It is likely due to your automation not catching those rapid state changes. When the door opens and closes in quick succession, the physical sensor might correctly switch states, but the automation might not execute fast enough to keep up, resulting in the helper sensor getting stuck in an incorrect state (often remaining "on" even when the door is actually closed).
To fix this, you can introduce a delay condition in your automation logic. This ensures that the door must stay in the "open" or "closed" state for a few seconds before the automation reacts. For example, you can require the door to remain open or closed for at least 5 seconds before updating the helper sensor. This prevents brief, unintended or transient state changes from triggering your actions. In YAML, this is done by adding a for: 5s parameter to the starters block, so the system only runs the command if the door remains in that state for a full 5 seconds.
If your automation platform doesn’t support using for directly in the trigger, you can instead introduce a short delay within the action block itself. This way, after the door changes state, the automation waits a few seconds before updating the helper sensor. During that wait, if the door changes back, the action will not execute, effectively filtering out quick toggles.
Lastly, make sure all devices involved are reliably connected and responsive. Missed or delayed automation steps can also happen due to network lags or hardware performance issues. Checking logs or event history might help identify if the automation is even firing when expected.
