Work to look at solving Gap Hits

1
Work to look at solving Gap Hits

• Question 1 Can we solve what hits should go
  together in a gap by looking at the entire gap or
  all gaps in a station
• Put code in to look at all combinations of gap
  coordinates
• For each coordinate in a plane, find out how many
  gap coordinates it made (in fiducial area, etc.)
• Loop over all combinations of gap coordinates
  (each coordinate paired up with ONE coordinate in
  same gap to make an x,y hit)
• For each combination
• Count how many coordinates in neighboring plane
  are used more than once (should never happen if
  solved correctly)
• Look at sum ( q1 q2 / ((q1q2)/2) ) (if
  incorrectly matched pairs, q values will not
  necessarily match
• See if a correct solution seems to stand out
  among other possible solutions
• Question 2 Can we determine what clusters were
  fit to the wrong number of tracks and refit to
  get better hit positions (using above or
  something simpler)?
• Trying to decide what we can do here
• If fit too few tracks
• Should have q1-q2 wrong for hits in the cluster
• Should have at least one unmatched hit in
  neighboring plane which crosses this strip
• If too many tracks fit, similar problems
• Do we need to do (1) to address (2)?

2
Some Del(q) Distributions

• Plot on left is sum(delq), Plot on right is
  sum(delq)/nfit
• Events are events where the same number of hits
  fit on both planes (more likely to be correct
  fit) and no clusters were used twice
• Certainly looks like there is a class of events
  where the solution is much better than others
• Need to see how well this corresponds to correct
  solution (HIJING), how to best use this info

3
UsedTwice Distributions

• Certainly most combinations dont look good (1 or
  more clusters used twice)

4
Look at x,y positions in 3 gaps

• J/psi mixed with real data, look at station 1
  octant 0
• From nfit in each plane, looks like 3 tracks in
  each half octant here
• Looks like a few areas clearly match up some
  ambiguity because not exactly the same hits in
  all gaps. Just because clusters fit incorrectly
  or because of dead areas? Need to figure out
  how to handle this if we use this info

5
50 Central HIJING Events

• Even without solving the problem absolutely, can
  we get a handle on whether something has been fit
  wrong?
• Look at number of fit coordinates in plane 1
  nfit in plane 2 for
• Both Planes fit Correctly At least
  One Fit Incorrectly