Abstract#

Prime numbers are often modeled using probabilistic frameworks, such as Cramér's random model, to understand macro-level distribution patterns. However, local prime behavior is strictly governed by deterministic arithmetic properties. This paper presents empirical evidence from a computational analysis of prime pairs up to 2 million, examining the distributional patterns of twin primes (gap 2), cousin primes (gap 4), and sexy primes (gap 6) relative to a harmonic lattice of step 6. Our findings reveal that approximately $99\%$ of twin and cousin primes, and approximately $49\%$ of sexy primes (with the remaining $49\%$ structurally excluded by modulo 3 constraints), have their information potential value $I = 2p + (g - 1)$ land within $\pm 48$ of another prime at multiples of 6. The data exhibits systematic harmonic preferences, with full-step multiples of 12 showing significantly higher frequencies than half-step intervals. These results highlight the powerful, localized deterministic constraints that operate within the pseudo-random macro-distribution of primes, demonstrating that local configurations of prime pairs are heavily governed by strict, underlying modular architectures.

Introduction

The distribution of prime numbers has fascinated mathematicians for centuries. While the Prime Number Theorem establishes that $\pi (N)\sim N / \log N$, the local behavior of primes particularly the gaps between consecutive primes -- has resisted simple characterization [1]-[3].

The dominant macro-level conceptual framework for understanding prime gaps has been probabilistic, originating with Cramér's 1936 random model. This model treats primes as independent random events with probability $1 / \log n$, successfully predicting the scale of typical prime gaps. However, number theorists have long recognized that this "randomness" is a heuristic modeling tool; primes are inherently deterministic, and probabilistic models must be refined by arithmetic constraints, such as divisibility by small primes [3].

Within this paradigm, special prime pairs -- twin primes (gap 2), cousin primes (gap 4), and sexy primes (gap 6) -- have been studied extensively. Previous research has noted certain regularities in prime gaps. Szpiro [4] observed that gaps of size 6 and multiples of 6 occur more frequently than expected under unrefined random models, attributing this to the fact that all primes greater than 3 are of the form $6k \pm 1$.

The present study extends this line of inquiry by examining not merely the gaps between primes themselves, but the relationship between prime pairs and a fixed harmonic lattice of step 6. Specifically, we investigate whether for each prime pair $(p, p + g)$ with $g \in \{2, 4, 6\}$, the quantity $I = 2p + (g - 1)$ tends to land near another prime at a multiple-of-6 offset.

Our results reveal that this relationship holds with remarkable consistency: approximately $99\%$ of twin and cousin primes, and approximately $49\%$ of sexy primes, satisfy this property within the range $\pm 48$. Rather than contradicting probabilistic models, this near-universal participation illustrates the intense, deterministic rigidity of the local structural corridors in which prime pairs reside.

Context and Novelty of the Present Study#

The theoretical foundations of deterministic constraints within prime distributions are well-established. Dirichlet's theorem on arithmetic progressions (1837) proved that sequences of the form $ak + b$ (where $a$ and $b$ are coprime) contain infinitely many primes, establishing the deterministic "corridors" in which primes must reside. In the 20th century, the Hardy-Littlewood conjectures formalized the concept of "admissible sets" -- constellations of numbers that avoid forced divisibility by small primes, thereby dramatically increasing the local density of prime occurrences.

However, while these macro-level theories are foundational, computational literature has primarily focused on either the asymptotic frequencies of specific prime gaps (e.g., Szpiro's observations on multiples of 6) or the exhaustive search for massive prime $k-tuples$.

The novelty of this present study lies in the introduction of a specific, localized heuristic construct: the Information Potential, defined as $I = 2p + (g - 1)$. To the author's knowledge, prior computational sieves have not utilized this specific combinatorial pivot to map local prime densities. By anchoring a step-6 harmonic lattice to this exact potential, this study empirical demonstrates that the theoretical concept of admissible sets translates into a remarkably highly efficient, near-certain predictive bounding.

Specifically, we demonstrate that this novel framing captures a prime $99\%$ of the time within an exceptionally tight window of just 16 offsets $(\pm 48)$ for twins and cousins. While the underlying modular mechanics (such as modulo 2 and 3 constraints) are established mathematics, the empirical demonstration of this specific algorithmic trap -- and its staggering $99\%$ efficiency -- represents a uniquely high-resolution mapping of local prime architecture that has not been previously documented in computational number theory.

Methodology

Definitions#

Following standard terminology:

• Twin primes: Pairs of primes $(p, p + 2)$

• Cousin primes: Pairs of primes $(p, p + 4)$

• Sexy primes: Pairs of primes $(p, p + 6)$

For each prime pair, we define the information potential as: $I = p + (p + g) - 1 = 2p + (g - 1)$ where $g$ is the gap (2, 4, or 6). This quantity represents the sum of the endpoints minus one.

Harmonic Offsets#

We consider offsets that are multiples of 6 from $-48$ to $+48$ (excluding zero), corresponding to 16 discrete positions: $O = \{\pm 6,\pm 12,\pm 18,\pm 24,\pm 30,\pm 36,\pm 42,\pm 48\}$ For each prime pair, we test whether $I + o$ is prime for any $o \in O$.

Computational Approach#

We generated all primes up to 4,000,049 using the Sieve of Eratosthenes, yielding 283,149 primes. For each prime pair type, we identified all pairs with $p \leq 2,000,000$, computed $I$, checked primality of $I + o$ for all offsets, and recorded frequencies. The analysis was performed using Python.

Rationale for Offset Selection#

The choice of multiples of 6 is motivated by the fact that all primes greater than 3 are congruent to $\pm 1$ (mod 6) [4]. Consequently, numbers that are multiples of 6 from any starting point will preserve the residue class modulo 6 of the base, forming a natural harmonic lattice for prime distribution.

Results

Twin Primes (Gap 2)#

For twin primes, $I = 2p + 1$. The analysis covered 14,871 twin primes up to 2 million. Overall confirmation rate: 14,701 of 14,871 twin primes (98.86%) had at least one prime at a multiple-of-6 offset within $\pm 48$.

The three most frequent offsets are $-36$ (32.75%), $+24$ (29.57%), and $+48$ (29.10%). Offsets that are multiples of 12 (even $k$) generally show higher frequencies. A sign alternation pattern emerges among even $k$: $-36$ dominates over $+36$, while $+24$ dominates over $-24$.

Cousin Primes (Gap 4)#

For cousin primes, $I = 2p + 3$. Overall confirmation rate: 14,530 of 14,741 cousin primes (98.57%). The most frequent offsets shifted to $-30$ (29.45%), $+12$ (29.22%), and $+30$ (25.97%), suggesting that preferred offsets shift with gap size.

Sexy Primes (Gap 6)#

For sexy primes, $I = 2p + 5$. Overall confirmation rate: 11,983 of 24,531 sexy primes (48.85%). This dramatically lower confirmation rate suggests a structural bifurcation within the sexy prime population, fundamentally altering the available offsets.

Offset-specific frequencies for Twin Primes#

Discussion

The Modulo 6 Constraint and Sexy Prime Bifurcation#

The striking difference in confirmation rates between twins/cousins ($\sim 99\%$) and sexy primes ($\sim 49\%$) is a direct consequence of modulo 6 arithmetic. For any prime $p > 3$, $p \equiv 1$ or 5 (mod 6).

For sexy primes ($g = 6$), two cases arise:

If $p \equiv 1$ (mod 6), then $p + 6 \equiv 1$ (mod 6), and $I = 2p + 5 \equiv 2(1) + 5 = 7 \equiv 1$ (mod 6). If $p \equiv 5$ (mod 6), then $p + 6 \equiv 5$ (mod 6), and $I = 2p + 5 \equiv 2(5) + 5 = 15 \equiv 3$ (mod 6).

When $I \equiv 3$ (mod 6), every number of the form $I + 6k$ is also $\equiv 3$ (mod 6) and therefore divisible by 3. Thus, none of the offsets can be prime. This perfectly accounts for half of all sexy primes being structurally excluded, explaining the $\sim 49\%$ confirmation rate.

Refining the Probabilistic Expectation#

At first glance, finding a prime in $99\%$ of cases seems to conflict with unrefined probabilistic models. For $I \approx 10^{6}$, $\ln (I) \approx 14.5$. The unadjusted probability of a random number being prime is roughly $1 / 14.5$. Testing 16 independent offsets at this rate yields an expected success rate of $1 - (1 - 1 / 14.5)^{16} \approx 68\%$.

However, our target $I$ is strictly constructed from prime pairs. It is not a random integer. By searching only at $6k$ offsets, we strictly search within a valid modulo 6 class, eliminating divisibility by 2 and 3 and immediately doubling the baseline probability to $\sim 13.8\%$. Furthermore, because $p$ and $p + g$ are primes, $I$ is heavily skewed to be coprime to 5 and 7 as well. Incorporating these localized deterministic constraints pushes the local density of primes in these specific offsets to approximately 20-25%.

Given a $\sim 25\%$ localized probability of any specific offset being prime, the chance of finding at least one prime across 16 offsets is $1 - (1 - 0.25)^{16} \approx 0.99$. Therefore, the 99% empirical confirmation is a rigorous, highly accurate validation of probabilistic models once they are correctly refined by local modular architecture.

Harmonic Preferences and Sign Asymmetry#

Across all three types, multiples of 12 (even $k$) consistently show higher frequencies. Furthermore, the sign asymmetry (e.g., $-36$ appearing at $32.75\%$ while $+36$ is $18.57\%$) heavily implies modulation by higher moduli, such as modulo 7 or modulo 11. These underlying deterministic constraints bias the distribution of successful $k$ values depending on the specific residues of $I$.

Conclusion

This study demonstrates that twin, cousin, and sexy primes exist within highly structured, deterministic corridors. Approximately $99\%$ of twin and cousin primes, and $49\%$ of sexy primes (with the remainder systematically excluded by modulo 3), find a prime within $\pm 48$ at a multiple-of-6 offset. Rather than negating probabilistic models, these findings powerfully illustrate how macro-level pseudo-randomness is built upon strictly ordered, local modular arithmetic. Future work should mathematically map the modulo 7 and modulo 11 constraints to formally prove the observed sign asymmetries and offset preferences.

References#

1. E. Hibbs, "Applying the new primer on prime numbers," WSEAS Transactions on Mathematics, April 2010.
2. "Prime k-tuple," Wikipedia.
3. A. Dixit and S. Pathak, "Patterns In Primes Via Probability," IISER Kolkata, December 2025.
4. G.G. Szpiro, "Peaks and gaps: Spectral analysis of the intervals between prime numbers," Physica A, vol. 384, no. 2, pp. 202-206, 2007.
5. OEIS Foundation Inc., "Number of cousin primes < 10^n," OEIS, sequence A080840.
6. L. Wang and L. Wang, "Pseudo Random test of prime numbers," arXiv:math/0603450, 2006.
7. Vidvashree H R et al., "Patterns in Primes..." Int. J. of Technology, Knowledge and Society, vol. 15, no. 1, 2025.