Sharp Asymptotics for the Largest Component in the Subcritical Regime of Preferential Attachment Without Vertex Growth
Pith reviewed 2026-07-02 07:12 UTC · model grok-4.3
The pith
In the subcritical regime of preferential attachment on a fixed vertex set, the largest component L1 is (1+o_p(1)) times 2(α+2)/(α+1) times ε^{-2} log(ε³ n).
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
We prove that if m = m_c(1-ε), ε=ε(n)→0, ε³n→∞, then L1=(1+o_p(1)) * 2(α+2)/(α+1) * ε^{-2} log(ε³ n) for every fixed α>0. The same asymptotic holds when α=α(n)→a∈(0,∞]. In particular the constant converges to the Erdős–Rényi value 2 as α→∞. Moreover if m=⌊n/2(1-ε)⌋ and αε→∞ then L1=(2+o_p(1))ε^{-2}log(ε³ n). The upper bound uses that after conditioning on the degree sequence the graph can be treated through the corresponding configuration model; the lower bound follows from tree component asymptotics and a second moment argument.
What carries the argument
Conditioning on the degree sequence to treat the graph as a configuration model for the upper bound on L1, together with tree-component asymptotics and the second-moment method for the lower bound.
If this is right
- The leading constant 2(α+2)/(α+1) approaches the Erdős–Rényi value of 2 as α tends to infinity.
- The asymptotic continues to hold when the parameter α grows or shrinks with n as long as it tends to a positive limit.
- In the special case where αε tends to infinity the size simplifies to (2 + o_p(1)) ε^{-2} log(ε³ n).
Where Pith is reading between the lines
- The comparison to the configuration model indicates that degree conditioning captures the essential randomness for component structure in this attachment process.
- Similar sharp asymptotics might be derivable for other attachment kernels that produce comparable degree sequences.
- Numerical checks for moderate n could verify the accuracy of the logarithmic term in the formula.
Load-bearing premise
After conditioning on the degree sequence the preferential attachment graph behaves sufficiently like the configuration model for the purpose of bounding the largest component.
What would settle it
A computation or simulation in which L1 divided by ε^{-2} log(ε³ n) fails to converge in probability to 2(α+2)/(α+1) for sequences where ε→0 and ε³n→∞ would falsify the stated asymptotic.
read the original abstract
We study the size of the largest component in Pittel's preferential attachment process without vertex growth. Starting from the empty graph on a fixed vertex set $[n]$, edges are added one by one with probabilities proportional to $(d_u+\alpha)(d_v+\alpha)$, where $d_u$ and $d_v$ are the current degrees of $u$ and $v$, and $\alpha>0$. Let $L_1$ denote the size of the largest component, and set $m_c:=\frac{\alpha n}{2(\alpha+1)}.$ We prove that if $m=m_c(1-\varepsilon), \varepsilon=\varepsilon(n)\to0, \varepsilon^3 n\to\infty,$ then \[ L_1=(1+o_p(1))\frac{2(\alpha+2)}{\alpha+1}\varepsilon^{-2}\log(\varepsilon^3 n) \] for every fixed $\alpha>0$. More generally, the same asymptotic holds whenever $\alpha=\alpha(n)\to a\in(0,\infty]$. In particular, the constant $2(\alpha+2)/(\alpha+1)$ converges to the Erd\H{o}s--R\'enyi value $2$ as $\alpha\to\infty$. Moreover, if $m=\left\lfloor \frac n2(1-\varepsilon)\right\rfloor$ and $\alpha\varepsilon\to\infty$, then \[ L_1=(2+o_p(1))\varepsilon^{-2}\log(\varepsilon^3 n). \] The subcritical asymptotics for \(L_1\) resolve the problem left open by Janson and Warnke. The upper bound argument relies on the observation that, after conditioning on the degree sequence, the graph can be treated through the corresponding configuration model, the lower bound follows from tree component asymptotics and a second moment argument.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript establishes sharp asymptotics for the largest component L1 in Pittel's preferential attachment process on n fixed vertices (no vertex growth). For m = m_c(1-ε) with ε=ε(n)→0 and ε³n→∞, it claims L1=(1+o_p(1)) * [2(α+2)/(α+1)] ε^{-2} log(ε³n) for fixed α>0 (and the same form when α(n)→a∈(0,∞]). The upper bound proceeds by conditioning on the degree sequence and reducing to the configuration model; the lower bound uses tree-component asymptotics plus a second-moment argument. The result resolves an open question of Janson and Warnke and recovers the Erdős–Rényi constant 2 in the α→∞ limit.
Significance. If the configuration-model approximation is controlled with error small enough not to affect the leading constant, the result supplies the first explicit sharp asymptotic for this model in the subcritical window, with a clean dependence on α that interpolates to the known ER case. The combination of degree conditioning, configuration-model tail bounds, and second-moment methods on trees is a standard and potentially reproducible approach once the error terms are verified.
major comments (1)
- [Upper bound argument] Upper bound argument (as described after the statement of the main theorem): the assertion that, after conditioning on the degree sequence, “the graph can be treated through the corresponding configuration model” is used to transfer component-size tail bounds that deliver the precise prefactor 2(α+2)/(α+1). No quantitative total-variation or coupling bound is supplied showing that the difference in the probability of long paths (or in the exponential decay rate of component sizes) is o(1) uniformly in the regime ε→0, ε³n→∞. Because the target asymptotic is sharp, any discrepancy of order ε or larger in the decay rate would change the leading constant; an explicit error estimate (e.g., via the evolving attachment probabilities versus the uniform configuration measure) is therefore load-bearing.
minor comments (2)
- [Main theorem] The statement of the result for α(n)→a should clarify whether the o_p(1) term is uniform in a or requires a separate argument when a=∞.
- [Introduction] Notation: the critical value m_c is defined as αn/(2(α+1)); it would help to recall the exact relation between m_c and the emergence of a giant component in the model.
Simulated Author's Rebuttal
We thank the referee for the careful reading and for identifying the need for quantitative error control in the upper-bound argument. We address the comment below.
read point-by-point responses
-
Referee: [Upper bound argument] Upper bound argument (as described after the statement of the main theorem): the assertion that, after conditioning on the degree sequence, “the graph can be treated through the corresponding configuration model” is used to transfer component-size tail bounds that deliver the precise prefactor 2(α+2)/(α+1). No quantitative total-variation or coupling bound is supplied showing that the difference in the probability of long paths (or in the exponential decay rate of component sizes) is o(1) uniformly in the regime ε→0, ε³n→∞. Because the target asymptotic is sharp, any discrepancy of order ε or larger in the decay rate would change the leading constant; an explicit error estimate (e.g., via the evolving attachment probabilities versus the uniform configuration measure) is therefore load-bearing.
Authors: We agree that an explicit quantitative bound is required to justify transferring the sharp tail asymptotics from the configuration model, since any discrepancy of order ε in the decay rate would alter the leading constant. The current manuscript relies on a qualitative statement that the conditioned preferential-attachment graph can be treated via the configuration model but does not supply total-variation or coupling estimates. In the revision we will add a dedicated section deriving such bounds: we will compare the sequential attachment probabilities to the uniform stub-matching measure and show that the difference in the logarithmic asymptotics of the probability of long paths (or of component-size tails) is o(ε) uniformly on the event that the degree sequence lies in the typical set for the subcritical regime. This error is small enough not to affect the prefactor 2(α+2)/(α+1) when ε³n→∞. revision: yes
Circularity Check
No circularity; derivation self-contained from model definition
full rationale
The paper states its main result as a theorem proved from the preferential attachment process definition. The upper bound uses conditioning on the degree sequence followed by treatment as a configuration model; the lower bound uses tree-component asymptotics and the second-moment method. These steps are presented as direct applications of the model without any reduction of the target asymptotic to a fitted parameter, self-referential definition, or load-bearing self-citation chain. The cited prior work (Janson-Warnke) is external and the result is obtained via standard probabilistic arguments on the given process, rendering the derivation self-contained.
Axiom & Free-Parameter Ledger
axioms (1)
- standard math Standard results on configuration models and second-moment calculations for component sizes in random graphs with given degree sequences.
Reference graph
Works this paper leans on
-
[1]
Brownian excursions, critical random graphs and the multiplicative coalescent
David Aldous. Brownian excursions, critical random graphs and the multiplicative coalescent. The Annals of Probability, 25(2):812–854, 1997. doi:10.1214/aop/1024404421
-
[2]
Cambridge University Press, Cambridge, 2016
Albert-L´ aszl´ o Barab´ asi.Network Science. Cambridge University Press, Cambridge, 2016
2016
-
[3]
Emergence of scaling in random networks.Science, 286(5439):509–512, 1999
Albert-L´ aszl´ o Barab´ asi and R´ eka Albert. Emergence of scaling in random networks.Science, 286(5439):509–512, 1999. doi:10.1126/science.286.5439.509
-
[4]
Eli Ben-Naim and P. L. Krapivsky. Popularity-driven networking.Europhysics Letters, 97(4):48003, 2012. doi:10.1209/0295-5075/97/48003
-
[5]
Noam Berger, Christian Borgs, Jennifer T. Chayes, and Amin Saberi. Asymptotic behavior and distributional limits of preferential attachment graphs.The Annals of Probability, 42(1):1– 40, 2014. doi:10.1214/12-AOP755
-
[6]
Shankar Bhamidi, Remco van der Hofstad, and Sanchayan Sen. The multiplicative coalescent, inhomogeneous continuum random trees, and new universality classes for critical random graphs.Probability Theory and Related Fields, 170(1–2):387–474, 2018. doi:10.1007/s00440- 017-0760-6
-
[7]
Cambridge University Press, Cambridge, second edition,
B´ ela Bollob´ as.Random Graphs. Cambridge University Press, Cambridge, second edition,
-
[8]
doi:10.1017/CBO9780511814068
-
[9]
B´ ela Bollob´ as, Svante Janson, and Oliver Riordan. The phase transition in inhomogeneous random graphs.Random Structures & Algorithms, 31(1):3–122, 2007. doi:10.1002/rsa.20168
-
[10]
Random graphs and branching processes
B´ ela Bollob´ as and Oliver Riordan. Random graphs and branching processes. InHandbook of Large-Scale Random Networks, pages 15–115. Springer, Berlin, 2009. doi:10.1007/978-3-540- 69395-6 1
-
[11]
The phase transition in the Erd˝ os–R´ enyi random graph process
B´ ela Bollob´ as and Oliver Riordan. The phase transition in the Erd˝ os–R´ enyi random graph process. InErd˝ os Centennial, pages 59–110. J´ anos Bolyai Mathematical Society, Budapest,
-
[12]
doi:10.1007/978-3-642-39286-3 3
-
[13]
B´ ela Bollob´ as, Oliver Riordan, Joel Spencer, and G´ abor Tusn´ ady. The degree sequence of a scale-free random graph process.Random Structures & Algorithms, 18(3):279–290, 2001. doi:10.1002/rsa.1009. 31
-
[14]
S´ os, and Katalin Vesztergombi
Christian Borgs, Jennifer Chayes, L´ aszl´ o Lov´ asz, Vera T. S´ os, and Katalin Vesztergombi. Limits of randomly grown graph sequences.European Journal of Combinatorics, 32(7):985– 999, 2011. doi:10.1016/j.ejc.2011.03.015
-
[15]
Tom Britton, Maria Deijfen, and Anders Martin-L¨ of. Generating simple random graphs with prescribed degree distribution.Journal of Statistical Physics, 124(6):1377–1397, 2006. doi:10.1007/s10955-006-9168-x
-
[16]
Fan Chung and Linyuan Lu. Connected components in random graphs with given expected degree sequences.Annals of Combinatorics, 6(2):125–145, 2002. doi:10.1007/PL00012580
-
[17]
Matthew Coulson and Guillem Perarnau. Largest component of subcritical random graphs with given degree sequence.Electronic Journal of Probability, 28:34, 1–28, 2023. doi:10.1214/23-EJP921
-
[18]
Souvik Dhara, Remco van der Hofstad, Johan S. H. van Leeuwaarden, and Sanchayan Sen. Critical window for the configuration model: finite third moment degrees.Electronic Journal of Probability, 22:1–33, 2017. doi:10.1214/17-EJP29
-
[19]
Jian Ding, Jeong Han Kim, Eyal Lubetzky, and Yuval Peres. Anatomy of a young giant component in the random graph.Random Structures & Algorithms, 39(2):139–178, 2011. doi:10.1002/rsa.20342
-
[20]
Jian Ding, Eyal Lubetzky, and Yuval Peres. Anatomy of the giant component: The strictly supercritical regime.European Journal of Combinatorics, 35:155–168, 2014. doi:10.1016/j.ejc.2013.06.004
-
[21]
S. N. Dorogovtsev and J. F. F. Mendes.Evolution of Networks. Oxford University Press, Oxford, 2003. doi:10.1093/acprof:oso/9780198515906.001.0001
work page doi:10.1093/acprof:oso/9780198515906.001.0001 2003
-
[22]
S. N. Dorogovtsev, J. F. F. Mendes, and A. N. Samukhin. Principles of statistical mechanics of uncorrelated random networks.Nuclear Physics B, 666(3):396–416, 2003. doi:10.1016/S0550- 3213(03)00504-2
-
[23]
Cambridge University Press, Cambridge, 2010
Rick Durrett.Random Graph Dynamics. Cambridge University Press, Cambridge, 2010
2010
-
[24]
On random graphs
Paul Erd˝ os and Alfr´ ed R´ enyi. On random graphs. I.Publicationes Mathematicae Debrecen, 6:290–297, 1959
1959
-
[25]
On the evolution of random graphs.Publications of the Math- ematical Institute of the Hungarian Academy of Sciences, 5:17–61, 1960
Paul Erd˝ os and Alfr´ ed R´ enyi. On the evolution of random graphs.Publications of the Math- ematical Institute of the Hungarian Academy of Sciences, 5:17–61, 1960
1960
-
[26]
Percolation on sparse random graphs with given degree sequence
Nikolaos Fountoulakis. Percolation on sparse random graphs with given degree sequence. Internet Mathematics, 4(4):329–356, 2007. doi:10.1080/15427951.2007.10129148
-
[27]
Cambridge University Press, Cambridge, 2016
Alan Frieze and Micha l Karo´ nski.Introduction to Random Graphs. Cambridge University Press, Cambridge, 2016. doi:10.1017/CBO9781316339831. 32
-
[28]
Hamed Hatami and Michael Molloy. The scaling window for a random graph with a given de- gree sequence.Random Structures & Algorithms, 41(1):99–123, 2012. doi:10.1002/RSA.20394
-
[29]
Nongrowing preferential attachment random graphs.Internet Mathematics, 6(4):461–487, 2010
Tom´ aˇ s Hruˇ z and Uwe Peter. Nongrowing preferential attachment random graphs.Internet Mathematics, 6(4):461–487, 2010. doi:10.1080/15427951.2010.553143
-
[30]
Slightly subcritical hypercube percolation.Random Struc- tures & Algorithms, 56(2):557–593, 2020
Tim Hulshof and Asaf Nachmias. Slightly subcritical hypercube percolation.Random Struc- tures & Algorithms, 56(2):557–593, 2020. doi:10.1002/rsa.20853
-
[31]
Svante Janson. On percolation in random graphs with given vertex degrees.Electronic Journal of Probability, 14:87–118, 2009. doi:10.1214/EJP.v14-603
-
[32]
Svante Janson. Susceptibility of random graphs with given vertex degrees.Journal of Com- binatorics, 1(4):357–387, 2010. doi:10.4310/JOC.2010.v1.n4.a2
-
[33]
On edge exchangeable random graphs.Journal of Statistical Physics, 173(3– 4):448–484, 2018
Svante Janson. On edge exchangeable random graphs.Journal of Statistical Physics, 173(3– 4):448–484, 2018. doi:10.1007/s10955-017-1832-9
-
[34]
Svante Janson and Malwina J. Luczak. A simple solution to thek-core problem.Random Structures & Algorithms, 30(1–2):50–62, 2007. doi:10.1002/rsa.20147
-
[35]
Svante Janson and Malwina J. Luczak. A new approach to the giant component problem. Random Structures & Algorithms, 34(2):197–216, 2009. doi:10.1002/rsa.20231
-
[36]
Wiley, New York,
Svante Janson, Tomasz Luczak, and Andrzej Ruci´ nski.Random Graphs. Wiley, New York,
-
[37]
doi:10.1002/9781118032718
-
[38]
Svante Janson and Lutz Warnke. Preferential attachment without vertex growth: emer- gence of the giant component.The Annals of Applied Probability, 31(4):1523–1547, 2021. doi:10.1214/20-AAP1610
-
[39]
The component sizes of a critical random graph with given degree sequence
Adrien Joseph. The component sizes of a critical random graph with given degree sequence. The Annals of Applied Probability, 24(6):2560–2594, 2014. doi:10.1214/13-AAP985
-
[40]
Mihyun Kang and Tom G. Seierstad. The critical phase for random graphs with a given degree sequence.Combinatorics, Probability and Computing, 17(1):67–86, 2008. doi:10.1017/S096354830700867X
-
[41]
Michael Molloy and Bruce Reed. A critical point for random graphs with a given degree se- quence.Random Structures & Algorithms, 6(2–3):161–180, 1995. doi:10.1002/rsa.3240060204
-
[42]
Michael Molloy and Bruce Reed. The size of the giant component of a random graph with a given degree sequence.Combinatorics, Probability and Computing, 7(3):295–305, 1998. doi:10.1017/S0963548398003526
-
[43]
Critical percolation on random regular graphs.Random Structures & Algorithms, 36(2):111–148, 2010
Asaf Nachmias and Yuval Peres. Critical percolation on random regular graphs.Random Structures & Algorithms, 36(2):111–148, 2010. doi:10.1002/rsa.20277. 33
-
[44]
Advances in Applied Probability , author =
Ilkka Norros and Hannu Reittu. On a conditionally Poissonian graph process.Advances in Applied Probability, 38(1):59–75, 2006. doi:10.1239/aap/1143936140
-
[45]
On a random graph evolving by degrees.Advances in Mathematics, 223(2):619– 671, 2010
Boris Pittel. On a random graph evolving by degrees.Advances in Mathematics, 223(2):619– 671, 2010. doi:10.1016/j.aim.2009.08.015
-
[46]
Boris Pittel, Joel Spencer, and Nicholas Wormald. Sudden emergence of a giantk-core in a random graph.Journal of Combinatorial Theory, Series B, 67(1):111–151, 1996. doi:10.1006/jctb.1996.0036
-
[47]
Derek J. de Solla Price. A general theory of bibliometric and other cumulative advantage processes.Journal of the American Society for Information Science, 27(5):292–306, 1976. doi:10.1002/ASI.4630270505
-
[48]
Bal´ azs R´ ath. Time evolution of dense multigraph limits under edge-conservative pref- erential attachment dynamics.Random Structures & Algorithms, 41(3):365–390, 2012. doi:10.1002/rsa.20422
-
[49]
Bal´ azs R´ ath and L´ aszl´ o Szak´ acs. Multigraph limit of the dense configuration model and the preferential attachment graph.Acta Mathematica Hungarica, 136(3):196–221, 2012. doi:10.1007/s10474-012-0217-4
-
[50]
Oliver Riordan. The phase transition in the configuration model.Combinatorics, Probability and Computing, 21(1–2):265–299, 2012. doi:10.1017/S0963548311000666
-
[51]
Birth control for giants.Combinatorica, 27(5):587–628,
Joel Spencer and Nicholas Wormald. Birth control for giants.Combinatorica, 27(5):587–628,
-
[52]
doi:10.1007/s00493-007-2163-2
-
[53]
Remco van der Hofstad.Random Graphs and Complex Networks. Volume 1. Cambridge University Press, Cambridge, 2017. doi:10.1017/9781316779422
-
[54]
Remco van der Hofstad.Random Graphs and Complex Networks. Volume 2. Cambridge University Press, Cambridge, 2024. doi:10.1017/9781316795552
-
[55]
Remco van der Hofstad, Svante Janson, and Malwina J. Luczak. Component structure of the configuration model: barely supercritical case.Random Structures & Algorithms, 55(1):3–55,
-
[56]
doi:10.1002/rsa.20837. 34
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.