WIP - Update references for changes in associators

This MR contains an update to the Brunel reference files that seems to have been triggered by the change in VP associators (though I cannot be sure, as it was not spotted before those changes were merged). The only things I am sure of:

• The change happened some time between 10th January (when the reference files were last updated) and 20th January (when I first spotted the problem in the nightlies)
• The pattern recognition is not affected (according to PrChecker, number of tracks found in each category is unchanged)
• There are big changes in the number of tracks classified as ghosts, in all categories. Hence my suspicion that this has something to do with MC association.

@raaij @sstahl could the problem have been introduced by Lbcom!205 (merged)? (It can be reproduced by running the Brunel/master tests against lhcb-head nightly)

@graven @sponce @lgrillo @fpolci @gdujany FYI in case you have any other bright ideas...