While optimizing some functions last night, I managed to introduce a bug in the CENTITIME processing function that mangled times with centiseconds. I fixed the bug, but anyone who submitted a CENTITIME stat since last night may have been hit by it. If you submitted any stats with centiseconds since December 22nd at 11:29 PM EST (but before December 23rd at 6:45 PM EST), please double-check those stats to make sure that they are correct.
EDIT: In case anyone doesn't know what CENTITIME is:
TIME = 1:23
CENTITIME = 1:23:45 <-- this is the one that broke
MILLITIME = 1'23"456