Bootstrap Confidence Interval Coverage on Skewed Data: Percentile versus t Interval (2D)
Thirty repeated experiments drawn as intervals around the true mean, with Blender-rendered desk props and live coverage charts. The same shared model as the 3D board: t interval versus percentile bootstrap on skewed data.
Draw repeated samples from a skewed lognormal population and watch how often the t interval and the percentile bootstrap interval actually contain the true mean. Change skewness, sample size, confidence level and number of resamples, and see coverage and width against the promised level. A Blender-built data board with numerical checks (exact enumeration, Monte Carlo, table values) built in.
Thirty repeated experiments drawn as intervals around the true mean, with Blender-rendered desk props and live coverage charts. The same shared model as the 3D board: t interval versus percentile bootstrap on skewed data.
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install