Sampling Safari · MCMC Trails
Explore three probability landscapes and tune four real Markov-chain samplers to collect every golden landmark in as few samples as possible. Banana Bend, Ring Lagoon and Three Islands; random-walk Metropolis-Hastings, Hamiltonian Monte Carlo, Langevin and Gibbs; original animated proposals, accept/reject trails, density contours and marginal histograms. Pause, single-event step, fast trail, live parameter controls, gold/silver medals and device-local passport. Landmark collection is a playful exploration goal, not a convergence diagnostic. Static adaptation of Chi Feng’s MCMC Interactive Gallery. Source: https://github.com/chi-feng/mcmc-demo/tree/c7fd6fa79821a83c94d5835b977d5fb376a9916d (948 GitHub stars observed November 7, 2026; MIT; pushed August 28, 2026). Four source samplers and original target mathematics, linear algebra and renderer retained; native Safari controls replace dat.GUI and external seedrandom, popups removed, all assets local. Metadata tags: github-source, game.
Every public app is built from a security-scanned skill and must pass a clean scan — skill and frontend — before it can be listed. Have a skill of your own? Turn it into an app — or read the step-by-step walkthrough.