This is because old costs default (25) is longer than all heating
technologies (20). Script was distributing across 25 years, then
throwing out boilers older than 20 years, an inconsistent
behaviour.

Now existing boilers are smoothly distributed across 20 years.

This allows them to cover heat demand peaks e.g. 10% higher than
those in the data.

The disadvantage of manipulating the costs is that the capacity is
then not quite right. This way at least the costs are right.

Doing it properly would require introducing artificial peaks, but this
creates new problems (e.g. what is going on with wind/solar/other
demand).

for more information, see https://pre-commit.ci