture, reformulating the additive problem $2N = P + Q$ as a geometric search for an integer $M$ such that the area $N^2 - M^2$ is a semiprime $P \cdot Q$. By redefining $D_N$ to capture all achievable $M$ from straddling prime pairs, we analyzed its cardinality using computational data.

Our computational analysis establishes that $|D_N| \ge 1$ for all $N \le 2^{14}$, with the gap function $G(N)$ exhibiting a robust positive trend that strengthens across successive intervals. This provides strong empirical support for the conjecture that every even integer greater than or equal to 8 is the sum of two distinct prime numbers.

### Significance and Future Work

This work makes several contributions:

1. **Novel Geometric Framework:** We establish a rigorous equivalence between Goldbach partitions and semiprime areas in nested square constructions, offering fresh geometric intuition for a classical arithmetic problem.

2. **Computational Evidence:** Extensive verification up to $N = 2^{14}$ demonstrates the viability of our approach and reveals consistent patterns in the distribution of valid configurations.

3. **Conditional Proof Strategy:** We show that an analytical bound on $|D_N|$ would immediately yield a proof via the pigeonhole principle, reducing the problem to establishing a single inequality.

4. **Open Problem:** The central challenge is to prove analytically that $|D_N| > (N-3) - \log^{2} (2N)$ for all $N \geq 4$. Our heuristic arguments based on prime distribution theory (Bertrand's postulate refinements) suggest promising directions, but a rigorous proof remains elusive.

This work highlights how geometric reformulation combined with computational exploration can illuminate classical problems and suggest concrete paths toward resolution. The gap between computational evidence and universal proof underscores both the power and limitations of empirical methods in number theory.

---

*MSC (2020):* 11P32 (Goldbach-type theorems; other additive questions involving primes), 51M15 (Geometric constructions in real or complex geometry), 11A25 (Arithmetic functions; related numbers; inversion formulas), 11Y11 (Primality testing; factorization)

---

## Documentation
Available as PDF at [Geometric Insights into the Goldbach Conjecture](https://github.com/frankvegadelgado/goldbach).