@@ -277,8 +277,8 @@ Until that sensor has a week of history, for example just after upgrading, the o
 The history is scaled by `pv_scaling` before the comparison, so both sides are on the same basis - to sanity-check the calibration factor by hand, divide your actual generation by the raw provider forecast **multiplied by** `pv_scaling`
 (equivalently, divide your actual-over-raw ratio by `pv_scaling`); comparing against the raw figure on its own leaves your answer out by a factor of `pv_scaling`.<BR>
 Do not use if you are using the [Solcast integration and have turned on the integration's auto dampening](https://github.com/BJReplay/ha-solcast-solar?tab=readme-ov-file#dampening-configuration).<BR>
-Predbat relies upon your solar generation being accurate so if your export generation can be curtailed by your solar inverter or your electricity supplier in periods when there is excess electricity in the grid,
-then you must turn PV calibration Off as otherwise Predbat will model the chopped solar generation as a PV calibration factor and will significantly reduce your forecast PV generation, leading to a very inaccurate plan.
+PV production calibration relies upon your actual solar generation reflecting the maximum your panels could produce. If your inverter can curtail or clip your solar generation, either through export limits (including zero-export setups) where generation is throttled because your battery is full and there is nowhere for the excess to go, or under instruction from your electricity supplier in periods when there is excess electricity in the grid,
+then you must turn PV calibration Off. Otherwise, Predbat will model the chopped solar generation as a PV calibration factor and will significantly reduce your forecast PV generation, leading to a very inaccurate plan.
 
 Note: If you change the PV calibration enable switch (to On or Off), you will need to restart Predbat for the change to take effect.