Category
Algorithms & Models
Mathematical and computational ranking methods — link analysis, pairwise comparison models, and voting or aggregation systems, explained conceptually.
- Borda Count — How Ranked Ballots Become a Ranking How the Borda count works: points per rank position summed across ballots, consensus-seeking behavior, vulnerability to strategic voting, and where it's used.
- Bradley–Terry Model — How Pairwise Comparisons Become Rankings How the Bradley–Terry model works: win probabilities from strength ratios, maximum-likelihood fitting, and its role from chess ratings to AI preference ranking.
- Condorcet Voting Method — How Pairwise Majority Ranking Works How Condorcet methods rank candidates: head-to-head majority duels from ranked ballots, the Condorcet paradox and cycles, and resolution rules like Schulze.
- Eigenvector Centrality — How It Ranks Nodes in a Network How eigenvector centrality works: importance from connections to important neighbors, the eigenvector equation, convergence via power iteration, and limitations.
- Glicko Rating System — How It Adds Uncertainty to Elo How the Glicko rating system works: ratings deviation (RD) beside every rating, volatility in Glicko-2, inactivity effects, and use by major chess platforms.
- HITS Algorithm — How It Ranks Hubs and Authorities How the HITS algorithm works: mutually reinforcing hub and authority scores, query-dependent ranking, eigenvector computation, and how it differs from PageRank.
- PageRank Algorithm — How It Ranks Web Pages How the PageRank algorithm works: the random-surfer model, link-based importance, the damping factor, update history, and well-documented limitations.
- TrueSkill — How Bayesian Skill Ranking Works How the TrueSkill ranking system works: skill as a Gaussian belief (mu and sigma), team and multiplayer support, factor-graph updates, and use in Xbox Live.