I can confirm that I have as well such weird numbers for data objects, checked some spools from 2008 and 2011, some were pretty close to the archived batches just 1 record difference, others were apart by many thousand.
I personally look more at the table statistics, how much records got removed from MCH1 and MCHA etc.
I think this numbers are retrieved with function module ARCHIVE_GET_INFORMATION
Open it in SE37 and click the button for documentation on top right.
then scroll down and click the hyperlink for
SAP explains there how they get the object number, they talk about a "handle", so I think this object number is pretty useless information for a user.