---
title: Polyakov–KNO Multiplicity Distributions in QCD
url: https://www.emergentmind.com/topics/polyakov-kno-multiplicity-distributions
type: topic
---

# Polyakov–KNO Multiplicity Distributions in QCD

Polyakov–KNO multiplicity distributions are continuous scaling distributions \(\Psi(\nu,Q)\) associated with the discrete multiplicity probabilities \(P_n(Q)=\sigma_n(Q)/\sigma_{\rm tot}(Q)\), defined through \(N(Q)P_n(Q)\simeq \Psi(\nu,Q)\) with \(N(Q)\equiv \langle n\rangle(Q)\) and \(\nu=n/N(Q)\). In the strict KNO limit, \(\Psi\) becomes \(Q\)-independent; in QCD jets one instead expects approximate P-KNO scaling with slow violations controlled by the running coupling. A universal QCD-motivated expression has been proposed for hadrons in jets, and its moments and overall shape in full events and quark and gluon jets are described with reasonable quantitative precision over a range of energies; in particular, the scaling violation predicted by QCD is seen clearly in the moments and high-multiplicity fluctuations [2507.07691].

## 1. Scaling definition and kinematic variables

The hardness scale \(Q\) is the organizing variable for the jet multiplicity problem. For \(e^+e^-\) at centre-of-mass \(\sqrt{s}\), one has \(Q=\sqrt{s}\) for full events or \(Q=\sqrt{s/2}\) per hemisphere; more generally, for a jet of energy \(E\) and opening angle \(\Theta\), \(Q=2E\sin(\Theta/2)\). The Polyakov–KNO scaling variable is
\[
\nu=\frac{n}{N(Q)}, \qquad N(Q)\equiv \langle n\rangle(Q),
\]
and the continuous scaling distribution is defined by
\[
P_n(Q)\equiv \sigma_n(Q)/\sigma_{\rm tot}(Q), \qquad N(Q)P_n(Q)\simeq \Psi(\nu,Q).
\]

The P-KNO hypothesis, identified in the jet analysis as Polyakov 1971, states that the full energy dependence of \(P_n\) is absorbed into \(N(Q)\), so that \(N(Q)P_n\to \Psi(\nu,Q)\). In this formulation, strict scaling would require \(\Psi(\nu,Q)\to \Psi(\nu)\), whereas finite-energy QCD predicts scaling violation through the residual \(Q\)-dependence of the scaling function.

The physical picture is a cascading QCD parton shower. Hadron multiplicities in jets arise from this cascade, whose universal scaling properties were first anticipated by Polyakov in a conformal field theory context. This places Polyakov–KNO distributions at the intersection of multiparticle phenomenology, parton-cascade dynamics, and asymptotic scaling theory [2507.07691].

## 2. Universal jet distribution in the Dokshitzer–Webber framework

For a gluon-initiated jet, the proposed universal form is
\[
\nu\,\Psi_g(\nu;Q)
\simeq
\frac{2\,\mu(\gamma)^2\,[\kappa(\nu,\gamma)]^2\,\exp[-\kappa(\nu,\gamma)]}
{\Bigl[1+\frac{1}{p_1\,\kappa(\nu,\gamma)}\Bigr]^{p_1}\,
\chi\!\bigl(\kappa(\nu,\gamma),\gamma\bigr)},
\qquad p_0=1,\quad p_1=4.
\]

Its ingredients are specified by the multiplicity anomalous dimension \(\gamma\equiv \gamma(\bar\alpha_s(Q))\), the Polyakov exponent \(\mu(\gamma)=1/(1-\gamma)\), the tail-normalisation factor
\[
D(\gamma)=C\,\gamma^\gamma(1-\gamma)^{1-\gamma}/\Gamma(1+\gamma), \qquad C\approx 2.552,
\]
the high-\(\nu\) variable
\[
\kappa(\nu,\gamma)=[D(\gamma)\nu]^{\mu(\gamma)},
\]
and the slowly-varying prefactor
\[
\chi(k,\gamma)=
\frac{\Gamma(1+k)\,[\gamma^\gamma(1-\gamma)^{1-\gamma}]^k}
{\Gamma(1+\gamma k)\,\Gamma(1+(1-\gamma)k)}.
\]
The anomalous dimension is given to two loops by
\[
\gamma(Q)\approx \sqrt{\frac{2N_c\,\bar\alpha_s(Q)}{\pi}
-\Bigl(\frac{\beta_0}{4}+\frac{10n_f}{3N_c^2}\Bigr)\frac{\bar\alpha_s^2(Q)}{2\pi}},
\]
with \(\bar\alpha_s=\alpha_s/\pi\) in the CMW (physical) scheme.

For a more general source \(S\), such as a quark jet or a full \(e^+e^-\) event, the distribution is written in terms of the hadron-production “power” \(\rho_S=\langle n_S\rangle/\langle n_g\rangle\):
\[
\nu\,\Psi_S(\nu;Q)
\simeq
\bigl[\nu\,\Psi_g(\nu;Q)\bigr]^{\rho_S}\,\sqrt{\rho_S}\,
G(\kappa,\gamma)^{\,\rho_S-1},
\]
where \(G\) is a slowly varying function of \(\kappa\) and \(\gamma\). Apart from two ad hoc constants \((p_0,p_1)\), the entire shape and its energy dependence are determined by one function \(\gamma(Q)\), itself fixed by the two-loop running coupling; quark/gluon differences enter only via a multiplicative \(\rho\) factor [2507.07691].

## 3. Resummed derivation from moments to the full distribution

The construction begins with the P-KNO hypothesis and then uses generating functions and Laplace transforms. In the leading double-logarithmic approximation (DLA), one obtains an asymptotic form for the Laplace transform \(\Phi(\beta)=\int d\nu\,\Psi(\nu)e^{-\beta\nu}\) and therefore for the factorial and power moments \(g_k\equiv \langle n^k\rangle/N^k\). For \(k\gg1\), the gluon-jet moments behave as
\[
g_k^g\approx (2\,k!/C^k)\Bigl(k+\frac{1}{3k}\Bigr).
\]

The modified DLA (MDLA), identified with Dokshitzer ’93 in the summary, resums all terms of order \((k\gamma)^m\) in the perturbative series and restores exact energy-momentum balance in the cascade. This yields
\[
g_k^g(Q)=g_k^{\rm DLA}\times \frac{[\Gamma(1+\gamma)]^k}{\Gamma(1+\gamma k)}.
\]
Laplace inversion by steepest descent reconstructs \(\Psi_g(\nu)\) from \(\Phi(\beta)\), matching the large-\(k\) behaviour of the moments to the moments of the scaling function. The resulting high-\(\nu\) tail is
\[
\nu\,\Psi_g(\nu)\sim 2\,\mu^2\,(\kappa^2-\kappa)\,\chi(e^{-\kappa}).
\]

A two-parameter completion then introduces subleading shifts \(p_0,p_1\) so as to restore positivity for \(\nu<1\) and exactly satisfy \(g_0=1,g_1=1\). The minimal choice \(p_0=1,p_1=4\) fixes the full-range ansatz. Extension to quark jets and full \(e^+e^-\) events follows from the DLA relation
\[
\Phi_S(\beta)=\bigl[\Phi_g(\beta/\rho_S)\bigr]^{\rho_S},
\]
which is then re-inverted to obtain \(\Psi_S\). The overall procedure moves from moments to a full scaling distribution while preserving the perturbatively controlled high-multiplicity structure [2507.07691].

## 4. Moments, cumulants, and QCD scaling violation

The mean multiplicity is determined by
\[
\frac{d\ln N}{d\ln Q}=\gamma\bigl(\bar\alpha_s(Q)\bigr),
\]
and the power moments are defined by
\[
g_k(Q)\equiv \frac{\langle n^k\rangle}{N(Q)^k}
=\left.\Bigl(-\frac{d}{d\beta}\Bigr)^k\Phi(\beta,Q)\right|_{\beta=0}.
\]
In MDLA, the gluon moments retain the form
\[
g_k^g(Q)=g_k^{\rm DLA}\,
\frac{[\Gamma(1+\gamma)]^k}{\Gamma(1+\gamma k)}.
\]
For a quark jet \((\rho_q=2/3)\) or a full \(e^+e^-\) event \((\rho_{ee}=4/3)\), the corresponding moments are obtained by repeated differentiation of \([\Phi_g(\beta/\rho)]^\rho\).

Cumulants follow in the usual way from the \(g_k\). In particular, the dispersion \(D=\sqrt{\langle n^2\rangle-\langle n\rangle^2}\) satisfies
\[
g_2(Q)=\frac{\langle n^2\rangle}{N^2}, \qquad
\frac{N}{D}=\frac{1}{\sqrt{g_2-1}}.
\]
At low \(Q\), DLA fails because \(g_2\to1\) and hence \(N/D\to\infty\), whereas MDLA cures this for \(Q\gtrsim30\,\mathrm{GeV}\) and predicts a strong KNO-violation for \(Q\lesssim40\,\mathrm{GeV}\) as \(\gamma\to0.56\). The energy dependence of the moments is driven entirely by \(\gamma(Q)\): as \(Q\) increases, \(\gamma\) decreases slowly, the MDLA suppression factor tends to \(1\), and the asymptotic DLA values are recovered.

Typical moments extracted from the summary are:

| \(k\) | \(91\,\mathrm{GeV}\): \(g_k^{\rm exp}/g_k^{\rm MDLA}\) | \(29\,\mathrm{GeV}\): \(g_k^{\rm exp}/g_k^{\rm MDLA}\) |
|---|---|---|
| 2 | \(\approx 1.31 / 1.29\) | \(\approx 1.05 / 1.08\) |
| 3 | \(\approx 1.88 / 1.85\) | \(\approx 1.15 / 1.17\) |
| 4 | \(\approx 3.0 / 2.9\) | \(\approx 1.3 / 1.33\) |

These values were read off Fig. 1 with uncertainties \(\lesssim 5\%\). A common misconception is that KNO scaling in QCD should be exact at accessible energies. In the jet calculation, the asymptotic regime is recovered only slowly because \(\gamma(Q)\to0\) only logarithmically with \(Q\); at finite \(Q\), the high-\(\nu\) region exhibits noticeable scaling violation, including the narrowing of the distribution as \(Q\) increases [2507.07691].

## 5. Comparison with \(e^+e^-\) events and identified jet samples

The distributions are compared in the form \(\Psi(\nu,Q)=N\,P_n\) versus \(\nu=n/N\), in both linear and logarithmic scale. For full \(e^+e^-\) events at LEP-1, with \(Q=M_Z=91.2\,\mathrm{GeV}\), ALEPH, OPAL, DELPHI, and L3 all agree, and MDLA with \(\rho=4/3\) reproduces the shape and high-\(\nu\) tail to within experimental errors. OPAL data at LEP-2, \(Q=161\) and \(189\,\mathrm{GeV}\), were also compared with the MDLA form. At lower energies, including TASSO at \(44\,\mathrm{GeV}\), HRS at \(29\,\mathrm{GeV}\), ARGUS at \(10\,\mathrm{GeV}\), and TASSO at \(14\) and \(22\,\mathrm{GeV}\), the distributions are broader than MDLA for \(Q\lesssim40\,\mathrm{GeV}\), consistent with the proximity to \(\gamma\to0.56\).

For quark jets, DELPHI at \(91\,\mathrm{GeV}\) and HRS at \(29\,\mathrm{GeV}\) show that the MDLA curves for \(Q=\sqrt{s}\) and, optionally, \(Q=\sqrt{s/2)\), bracket the data within uncertainties. For gluon jets, OPAL unbiased gluon jets in three-jet events, with \(E_g\sim5\)–\(18\,\mathrm{GeV}\) and mean \(E_g^\ast\sim14\,\mathrm{GeV}\), corresponding to \(Q=28\,\mathrm{GeV}\), prefer the \(\rho=1\) gluon curve over the \(\rho=2/3\) quark curve, especially in the tail \(\nu>1\). OPAL gluon jets recoiling against \(q\bar q\), with \(E_g\sim40\,\mathrm{GeV}\) and restricted rapidity \(y<2\), show reasonable agreement at mid-rapidity.

The high-multiplicity tail is central to the phenomenology. In pure KNO scaling one would have \(\Psi(\nu;Q)\to\Psi(\nu)\) as \(Q\to\infty\). In QCD, however, the tail falls faster than exponentially:
\[
\nu\,\Psi(\nu)\sim \exp[-\kappa], \qquad
\kappa=(D\nu)^\mu,\qquad
\mu=\frac{1}{1-\gamma}>1.
\]
As \(Q\to\infty\), \(\gamma\to0\), so \(\mu\to1\) and \(D(\gamma)\to1\), giving \(\kappa\to\nu\) and the pure exponential tail \(\Psi(\nu)\sim \exp(-\nu)/\nu\). At finite \(Q\), the super-exponential suppression is therefore stronger, and the narrowing of the distribution in the high-\(\nu\) region becomes a direct manifestation of scaling violation [2507.07691].

## 6. Other QCD realizations, limitations, and interpretation

Polyakov–KNO-type behaviour also appears in other high-energy QCD limits. In the BFKL approach, the multiplicity generating function for real gluon emissions in a cut ladder takes the form
\[
P(u)=\exp\{\omega(0)Y(u-1)\},
\]
with mean multiplicity
\[
\langle n\rangle=\omega(0)Y\equiv \omega_p Y,
\qquad
\omega_p=4\,\alpha_s N_c\ln 2/\pi
\]
in LLA, and therefore
\[
P_n=e^{-\langle n\rangle}\frac{\langle n\rangle^n}{n!}.
\]
This is exactly a Poisson distribution. Defining \(z=n/\langle n\rangle\) and \(\psi(z)=\langle n\rangle P_n\), Stirling’s formula gives an asymptotic form in which, up to a slowly-varying prefactor \(1/\sqrt{\langle n\rangle}\), the exponent depends only on \(z\). The analysis stresses that this Poisson law describes the number of BFKL “cells” in the cut diagram rather than fully dressed physical gluons, and that realistic cut-offs or restricted rapidity windows produce modest corrections while leaving the leading Poisson behaviour intact [2003.03275].

A distinct realization arises in deep inelastic scattering in the CGC/saturation framework at large \(z=\ln(Q_s^2/Q^2)\gg1\). In that regime one finds \(\bar n(z)\propto \exp(z^2/(2\kappa))\) with \(\kappa=4.88\) at LO BFKL, and for \(n>\bar n\) the multiplicity distribution almost reproduces KNO scaling with
\[
\Psi\!\left(\frac{n}{\bar n}\right)=\exp\!\left(-\frac{n}{\bar n}\right).
\]
For \(n<\bar n\), however, the distribution is instead
\[
\sigma_n \propto \frac{z-\sqrt{2\kappa\ln(n-1)}}{n-1},
\]
and these small-\(n\) terms determine the entropy
\[
S_E=0.3\,\frac{z^2}{2\kappa}
\]
at large \(z\), where the factor \(0.3\) stems from non-perturbative corrections associated with the large-\(b\) behaviour of the saturation momentum. The derivation relies on LO pQCD with fixed coupling, geometric scaling, AGK cutting rules for CGC Pomeron fan diagrams, and a smooth nonperturbative decay of \(Q_s(b)\); running-coupling effects, NLO corrections, and the model for \(Q_s(b)\) affect both the small-\(n\) region and the entropy coefficient [2306.12055].

Taken together, these results indicate that Polyakov–KNO scaling in QCD is not a single universal law with identical microscopic content in all regimes. In jets, the salient feature is a resummed, largely parameter-free analytic form whose deviations from strict scaling are governed by \(\gamma(Q)\). In BFKL ladders, the leading multiplicity law is Poissonian. In deep saturation, the large-\(n\) tail is exponential in \(n/\bar n\), while the small-\(n\) sector obeys a different law. This suggests that the Polyakov–KNO concept is best understood as a scaling framework whose concrete realization is determined by the underlying QCD dynamics, the relevant hardness variable, and the treatment of energy conservation, rapidity restrictions, and impact-parameter structure.

Source: https://www.emergentmind.com/topics/polyakov-kno-multiplicity-distributions