Pointwise multiple ergodic averages for mixing transformations

1 · The essentials

Imagine repeatedly applying a rule that moves points around a space while leaving their overall probability distribution unchanged. The manuscripts study reversible rules for which events become increasingly unrelated when separated by many steps. At each round, one takes bounded readings after one, two, three, or more multiples of the round number, multiplies those readings, and averages the products over rounds. The main manuscript claims that, for any fixed finite number of readings at least two, this average approaches the product of the individual space-wide averages. This happens for almost every starting point, meaning that the exceptions together have probability zero. The claim covers every fixed choice of readings and every positive integer averaging length, without requiring a specified speed at which events become unrelated. The main result is not formalized in Lean and has not been formally verified.

A picture to help (an imperfect image, not to be taken literally)

As an imperfect image, imagine a reversible shuffle and several bounded scores assigned to its configurations: at round $k$ , you record scores after $k$ , $2 k$ , and $3 k$ shuffles and multiply them. The claimed conclusion says that averaging these products over rounds approaches the product of the separate average scores, provided the shuffle satisfies the manuscripts’ precise assumptions; an ordinary shuffle need not satisfy them.

Why it matters

The claim concerns what happens along the evolution of almost every individual starting point, not merely a space-wide average. Its scope includes every finite consecutive length at least two, without a prescribed mixing rate or a standardness assumption on the probability space. The companion manuscripts also address fourfold and triple averages, including triples with distinct integer slopes.

2 · Going further

Let $T$ be an invertible mixing probability-preserving transformation on a probability space $(X, F, μ)$ . For an integer $r \geq 2$ and a fixed tuple of bounded measurable functions $f_{1}, \dots, f_{r}$ , the main manuscript claims that $A_{N} (x) = \frac{1}{N} \sum_{k = 1}^{N} \prod_{j = 1}^{r} f_{j} (T^{j k} x)$ converges to $\prod_{j = 1}^{r} \int_{X} f_{j} d μ$ for $μ$ -almost every $x$ . The limit is taken as $N$ runs through all positive integers, rather than through a selected subsequence.

The stated hypotheses require mixing but impose no quantitative rate of mixing. The probability space need not be standard. The quantifiers concern each fixed tuple of bounded measurable inputs; the abstracts do not state that a single exceptional set works simultaneously for every tuple. The fourfold and consecutive-triple abstracts explicitly require a measurable inverse, with the fourfold abstract describing the transformation as bimeasurable.

The companion manuscripts describe overlapping and related claims. The fourfold manuscript treats the times $k, 2 k, 3 k, 4 k$ , and the consecutive-triple manuscript gives the formula for $k, 2 k, 3 k$ , explicitly allowing complex-valued inputs. The distinct-slopes manuscript claims the corresponding triple conclusion for times $a_{1} k, a_{2} k, a_{3} k$ , where the integers $a_{1}, a_{2}, a_{3}$ are nonzero and pairwise distinct; positive and negative slopes are allowed.

The supplied abstracts state conclusions and assumptions but do not describe an argument or proof mechanism, so no main argumentative idea can be reported from this source. They also provide no historical account or named open problem. The main result has not been formalized in Lean or formally verified.

Useful notions

Probability spaces and integration
Measurable functions and boundedness
Iterating an invertible transformation
Probability preservation and mixing
Almost-everywhere convergence

What was known before / what is new

Before

The source supplies no account of earlier established results or of the historical status of this problem. It lists companion manuscripts on fourfold averages and two kinds of triple averages, but does not give a chronology or say that these results were previously established.

New

The main manuscript claims almost-everywhere convergence for consecutive multiple averages of every finite length at least two, with the limit equal to the product of the input integrals. The claim requires neither a quantitative mixing rate nor a standard probability space and covers all positive integer averaging lengths. A companion manuscript also claims a triple-average result for any three pairwise distinct nonzero integer slopes, including negative slopes.

Limits and assumptions

The main claim assumes an invertible, mixing, probability-preserving transformation and a fixed finite tuple of bounded measurable functions. The fourfold and consecutive-triple abstracts explicitly include measurability of the inverse. Convergence is almost everywhere, not at every starting point, and the source does not assert a common full-measure set for all input tuples. Arbitrary distinct nonzero integer slopes are covered by the supplied claim for triples, not for every finite length. The abstracts supply no argument details, and the main result is not formalized in Lean and has not been formally verified. No remaining open problem is identified in the source; cases beyond the stated assumptions are not addressed.

Mini glossary

Probability-preserving transformation: A measurable map $T$ that satisfies $μ (T^{- 1} A) = μ (A)$ for every measurable set $A$ .
Mixing: The property that $μ (A \cap T^{- k} B)$ tends to $μ (A) μ (B)$ as $k$ tends to infinity for measurable sets $A$ and $B$ .
Multiple ergodic average: An average of products of function values evaluated at several iterates of the same transformation.
Almost-everywhere convergence: Convergence at every point outside an exceptional set of probability zero.
Integer slope: An integer multiplier $a$ specifying observation times $ak$ in an average indexed by $k$ .
Bimeasurable: For an invertible transformation, the property that both the transformation and its inverse are measurable.

Keywords : Mixing transformations · Multiple ergodic averages · Almost-everywhere convergence

AI-generated explanation based on the manuscript. Unofficial. Refer to the manuscript for the authoritative version.

3 · Technical details and sources

Claim made in the manuscript (source text)

Proves almost-everywhere convergence of consecutive multiple ergodic averages of every finite length for invertible mixing probability-preserving transformations. For each fixed tuple of bounded functions, the limit is the product of their integrals, along all positive averaging lengths. No mixing rate or standardness assumption on the probability space is required.

To be verified: the proof is written in the manuscript, but no software has checked it and no validation by independent mathematicians is established. It may contain errors.

Manuscripts in this family

★

Pointwise Multiple Ergodic Averages for Mixing Transformations

04/10/2026

To be verified

Let T be an invertible mixing probability-preserving transformation. For every integer n ≥ 2 and every fixed tuple of bounded measurable functions, we prove that the consecutive multiple ergodic averages of length n converge almost everywhere to the product of the integrals, as the averaging length tends to infinity through all positive integers. The probability space need not be standard, and no rate of mixing is required.

PDF GitHub folder

Pointwise convergence of fourfold ergodic averages for mixing transformations

04/10/2026

To be verified

We prove that fourfold ergodic averages along the times $n, 2 n, 3 n, 4 n$ converge almost everywhere to the product of the integrals for every invertible, bimeasurable, mixing probability-preserving transformation and every fixed choice of bounded measurable inputs. Convergence holds along all positive integer averaging lengths. No quantitative mixing rate is assumed, and the probability space need not be standard.

PDF GitHub folder

Triple ergodic averages with distinct integer slopes

04/10/2026

To be verified

For every invertible mixing probability-preserving transformation, triple ergodic averages of bounded measurable functions along any three pairwise distinct nonzero integer slopes converge almost everywhere to the product of their integrals. The slopes may be positive or negative.

PDF GitHub folder

Pointwise convergence of triple ergodic averages for mixing transformations

04/10/2026

To be verified

We prove pointwise convergence of triple ergodic averages for every invertible mixing probability preserving transformation T of an arbitrary probability space $(X, F, μ)$ , with measurable inverse. For every triple of bounded measurable functions $f_{1}, f_{2}, f_{3} : X \to C$ , $\frac{1}{N} n = 1 \sum N f_{1} (T^{n} x) f_{2} (T^{2 n} x) f_{3} (T^{3 n} x) ⟶ j = 1 \prod 3 \int_{X} f_{j} d μ$ for μ-almost every x as $N \to \infty$ through all positive integers.

PDF GitHub folder