Training load¶

Training load is a single number for how much a workout took out of you. fitdocs computes it with a built-in calculator called threshold, and shows its work on every page.

The scale¶

All of fitdocs' load numbers use the standard scale: one hour at your threshold intensity = 100. An easy 45-minute run might score 40. A hard two-hour ride might score 150. Because every sport uses the same scale, a week of running, cycling, and hiking adds up to one meaningful total.

Three channels, one pick¶

For each activity, the threshold calculator computes a load from every channel that has enough data:

Then it picks one using a per-sport priority order, and uses only that number for totals, history, and training blocks. If load is computed from multiple channels, only the first is used., but the others are visible on the workout page. Default orders:

If the first channel in line can't be computed (no power meter that day, or a threshold you haven't recorded), the reason is recorded and the next one is used.

Change the order in fitdocs.toml:

[load.priority]
run = ["power", "pace", "heart_rate"]   # trust the footpod over GPS pace

Custom load calculators can be added as plugins.

What it needs from you¶

Load is measured against your thresholds, which fitdocs stores in athlete.toml in your data root:

• Resting and max heart rate
• Threshold pace, FTP, and threshold heart rate, per sport

There are three ways to fill them in:

1. Answer the prompts. When fitdocs sync or fitdocs load needs a value it doesn't have, it asks, checks your answer, and saves it. If the activity is older than your answer, it also asks whether that answer should apply to earlier activities too.
2. Let fitdocs work them out from your races. Tag your races and hard efforts, and fitdocs derive-benchmarks turns them into dated thresholds. See Benchmarks and race tags.
3. Write them by hand:

# <data-root>/athlete.toml
[[benchmarks.athlete.max_hr_bpm]]
value = 190
measured_on = 2026-06-01

[[benchmarks.athlete.resting_hr_bpm]]
value = 48
measured_on = 2026-06-01

[[benchmarks.run.threshold_pace_s_per_km]]
value = 240            # 4:00 /km
measured_on = 2026-09-01

[[benchmarks.ride.ftp_watts]]
value = 270
measured_on = 2026-08-15

Every benchmark has a date, and each activity is scored with the value that applied on that day. A threshold you set this week doesn't change how your runs from 2022 are scored.

Data Quality¶

If you’ve worn a fitness wearable for any amount of time, you’ve likely encountered bad data. Using poor quality training data to plan future workouts leads to confidently wrong decisions. fitdocs evaluates the data quality of each channel, and adds a warning to the page of any flagged workout:

• Cadence lock. Wrist heart-rate sensors sometimes pick up your running cadence and report it as your heart rate. fitdocs checks whether "heart rate" is following cadence suspiciously closely.
• Channels disagree. If the picked channel and heart rate disagree sharply about how hard the session was, the page flags it.
• Aerobic drift. Heart rate rising steadily against steady output is flagged when it passes a threshold.
• Stale threshold. A benchmark older than 12 weeks (configurable) gets flagged. The load is still computed, but you'll know it rests on an old number.

Running the load pass yourself¶

Load runs automatically at the end of every sync and regen. You can also run it on its own:

fitdocs load                    # fill in pages that don't have a load yet
fitdocs load --recompute        # recompute pages that already have one
fitdocs load --no-prompt        # never ask; skip pages that would need an answer
fitdocs load --calculator ID    # use one specific calculator

Run fitdocs load after adding new benchmarks to fill in pages that were missing a number.

Tuning¶

Everything the calculator uses can be set in fitdocs.toml: how much data a channel needs before it counts, the priority order, the warning thresholds, and when a benchmark counts as stale. See Configuration.