Skip to content
@northguild/gmt

A Delay Across the Fall-Back

  • Legs, dwell, cut-offs and punctuality, shared by road, rail, sea and air.

A Delay Across the Fall-Back

A train is due in New York at 01:30 EDT on Sunday 3 November 2024. At 02:00 the clocks go back to 01:00, and the train arrives at 01:30 EST. How late was it?

The naive approach

// The naive approach: subtract the times on the arrivals board
const minutes = (hhmm) => { const [h, m] = hhmm.split(":").map(Number); return h * 60 + m; };
minutes("01:30") - minutes("01:30"); // 0 — "on time", across the fall-back
minutes("03:30") - minutes("01:30"); // 120 — "two hours late", across the spring-forward

Watch it break

The naive approach silently produces wrong results — there is no GMT equivalent to demonstrate because the failure is not using GMT.

Why

A wall-clock subtraction assumes every hour on the clock happened once. On a fall-back night 01:00 to 02:00 happens twice, so the same printed 01:30 names two instants an hour apart. On a spring-forward night 02:00 to 03:00 never happens. `scheduleDeviation` subtracts instants, so the delay is 60 minutes both times, and `classifyPunctuality` judges that against your tolerance.

The gmt approach

import { classifyPunctuality, scheduleDeviation } from "@northguild/gmt/transport";

scheduleDeviation("2024-11-03T01:30:00-04:00[America/New_York]", "2024-11-03T01:30:00-05:00[America/New_York]");
// "PT1H" — the second 01:30 is an hour after the first
classifyPunctuality("2024-11-03T01:30:00-04:00[America/New_York]", "2024-11-03T01:30:00-05:00[America/New_York]", { late: "PT15M" });
// "late"
scheduleDeviation("2024-03-10T01:30:00-05:00[America/New_York]", "2024-03-10T03:30:00-04:00[America/New_York]");
// "PT1H" — 03:30 EDT is an hour after 01:30 EST

Same widget, working

output 

Open the fall-back night in the Punctuality Board: read off the wall clocks, 3 of 4 trains are on time under a 15-minute tolerance; measured, 1 of 4 is. The rules are in Transport: Punctuality and Timestamps.