Shen, Uspensky & Vereshchagin — Chapter 9

Frequency & Game
Approaches

Von Mises, Ville, and martingales: alternative roads to randomness through selection rules and betting strategies.

Prerequisites: Chapter 3 (Martin-Löf randomness). Basic probability (optional: Chapter 7 on entropy).
10
Chapters
2
Simulations
10
Quizzes

Chapter 0: Von Mises' Idea

Before Martin-Löf, before Kolmogorov complexity, the first serious attempt to define random sequences was by Richard von Mises (1919). His idea was intuitive and beautiful, though ultimately flawed.

Von Mises' definition (informal): An infinite binary sequence is random if:
(1) The frequency of ones converges to 1/2, and
(2) The same is true for every subsequence selected by an "admissible" selection rule.

A selection rule is a procedure that looks at a sequence and decides which positions to select. For example: "take every third bit," or "take bits after a run of two zeros." Von Mises called a sequence satisfying both conditions a Kollektiv.

The idea is: a random sequence should remain random even after you cherry-pick a subsequence. You shouldn't be able to find a pattern that, when exploited, reveals a bias.

Gambling interpretation: A selection rule is a betting strategy that decides when to bet, but always bets the same way. If the sequence is random, no strategy for choosing when to bet should find a bias.
Check: Von Mises' Kollektiv requires that frequency remains 1/2 even when...

Chapter 1: Selection Rules

A selection rule is a function σ: {0,1}* → {0, 1} that, given the bits seen so far, decides whether to "select" the next bit (1 = select, 0 = skip). The selected subsequence consists of the bits at positions where σ returns 1.

Crucially, the decision to select position n depends only on the bits before position n, not on the bit at position n itself. The gambler decides to bet before seeing the card.

Examples:
• σ(x) = 1 for all x: select every bit (trivial rule).
• σ(x) = 1 iff |x| is even: select every other bit.
• σ(x) = 1 iff x ends in 00: select bits after two zeros.

Von Mises never precisely specified which selection rules are "admissible." This vagueness was the source of much debate. Different choices lead to different notions of randomness.

Check: A selection rule σ must decide whether to select position n based on...

Chapter 2: Mises-Church Randomness

Alonzo Church (1940) proposed a natural fix for Von Mises' vagueness: take the admissible selection rules to be the computable ones.

Mises-Church randomness: A sequence ω is Mises-Church random if for every computable selection rule σ that selects infinitely many bits, the selected subsequence has frequency of ones equal to 1/2.

This is a clean, precise definition. It says: no algorithm can select a biased subsequence from ω. The sequence defeats all computable "frequency-finding" strategies.

However, Mises-Church randomness turns out to be too weak. It is strictly weaker than Martin-Löf randomness. Ville showed why.

Check: Mises-Church randomness restricts selection rules to be...

Chapter 3: Ville's Counterexample

In 1939, Jean Ville proved a devastating result:

Ville's theorem: There exists a sequence ω that is Mises-Church random (frequency 1/2 is preserved under all computable selection rules) but has the property that the running frequency of ones is always ≥ 1/2. In other words, ω is "biased above" even though no computable test detects the bias through subsequence selection.

This sequence passes all frequency tests but fails a different kind of test: the running average never dips below 1/2. A truly random sequence should cross 1/2 infinitely often (by the Law of the Iterated Logarithm).

The lesson: Selection rules only test for frequency bias in subsequences. They miss other kinds of non-randomness, like biased running averages. To capture all effective tests, we need something stronger than selection rules — we need martingales.
Check: Ville's counterexample shows that Mises-Church randomness is...

Chapter 4: Martingales

A martingale is a betting strategy. Formally, a martingale is a function d: {0,1}* → [0, ∞) satisfying the fairness condition:

d(x) = (d(x0) + d(x1)) / 2 for all x

Think of d(x) as the gambler's capital after seeing the sequence x. The fairness condition says: no matter what the next bit is, the expected capital doesn't change. The gambler starts with d(λ) dollars and bets on each bit.

Gambling interpretation: Before seeing bit n, the gambler decides how much to bet on "the next bit is 1." If right, capital increases; if wrong, it decreases. The fairness condition ensures this is a "fair" game — against a truly random sequence, the expected capital stays constant.

A martingale succeeds on a sequence ω if d(ω[0..n)) → ∞ as n → ∞. The gambler makes unbounded profit — the sequence has a pattern the gambler can exploit.

A sequence is random (in the martingale sense) if no martingale from a given class succeeds on it. Different classes give different randomness notions.

Check: A martingale succeeds on a sequence if...

Chapter 5: Lower Semicomputable Martingales

The key result connecting martingales to Martin-Löf randomness:

Schnorr's Theorem: A sequence ω is Martin-Löf random iff no lower semicomputable martingale succeeds on ω.

A martingale d is lower semicomputable if d(x) can be approximated from below by a computable sequence. Equivalently, the gambler can compute increasingly accurate lower bounds on the optimal bet at each step.

This gives a beautiful game-theoretic characterization of ML-randomness: a sequence is random iff no effective gambler can make unlimited money betting on it.

Universal martingale: There exists a maximal lower semicomputable martingale M (obtained by mixing all l.s.c. martingales with weights 2−i). M succeeds on ω iff ω is not ML-random. The growth rate of M on ω's prefixes is the randomness deficiency.
Check: Schnorr's theorem characterizes ML-randomness using...

Chapter 6: Schnorr Randomness

Schnorr (1971) proposed a weaker randomness notion using computable martingales (rather than lower semicomputable ones).

Schnorr randomness: ω is Schnorr random if no computable martingale succeeds on ω. Additionally, the success must be "effective": lim sup d(ω[0..n)) = ∞ must happen at a computably recognizable rate.

Schnorr randomness is strictly weaker than ML-randomness. There exist sequences that are Schnorr random but not ML-random. The difference: Schnorr randomness only defends against "constructive" gambling strategies, while ML-randomness also defends against semi-constructive ones.

NotionMartingale classStrength
Mises-ChurchComputable selection rulesWeakest
SchnorrComputable martingalesMiddle
Martin-LöfLower semicomputable martingalesStrongest
Check: Schnorr randomness uses which class of martingales?

Chapter 7: Martingales and Dimension

Martingales also characterize effective Hausdorff dimension. Instead of asking "does the martingale grow to infinity?", we ask "how fast does it grow?"

Theorem: The effective Hausdorff dimension of ω equals:
dim(ω) = inf{ s : some s-supermartingale succeeds on ω }

An s-supermartingale is like a martingale but with a different fairness condition: d(x) ≥ 2s−1 (d(x0) + d(x1)). The parameter s controls how "aggressively" the gambler bets. When s = 1, this is a standard martingale. When s < 1, the gambler has an advantage — the game is biased.

Sequences with dimension s < 1 are "partially random" — a gambler with an s-advantage can make money, but a fair gambler cannot. The dimension measures the boundary between exploit and no-exploit.

Check: A sequence with effective dimension 0.5 means...

Chapter 8: Martingale Betting Simulator

The simulation below shows a gambler using a martingale strategy against a sequence. Against a random sequence, the capital fluctuates but doesn't grow. Against a patterned sequence, the gambler can exploit the pattern.

Martingale Capital vs. Bit Number

The gambler bets on each bit. Against random sequences, capital stays bounded. Against biased or patterned sequences, the gambler's capital grows.

What to notice: Against random sequences, capital does a random walk — sometimes up, sometimes down, but on average flat. Against biased sequences, a gambler who knows the bias bets accordingly and profits steadily. Against periodic patterns, the gambler can predict and profit exponentially. This is the gambling characterization of randomness.
Check: Against a truly random sequence, a fair martingale's expected capital is...

Chapter 9: Summary

ApproachKey ideaRandomness notion
Von MisesSubsequence frequenciesKollektiv (informal)
ChurchComputable selection rulesMises-Church (weakest)
VilleShowed selection rules insufficientMotivated martingales
MartingalesBetting strategies / capital growthDepends on class
SchnorrComputable martingalesSchnorr randomness
Schnorr's thmL.s.c. martingalesML-randomness
Looking ahead: Chapter 10 develops the theory of information inequalities — linear inequalities relating entropies of groups of random variables (or complexities of groups of strings). This connects Kolmogorov complexity to combinatorics, polymatroid theory, and the structure of information itself.
Final check: The hierarchy from weakest to strongest randomness is...