Long term Statistics

Beyond the live sensors, the integration imports long-term statistics under the suunto_app: namespace. These exist specifically to solve a problem a normal sensor can't: retroactive backfill. A sensor only records the value present at poll time; if your watch syncs to the app hours (or days) late, those past hours would be lost. Statistics are imported with their real timestamps, so a late sync fills the missed hours retroactively, giving a gap-free curve.

These are external statistics, not entities. They will not appear in your entity list, won't add to the sensor count, and are viewed in a Statistics Graph card (or ApexCharts) - see Dashboard Examples.

---

Why statistics (and not just sensors)?

• The dense workout heart-rate samples aren't in the live 24/7 stream, and they are backdated + sub-hourly. HA's only supported way to ingest backdated data is hourly statistics.
• A delayed watch→app sync means data arrives late. The import re-runs over a rolling ~5-day window each history cycle, so late hours are filled idempotently. (Data older than the window won't backfill.)

---

The statistics

Hourly (one bucket per hour)

Statistic ID	Type	Unit	Source
suunto_app:hr	mean / min / max	bpm	24/7 stream (~10 min) + dense workout HR (~25 s) folded in, so workout peaks show.
suunto_app:steps	sum	steps	24/7 only (workout totals already live in the 24/7 stream - adding them would double-count).
suunto_app:energy	sum	kcal	24/7 only.
suunto_app:recovery_balance	mean	%	Recovery stream (~30 min).
suunto_app:stress	mean	-	Recovery stream.

Daily (one point per night/day)

Statistic ID	Unit	From
suunto_app:sleep_duration	h	sleep series
suunto_app:sleep_hrv	ms	sleep series
suunto_app:sleep_rhr	bpm	sleep series
suunto_app:sleep_quality	%	sleep series
suunto_app:sleep_spo2	%	sleep series
suunto_app:readiness	- (0-100)	Derived Metrics
suunto_app:fitness_ctl	-	training-load model
suunto_app:fatigue_atl	-	training-load model
suunto_app:form_tsb	-	training-load model

The daily metrics come from data already fetched (60-day sleep series, 90-day workouts), so a late-synced night or day fills in retroactively too. The CTL/ATL/TSB trend emits the recent ~60 days (stable past the seeded ramp).

---

Viewing them

Add a Statistics Graph card (Dashboard → + Add Card → Statistics Graph) and pick the suunto_app:* ids. The HR statistic is the way to see a gap-free intraday HR curve with workout peaks - the live current_hr sensor cannot be filled backwards.

Minimal example (full set in Dashboard Examples):

type: statistics-graph
title: Heart rate (intraday)
chart_type: line
period: hour
stat_types:
  - mean
  - min
  - max
entities:
  - suunto_app:hr

type: statistics-graph
title: Fitness / Fatigue / Form
chart_type: line
period: day
stat_types:
  - mean
entities:
  - suunto_app:fitness_ctl
  - suunto_app:fatigue_atl
  - suunto_app:form_tsb

You can also inspect/clean them in Developer Tools → Statistics.

---

Notes & limits

• Backfill window ~ 5 days. A sync delayed beyond that won't fill the part older than the window.
• The import runs in the history coordinator (hourly cadence is plenty) and is wrapped so an import hiccup never blanks the normal sensor update - but a genuine auth failure still triggers reauth.
• Requires the Recorder integration (declared as an after_dependencies).