- formatting
- images
- links
- math
- code
- blockquotes
- external-services
•
•
•
•
•
•
-
Online Estimation: The Memory of an Exponentially Weighted Estimator
Deriving the finite and limiting effective sample size of an exponentially weighted estimator step by step, revisiting the 100-versus-199 experiment, and stating carefully what the memory interpretation does and does not establish.
-
Building Intelligent Agents for Connect-4: Verification and Benchmarking
How independent opening-book answers, mirror invariants and driver agreement provide evidence of correctness, followed by a paired timing comparison with stated statistical assumptions.
-
Online Estimation: The Covariance of a Weighted Mean
Deriving the covariance of a weighted sample mean line by line, stating the independence assumptions it needs, and showing why the covariance denominator must not be used to normalize the weights of a mean.
-
The Hexadecimal Digit Canon Challenge: Solution
A complete solution to the Hexadecimal Digit Canon challenge, including the canonicalization rules, combinatorial reasoning, optimization strategy, and a Python implementation.
-
Building Intelligent Agents for Connect-4: MTD(f) and Null-Window Search
How a distance-aware score supplies game-derived bounds, and how MTD(f) and a binary null-window driver reconstruct exact values from narrow alpha-beta probes.