Wall-Touch Response algorithm
Comfort Memory
Learns a small time-of-day target bias from repeated safe wall choices and re-applies it later that hour.
These algorithms exist for the exact household fight AC Defender is built for: someone keeps raising the thermostat, but the room still needs to come back to your temperature without starting a visible duel.
The short version
Learns a small time-of-day target bias from repeated safe wall choices and re-applies it later that hour.
What it watches
The current hour and the offsets learned for it; the room temperature.
How it decides
Repeated safe touches teach a bounded offset (± max offset) for the current hour slot. On later checks in the same window it nudges the target by that learned offset. Learned memory expires after the retention hours and is skipped when the room is warm or upstairs needs cooling.
What it changes
Adjusts the defender target by the learned hourly bias.
Safety boundaries
- Uses the real inputs listed above. It does not invent thermostat, weather, usage, or sensor state.
- Changes only the output listed above. Thermostat-affecting work goes through Home Assistant or returns a real error.
- The global AC Defender rules still apply: the website target remains the floor for cooling commands, the worker keeps refreshing real Home Assistant state 24/7, and comfort/safety rules are not bypassed by decorative timing.
Settings
ComfortMemoryEnabledComfortMemoryLearningTouchesComfortMemoryRetentionHoursComfortMemoryMaxOffsetCelsiusComfortMemorySafetyBandCelsius
Where to see it
- Defense page: live card with state, verdict, evidence, and metrics.
- Guide page: generated from the same guard catalog entry.
- Source:
Guards/GuardCatalog.csdescribes this page; the implementation is coordinated byServices/DefenderStateStore.csandServices/AcDefenderService.cs.
Failure modes
If Comfort Memory cannot obtain one of its required real inputs, it reports a blocked, held, or unavailable result and leaves the background worker’s Home Assistant refresh running. It never fills a missing room reading, audit event, weather sample, usage value, or device state with a simulator value. If a real Home Assistant command is rejected, the user sees the service’s actual error and the article’s surface remains available for recovery.
Security considerations
This feature consumes only the configured Home Assistant entity data, local settings, and the audit context named above. Tokens and credentials stay in the server environment; the static documentation site does not collect analytics, transmit search text, or embed third-party assets. Logs and exports should be reviewed before sharing because real entity names and timestamps can identify a household.
Verification
Verify the shipped behavior at the feature’s live page or endpoint, then run the repository’s documented build and test commands. Confirm the real-input and real-error paths, keyboard access, reduced-motion behavior, and a 390 px viewport without horizontal overflow. Record the exact commit and workflow result when publishing a release; a static screenshot alone is not proof of a live Home Assistant command.
Suggested articles
- Feature briefs — find every documented surface and guard.
- Defender Logic — follow the complete decision cycle and its bypass rules.
- Settings — inspect persisted configuration, language modes, and safety limits.