Tracking4D starts out by going through the list of planes, calls the first detector with a cluster on it
reference_first and the last detector with a cluster
The track finding is based on a straight line between those two.
I have a case where the very first detector plane (a timing reference placed before the Mimosa telescope) has a much worse spatial resolution than the following planes used in the tracking. After experimenting a bit, I found that the track finding is much more successful if I make sure this plane is not used as
reference_first, so that the straight line reference track can be constructed from two telescope planes instead.
To steer this, I've added a parameter
exclude_from_seed to the config file.
Putting this here as a merge request in case you think it can also be useful for other users.