Correct. However, they usually get several readings a day for popular rides, so it’s enough for their sampling to be predictive.
For example, here is the data for yesterday at BTMRR (FYI: I believe you have to have a subscription to click the link):
As you can see, there were quite a few submitted wait times throughout the day (green dots), and pretty much all of them are well below the posted wait time (pink line). The article says BTMRR averages 27% of posted. I definitely hope that continues to be the case when I’m there! Not that I care necessarily if the posted waits are high, just that I hope the actual waits are low.