Skip to contents

Find Best Effort for Target Distance

Source: R/calculate_pbs.R
find_best_effort.Rd

Locates the fastest contiguous sub-segment that covers target_distance within a stream. Two robustness improvements over the original nearest-row approach:

Usage

find_best_effort(stream_data, target_distance)

Details

  1. Strictly-monotonic-distance filter: the input is reduced to rows where cumulative distance and time both strictly increase, so spurious backward jumps (GPS glitches, laps restarting the distance counter) no longer contribute fake short intervals.

  2. Linear interpolation at the target distance: the start and end points of each candidate window are interpolated between recorded samples, eliminating a systematic bias toward slower efforts that nearest-row lookup introduced on low-Hz streams (a ~0.1 Hz Garmin smart-recording sample could previously inflate a 5 km time by up to ~10 s on each side).

  3. O(n) two-pointer sweep: replaces the previous O(n^2) scan across start indices with a single left-to-right pass, making per-activity PB extraction tractable on multi-hour ultras.