Shifts in matching projections

I'm running into a strange problem with Frealign. It seems that the SHX and SHY parameters have no effect on moving the reference projection from the center of the box. Is that correct? If so, what do I do if my particles are not perfectly centered? I checked the matching projections generated by setting FMATCH = T for two different parameter files (see below). In both runs I do not refine any of the parameters, hence I'm only calculating reference projections. In both cases, the matching projections are centered, even though the shifts are changing.

For the first run I use the following parameters:

C           PSI   THETA     PHI     SHX     SHY    MAG   FILM      DF1      DF2  ANGAST  PRESA   DPRES
      1    0.00   90.00    0.00    0.04    3.43  42857.     1  15933.0  16162.0  -26.28  99.33   49.33
      2    0.00   90.00    0.00   10.04    3.43  42857.     1  15933.0  16162.0  -26.28  99.04   49.04
      3    0.00   90.00    0.00   20.04    3.43  42857.     1  15933.0  16162.0  -26.28  98.33   48.33
      4    0.00   90.00    0.00   30.04    3.43  42857.     1  15933.0  16162.0  -26.28 103.35   53.35
      5    0.00   90.00    0.00   40.04    3.43  42857.     1  15933.0  16162.0  -26.28 107.30   57.30
      6    0.00   90.00    0.00   50.04    3.43  42857.     1  15933.0  16162.0  -26.28  79.72   29.72
      7    0.00   90.00    0.00   40.04    3.43  42857.     1  15933.0  16162.0  -26.28  80.34   30.34
      8    0.00   90.00    0.00   30.04    3.43  42857.     1  15933.0  16162.0  -26.28  87.25   37.25
      9    0.00   90.00    0.00   20.04    3.43  42857.     1  15933.0  16162.0  -26.28  81.50   31.50
     10    0.00   90.00    0.00   10.04    3.43  42857.     1  15933.0  16162.0  -26.28  79.70   29.70

I'm merely changing the SHX parameter in 10 pixel increments. However, the matching projections are not changing.

For the second run I use the following parameters:

C           PSI   THETA     PHI     SHX     SHY    MAG   FILM      DF1      DF2  ANGAST  PRESA   DPRES
      1    0.00   90.00    0.00    0.04    3.43  42857.     1  15933.0  16162.0  -26.28  99.33   49.33
      2    0.00   90.00    0.00    0.04   13.43  42857.     1  15933.0  16162.0  -26.28  99.04   49.04
      3    0.00   90.00    0.00    0.04   23.43  42857.     1  15933.0  16162.0  -26.28  98.33   48.33
      4    0.00   90.00    0.00    0.04   33.43  42857.     1  15933.0  16162.0  -26.28 103.35   53.35
      5    0.00   90.00    0.00    0.04   43.43  42857.     1  15933.0  16162.0  -26.28 107.30   57.30
      6    0.00   90.00    0.00    0.04  -53.43  42857.     1  15933.0  16162.0  -26.28  79.72   29.72
      7    0.00   90.00    0.00    0.04  -43.43  42857.     1  15933.0  16162.0  -26.28  80.34   30.34
      8    0.00   90.00    0.00    0.04  -33.43  42857.     1  15933.0  16162.0  -26.28  87.25   37.25
      9    0.00   90.00    0.00    0.04  -23.43  42857.     1  15933.0  16162.0  -26.28  81.50   31.50
     10    0.00   90.00    0.00    0.04  -13.43  42857.     1  15933.0  16162.0  -26.28  79.70   29.70

I'm merely changing the SHY by 10 pixel increments. Again, the reference projections remain centered within each box. Is this correct, or is this a bug. Thanks.