An honest field instrument for reasoning about where a wildland fire started.
You walk the burn with a map, flag physical fire-pattern indicators, and record the direction each one points. Backtrace fuses those bearings — each with its own angular uncertainty — into a probability field for the origin: a heatmap with credible regions, not a single false-precision dot.
The restraint is the whole idea. Peer-reviewed testing found fire-pattern indicators carry roughly 103° of mean directional error (Parker & Babrauskas, 2024), so a tool that collapses to a confident coordinate would be lying. Backtrace stays deliberately broad when the signs disagree, splits into two regions when the data supports two origins, and never prints a bare point — only a candidate area, sized honestly.
Under the hood: a von Mises grid posterior over a local ENU tangent plane (all math in meters), rendered as stepped credible-region bands (50 / 68 / 95%). It runs entirely in the browser as an installable PWA — offline, no account, no server — and saves and loads investigations as plain JSON files.
Built by Charlie Ramus, a Colorado wildfire field contributor, as a study in making forensic geometry legible without overstating it.
Leaflet + OpenStreetMap · TypeScript · von Mises grid estimator · offline PWA
See NOW.md for status and CRESEARCH.md for the math and geodesy.