The buffering to dose against. This is the median of the brewer's own
readings once there are two, and the model estimate until then.
We use the median because brew-day mistakes tend to be large and
one-sided: a meter left in the wort while it heated, a dose only half
stirred in, a reading taken before the acid had mixed. Any of those gives a
figure far from the true one, and a mean would carry it into the next
brew's dose. One reading on its own can't be told apart from a good one, so
we wait for a second before using either.
The buffering to dose against. This is the median of the brewer's own readings once there are two, and the model estimate until then.
We use the median because brew-day mistakes tend to be large and one-sided: a meter left in the wort while it heated, a dose only half stirred in, a reading taken before the acid had mixed. Any of those gives a figure far from the true one, and a mean would carry it into the next brew's dose. One reading on its own can't be told apart from a good one, so we wait for a second before using either.