I'm worrying about..

MACHO coordinates..

Coordinate calibration using crossmatched USNOB1.0 stars

doesn't improve MACHO RA&Dec at all..

Residual between MACHO and the crossmatched USNOB1.0 stars (in one MACHO field) are completely random..

What the hell..

update #1. after a bit hard work, I got some improvements over the original results at outside fields of LMC. But.. still.. the center region of LMC are problems..