Radar charts

TL;DR: Compare world mean and median profiles on radar spokes normalized to each metric's p90. Full explanation at the end ↓

% const AXES = [ % “dangerousCreatures”, % “chests”, % “rareMaterials”, % “pouches”, % “hearts”, % “wands”, % “potions”, % “shops”, % “greatChests”, % “mimics”, % “hvSpells”, % ]; % // The 11 axes, in scheme order.

Four charts — mean and median overlaid

One chart per world grouping; each shows the mean and the median as two overlaid series. Where they diverge (the small, lumpy axes like great chests or rare materials, whose median is often zero), the gap is visible directly.

Eight charts — one per world × stat

The same data, fully separated: each world grouping rendered once for the mean and once for the median.

Full explanation and methodology

The chart compares population means and medians across the loot-and-feature axes. Each spoke is divided by that metric’s 90th-percentile value so metrics with very different raw count ranges can share one readable chart.

A spoke at the rim means the displayed mean or median equals the threshold exceeded by only the top 10% of seeds for that metric. Rare individual seeds can extend beyond the reference. Hover points to read the underlying raw values.