What changed
Independent errors are easier to average than errors that depend on the past. Azuma studies sequences whose next increment has zero conditional mean, together with boundedness or conditional exponential-moment assumptions. These conditions control weighted sums and how their fluctuations grow. The paper gives convergence and iterated-logarithm upper bounds under explicit restrictions on the weights. This adds a route for studying dependent data without pretending that every dependent sequence behaves like independent noise.
Scientific scope
Independent errors are easier to average than errors that depend on the past. Azuma studies sequences whose next increment has zero conditional mean, together with boundedness or conditional exponential-moment assumptions. These conditions control weighted sums and how their fluctuations grow. The paper gives convergence and iterated-logarithm upper bounds under explicit restrictions on the weights. This adds a route for studying dependent data without pretending that every dependent sequence behaves like independent noise. The 1967 weighted-sum and asymptotic contribution. The exponential estimates do not erase the martingale extension already noted in the 1963 Hoeffding record. Theorem2 uses D_n^2=sum a_i^2, a_n^2/D_n^2->0 and D_n^2->infinity; Theorem3 uses positive increasing weights with a_n/A_n=o(1/log log A_n). These are upper/convergence conclusions, not an unconditional sharp two-sided LIL. Tsuchikura supplied a separating weight example.
Evidence
The linked original works support the scoped statement and contribution attribution. WRS has not independently replayed experiments or reconstructed a full mathematical proof. The 1967 weighted-sum and asymptotic contribution. The exponential estimates do not erase the martingale extension already noted in the 1963 Hoeffding record. Theorem2 uses D_n^2=sum a_i^2, a_n^2/D_n^2->0 and D_n^2->infinity; Theorem3 uses positive increasing weights with a_n/A_n=o(1/log log A_n). These are upper/convergence conclusions, not an unconditional sharp two-sided LIL. Tsuchikura supplied a separating weight example.
Contributors
- The named authors in the research record · SCIENCE: Source-listed authors for the exact contribution; experimental subjects are not autonomous scientific authors. Source support and predecessor work retain separate roles. Individual effort shares and unresolved identities are not inferred.
- T. Tsuchikura · SCIENCE: Supplied the separating weight-sequence example explicitly credited in the source; not a coauthor or claimant of all martingale results.
- Azuma: exponential control of bounded dependent weighted sums — original scientific article
- Azuma: exponential control of bounded dependent weighted sums — publisher deposit
Provisional WRS assessment, edition 2026-10-23.3. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction
A