--- title: 'Long-Period Runs: Domain-Specific Persistence' url: https://www.emergentmind.com/topics/long-period-runs type: topic --- # Long-Period Runs: Domain-Specific Persistence Searching arXiv for the cited paper and closely related work on longest runs in continued fractions and run theory. Long-period runs are studied in several technically distinct senses. In continued-fraction dynamics, they are longest blocks of repeated partial quotients, either for a fixed symbol or after maximization over all symbols. In combinatorics on words, they are maximal repetitions whose minimal period exceeds a prescribed threshold. In irregularly sampled time-series analysis, they are contiguous intervals during which a locally estimated period or amplitude remains persistently displaced from a baseline value. Related stochastic work treats long runs of persistent increments in gambler’s ruin, while fluid-dynamical work on turbulence uses the adjacent notion of a longest-period periodic orbit rather than a run statistic in the combinatorial sense [2606.00505]. ## 1. Domain-specific meanings and basic distinctions The literature does not use a single universal definition of “long-period runs.” Instead, the object depends on the ambient structure being analyzed. | Setting | Run object | Principal statistic | |---|---|---| | Continued fractions | consecutive repeated partial quotients | \(L_n(x,\lambda)\), \(R_n(x)\) | | Strings | maximal repetition with minimal period \(p\) | \(Runs(w)\), threshold \(p>T\) | | Running-Sines time series | contiguous interval of persistent local period or amplitude excursion | segments \([t_1,t_2]\) | | Persistent gambler’s ruin | maximal consecutive \(+1\) or \(-1\) increments | \(R'_N\), \(U'_N\), \(L'_N\) | A recurrent source of confusion is to identify these objects with one another. The continued-fraction problem concerns constant-symbol blocks in a countable-alphabet symbolic expansion. The string-theoretic problem concerns maximal periodic substrings with arbitrary minimal period. The Running-Sines formulation does not define a run at the symbol level at all; it defines a run as a contiguous interval in a trajectory of locally estimated parameters. These distinctions are structural rather than terminological. A second distinction concerns typical versus exceptional behavior. In continued fractions, one line of work establishes almost-sure logarithmic laws for longest runs, while another constructs full Hausdorff dimensional exceptional sets on which a prescribed digit realizes the overall maximum at an admissible growth scale. The same object is therefore studied both metrically and dimensionally. ## 2. Continued fractions: prescribed longest runs For every irrational \(x\in[0,1)\), the continued-fraction expansion is \[ x=[a_1,a_2,a_3,\dots], \] where \[ a_n(x)=\Big\lfloor 1/T^{\,n-1}(x)\Big\rfloor\in\mathbb N, \qquad T(x)=\frac1x-\lfloor 1/x\rfloor. \] Fix \(\lambda\in\mathbb N\). A run of length \(k\) of \(\lambda\) in \((a_1,\dots,a_n)\) is a string of \(k\) consecutive entries equal to \(\lambda\). The fixed-symbol longest-run up to time \(n\) is \(L_n(x,\lambda)\), the maximal block-length of \(\lambda\) among the first \(n\) partial quotients, and the overall longest-run is \[ R_n(x)=\max_{\mu\in\mathbb N}L_n(x,\mu). \] An admissible growth scale is a function \(f:\mathbb N\to\mathbb N\) that is non-decreasing, unbounded, and self-neglecting in the sense that \[ \lim_{n\to\infty}\frac{f(n+f(n))}{f(n)}=1. \] The central result shows that both the growth rate and the maximizing symbol can be prescribed in advance [2606.00505]. For \[ \mathfrak E_{\lambda,f} = \Bigl\{x\in[0,1)\setminus\mathbb Q: \lim_{n\to\infty}\frac{L_n(x,\lambda)}{f(n)}=1 \ \text{and}\ \forall n,\ \forall \mu\ne\lambda,\ L_n(x,\lambda)>L_n(x,\mu) \Bigr\}, \] one has \[ \dim \mathfrak E_{\lambda,f}=1. \] This simultaneously yields three full-dimensional exceptional-set statements: fixed-symbol exact growth, overall exact growth, and prescribed dominance. In particular, the maximizing symbol in the overall longest-run problem can be fixed in advance without any loss of Hausdorff dimension. The construction is carried out inside bounded-digit Cantor sets \[ \mathcal F_B=\{x:1\le a_n(x)\le B\ \forall n\}, \] whose dimension tends to \(1\) as \(B\to\infty\). With parameters \(M,B\) and \(B\ge \lambda+2\), stage \(k\) begins at time \(t_{k-1}\), sets a target run length \(r_k=f(t_{k-1})\), introduces a free block of length \(g_k\approx M r_k\), inserts the separator \((\lambda+1,\lambda+2)\) repeatedly, and then forces a run of \(\lambda\) of length \(s_k=f(t_{k-1}+g_k)\sim r_k\). Another separator terminates the stage. By design, the forced \(\lambda\)-run has the prescribed scale, while accidental runs of \(\lambda\) or of any other symbol outside the forced blocks are broken by separators and have length at most \(O(\sqrt{r_k})\ll r_k\). The resulting Cantor-type set \(\mathcal K_{\lambda,f,M,B}\subset \mathfrak E_{\lambda,f}\) satisfies \[ \dim\mathcal K_{\lambda,f,M,B}\ge \frac{M}{M+1}\,\dim\mathcal F_B, \] and letting \(B\to\infty\) and \(M\to\infty\) yields full dimension. The admissible scales include logarithmic and polynomial examples. When \(f(n)=\lceil c\log n\rceil\), one recovers full-dimensional sets with longest run growing like \(c\log n\). When \(f(n)=\lceil n^\theta\rceil\) for \(\theta\in(0,1)\), one obtains full-dimensional sets with polynomial longest-run growth. In every case, no other symbol ever matches the length of the \(\lambda\)-runs at any finite stage. ## 3. Quantitative longest-run laws for partial quotients The exceptional-set theorem is complemented by quantitative almost-sure laws for continued fractions. In a general ergodic, measure-preserving system with a countable-alphabet observable \(X_k\), the fixed-symbol longest-run \(L_n(\omega,m)\) and the maximized run \(R_n(\omega)\) are studied under three hypotheses: quantitative mixing, an exponential cylinder estimate for one distinguished symbol, and a summed cylinder bound over all symbols [2602.11462]. Under these assumptions, for every \(c>1/2\) and almost every \(\omega\), there is \(n_{c,\omega}<\infty\) such that \[ \Bigl|L_n(\omega,m_*)-\frac{\log n}{2\log\rho}\Bigr| \le \frac{c\,\log\log n}{\log\rho} \] for all large \(n\). If the summed cylinder bound also holds, then the same two-sided estimate applies to \(R_n(\omega)\) with the same centring constant. For the Gauss dynamical system on \([0,1)\setminus\mathbb Q\), the observable is \(X_k(x)=a_k(x)\), the \(k\)th partial quotient. The explicit centring constants are \[ c_n(\lambda)=\frac{\log n}{2\log \tau(\lambda)}, \qquad \tau(\lambda)=\frac{\lambda+\sqrt{\lambda^2+4}}{2}, \] for a fixed digit \(\lambda\), and \[ d_n=\frac{\log n}{2\log\varphi}, \qquad \varphi=\frac{1+\sqrt5}{2}, \] for the overall longest run. The almost-sure first-order laws are \[ \lim_{n\to\infty}\frac{L_n(x,\lambda)}{\frac{\log n}{2\log\tau(\lambda)}}=1, \qquad \lim_{n\to\infty}\frac{R_n(x)}{\frac{\log n}{2\log\varphi}}=1. \] The refinement is the double-logarithmic error bound: for every \(c>1/2\) and almost every \(x\), there exists \(N_{c,x}\) such that for all \(n\ge N_{c,x}\), \[ \Bigl|L_n(x,\lambda)-c_n(\lambda)\Bigr| \le \frac{c\,\log\log n}{\log\tau(\lambda)}, \qquad \Bigl|R_n(x)-d_n\Bigr| \le \frac{c\,\log\log n}{\log\varphi}. \] These results quantify the statement that, for almost every continued-fraction expansion, the longest constant block among the first \(n\) digits is of order \(\log n\). In conjunction with the prescribed-realization theorem, they separate typical logarithmic behavior from full-dimensional exceptional behavior at arbitrary admissible scales. ## 4. Long-period runs in strings In combinatorics on words, a run is a maximal repetition rather than a constant-symbol block. For a string \(w\in\Sigma^n\), an integer \(p\) is a period of \(w\) if \[ w[i]=w[i+p]\quad\forall\,1\le i\le |w|-p. \] If \(p\) is the smallest period, the exponent is \[ e(w)=\frac{|w|}{p}. \] A run, or maximal repetition, is a triple \(r=(i,j,p)\) such that \(w[i..j]\) has minimal period \(p\), length at least \(2p\), and cannot be extended while preserving that period [1406.0263]. Two fundamental bounds govern the global combinatorics of runs: \[ \rho(n)T\) is immediate in \(O(m)\le O(n)\). If threshold queries are repeated, runs can be sorted by period, and counts or summed exponents above a threshold can be answered via binary search and prefix sums, or via a segment tree or Fenwick tree. This usage differs sharply from the continued-fraction longest-run problem: the period is intrinsic to the repeated word, not to a cycle-length estimate or to a constant symbol. ## 5. Running-Sines analysis of quasi-periodic time series In the time-series literature on nearly periodic signals with irregular sampling, “long-period runs” refer to extended excursions in local period, phase, or amplitude estimated by a sliding local sinusoidal fit [1308.1129]. Given data \(\{(t_i,y_i,\sigma_i)\}\), a nominal period \(P_0\), and central times \(t_0\), one selects a window half-width \(\Delta t\), retains only points with \(|t_i-t_0|\le \Delta t\), and defines weights \[ w_i = p\!\left(\frac{t_i-t_0}{\Delta t}\right)\!/\,\sigma_i^2. \] With the rectangular filter, \(p(z)=1\) for \(|z|\le 1\). The local model is \[ f(t;C_1,C_2,C_3)=C_1+C_2\cos[\omega(t-T_0)]+C_3\sin[\omega(t-T_0)], \qquad \omega=\frac{2\pi}{P_0}, \] equivalently \[ f(t)=a_0+A\cos[\omega(t-T_0)+\phi], \] with \[ a_0=C_1,\qquad A=\sqrt{C_2^2+C_3^2},\qquad \phi=\arctan2(-C_3,C_2). \] The coefficients are obtained by minimizing the weighted least-squares criterion \(\chi^2(t_0,\Delta t)\), with normal equations \(A_{\alpha\beta}=\sum w_i f_\alpha(t_i)f_\beta(t_i)\), \(B_\alpha=\sum w_i y_i f_\alpha(t_i)\), and solution \(C=A^{-1}B\). Formal uncertainties follow from \(\mathrm{cov}(C)=\sigma_0^2A^{-1}\), where \(\sigma_0^2=\chi^2/(M-m)\). If period drift is suspected, the fit may be repeated over a small grid of trial periods around \(P_0\), choosing the period that minimizes \(\chi^2(t_0,P)\). In practice, only small fractional deviations \(\Delta P/P_0\lesssim 0.3\)–\(0.5\) are usually allowed for stability. The extraction of long-period runs proceeds by computing a global periodogram, choosing \(\Delta t\) typically near \(0.5P_0\), and then examining the local trajectories \(A(t_0)\), \(\phi(t_0)\), and possibly \(P(t_0)\). A run may be defined as a contiguous interval \([t_1,t_2]\) during which \(P(t_0)\) stays \(>P_0+\delta P\) or \(