Good post and good point.

It is fair to make an initial impression on probability until better data is available. That is sound reasoning, and essential.

However, when better data comes in, one needs to replace the older, lower-quality, data with the new data.